Search

This work presents a dynamic posture tracking control strategy for wheel-legged systems on uneven surfaces. Based on the kinematic model of a wheel-legged robotic system, the expected positions for the end-effectors of wheel-legs are calculated according to posture references and sensor feedback. The position control problem for a general wheel-leg is investigated for the active mechanism to imitate a passive suspension and respond to the external contact forces. The position tracking accuracy of the wheel-leg is sacrificed to enhance the compliance performance under rough terrain. Because of the unique contact state with the uneven ground for each wheel-leg, the position responses are different. As a result, the forces from the wheel-legs to the fuselage are inconsistent, which leads to the risk of posture oscillations. Equipping the wheel-legs with an undirected communication network, a consensus scheme for the robotic system is developed with proven global asymptotic stability to improve the posture tracking property. A novel robotic system is established with Stewart-structured wheel-legs, which are connected by a user datagram protocol network. Comparative experimental results are carried out on the physical prototype to validate the effectiveness of the proposed approach. • An improved kinematic model avoiding geomagnetism disturbances is established for general wheel-legged robotic systems. Considering the contact forces, an active suspension controller is proposed for the single wheel-leg to track the desired position with expected compliance behaviors. • Based on an undirected communication topology, a consensus strategy is carried out for multiple ASCs. The CASC fulfills the coordination requirements and active suspension response demands with proved convergency property. • The proposed posture control scheme is applied to a wheel-legged physical prototype. The comparative experimental results between the proposed CASC and our previous studies are presented, which highlights the CASC in posture tracking accuracy and wheel-leg coordination performances. [ABSTRACT FROM AUTHOR]

In the framework of the backstepping algorithm, this article proposes a new function approximation technique (FAT)-based compound learning control law for electrically-driven robotic manipulators with output constraint. The Fourier series expansion is adopted in the learning-based design to approximate unknown terms in the system description. The accuracy of FAT approximation is also studied by defining an identification error, which is derived from a serial–parallel identifier. Furthermore, the output constraint is taken into account by integrating the error transformation, the performance function and the dynamic surface control in a compact framework. Following this idea, new compound adaptation laws are then constructed. The proposed compound learning controller confirms that all the signals of the overall system are uniformly ultimately bounded, ensuring the tracking error within the predefined bounds during operation. Different simulation scenarios applied to a robotic manipulator with motor dynamics illustrate the capability of the control algorithm. • A compound FAT-based prespecified performance learning control is proposed with output constraint. • The accuracy of FAT approximation is studied via identification errors obtained from the serial–parallel identifier. • The adaptive laws are proposed, which are derived from the error surfaces and the accuracy of FAT learning. • The system uncertainties are precisely followed by compound estimations, while the tracking errors remain within the prescribed bound. • Various simulations are given to validate the superiority of the proposed algorithm. [ABSTRACT FROM AUTHOR]

Technological advances have significantly fostered the use of robotics in neurosurgery. Due to their novelty, there is a need to develop training methods within neurosurgical residency programs that provide trainees the skills to utilize these systems in their future practices safely and effectively. We describe a detailed curriculum for trainees with significant responsibilities in the operating room, as well as hands-on and theoretical didactics. The curriculum for robot-assisted stereotactic electroencephalography (SEEG) and deep brain stimulation (DBS) electrode implantation technique and assessment tool has been designed based on Accreditation Council for Graduate Medical Education's (ACGME's) milestone requirement for surgical treatment of epilepsy and movement disorders. Residents were surveyed to assess their use of robotics in their surgical training. Since 2019, more than 100 patients have undergone robot-assisted SEEG and DBS depth electrode implantations at our institution. Residents and fellows were involved in all aspects of surgical planning and execution and were encouraged to take an active role during procedures. Didactic sessions led by experienced faculty are emphasized as important learning tools prior to hands-on experience in the operating room. The results of the survey show that residents receive more training intraoperatively as compared to training sessions, yet trainees would benefit from more instruction on informative cadaveric simulation sessions. Our curriculum was developed to become a structured tool for assessment of robotic education in neurosurgical training. This curriculum based on ACGME milestone requirements serve as a template for resident and fellow education in robotics in neurosurgery. [ABSTRACT FROM AUTHOR]

Gas leaks from subsea oil and gas facilities could cause significant ocean environment damage. Such leaks can cause fire and explosion, for example, a fire on the ocean surface west of Mexico's Yucatan peninsula. Detecting a gas leak is critical in managing fire and explosion risks. This study proposes using autonomous underwater vehicles -robotic fish- for gas leak plume detection. The robotic fish is equipped with advance two well-known deep learning models, Faster RCNN and YOLOV4. A physical experiment system of various sizes of underwater gas leaks is used to generate the benchmark dataset. The results demonstrated the YOLOV4 model has a stronger online real-time capability. It is 43 times faster than the Faster RCNN model with the same level of accuracy. This study verifies the feasibility of integrating deep learning models with the mobile vehicle for real-time autonomous gas leak detection. This contribution will enable the development of a safe and reliable digital twin of subsea emergency management. [ABSTRACT FROM AUTHOR]

• Scheduling and routing of autonomous mobile vehicles in robotic mobile fulfillment systems. • Integrated assignment and sequencing of inventory pods and orders to picking stations. • Development of MIP models for both the sequential and integrated solution approaches. • Investigation of impact of the chosen performance metric in the optimization process. • An integrated approach leads to more efficient fulfillment systems. In robotic mobile fulfillment systems (RMFS), mobile robots carry inventory shelves autonomously from the storage area to picking stations and back. The scheduling of these robots and the order picking activities can be modeled as a collection of interrelated optimization problems. In this paper, we focus on the following optimization problems: the order allocation to picking stations, order sequencing, the inventory pod selection, and the robot scheduling. We present new mixed-integer programming (MIP) models for these decision problems and extended on existing models from the literature. To improve the models further, we include interstation travel which means that inventory racks can be transported from one picking station straight to another station without returning it to the storage area in between, hence reducing the overall travel distance. Moreover, In previous research on RMFS, only some of these decision problems are integrated. Therefore, we developed an integrated model to study the interdependencies between the decision problems. The models are validated through simulations and different performance metrics are analyzed such as the number of pod visits, the total distance travelled, and the system makespan. Moreover, we introduce a cost metric to facilitate the objective evaluation of the system performance. From the computational experiments we conclude that the integration of the decision problems results in significantly better performing systems compared to sequentially optimizing them. However, this comes at the cost of more computational effort. Furthermore, including interstation visits can further improve the overall system performance. [ABSTRACT FROM AUTHOR]

This paper focuses on the impedance control for robotic manipulators with time-varying output constraints. High-order control barrier functions (HoCBFs) are firstly proposed for a nonlinear system with high relative-degree time-varying constraints. Then, the HoCBFs are introduced to impedance control for robotic manipulators, where the HoCBFs are employed to avoid the violation of time-varying output constraints in Cartesian space by quadratic program (QP), and the impedance control is designed to achieve compliance for human–robot interaction (HRI). In this way, the desired trajectory within the safety-critical region can be tracked without violating the output constraints due to the controller generated from QP, and the safe HRI can also be achieved because of the usage of impedance control method. Finally, simulation tests are conducted to verify the proposed control design methods. • The HoCBFs are proposed to handle the time-dependent safe set. • A HoCBFs-based impedance control is proposed for the constrained robotic systems. • The HoCBFs are employed to avoid the violation of time-varying output constraints. • The proposed scheme is designed to achieve the compliance of human-robot interaction. [ABSTRACT FROM AUTHOR]

The success of the developed software product largely depends on the correctly chosen architectural solution. Software architecture is a set of technical solutions that ensure that code meets the requirements of software. The growing adoption of cloud software development is bringing changes to current architectural decisions. The purpose of the article is to analyze modern architectural solutions, select the best architecture for developing a robotic system, as well as the concept of an improved solution for microservice architecture. This article covers the following architectural solutions: monolithic, service-oriented, and microservice. A solution to dividing the microservice architecture into two conceptual layers is presented, which eliminates some of the existing shortcomings. A new concept of building microservice architecture of the ALKETON robotic system is presented, which makes it easy to increase and modify the functionality, while maintaining "transparent" connections between the services of the system. A feature of this approach is the division of the microservice architecture into two conceptual layers API and APP. The developed architecture has shown high efficiency in the development of a robotic system. The designed architecture allows easy scalability due to transparent communication between microservices, which is critical in this rapidly developing industry. A robotic system ALKETON has been developed, aimed at reducing the cost of servicing front offices and attracting new customers of the age group. [ABSTRACT FROM AUTHOR]

Applications of socially emotional artificial intelligence are demanded in many practical areas, including psychological counselling and psychotherapy. A person who is potentially in trouble and needs help has to overcome an internal barrier before deciding to see a psychiatrist. On the other hand, obtaining an anonymous online consultation or taking an online test is psychologically easy and acceptable for the many. At the same time, state-of-the-art solutions of this sort are limited. Here a concept of an intelligent system is proposed that should help to alleviate the problem. The system can be implemented as an embodied actor based on a virtual environment, VR/XR, or a robotic platform with social-emotional capabilities. The actor is controlled by a cognitive architecture with a possibility of its replacement by a deep neural network model. Interaction with the user is based on a multimodal human-computer interface. Possibilities that will be provided by this virtual psychologist system will enable conducting an initial screening, the results of which can later be used at the next session with a specialist, if chosen by the user. The key issue of ethics, privacy and anonymity is discussed. Systems of this sort could be used beyond psychological counseling and are expected to have an impact on the healthcare system in general. [ABSTRACT FROM AUTHOR]

The paper presents the methodology and results of modeling the operation of the parallel mechanism - a 6-DOF robotic platform of the Stewart-Gough type. MSC Adams and Matlab software packages were used as modeling tools, as well as a general-purpose, high-level object-oriented programming language, Python. The digital layout, which has the properties of a parameterized simulation model, is built in the MSC Adams software package. The Python programming language is used as an alternative to the Adams View internal command language for creating and iteratively modifying modeling objects in Adams. Mutual integration of Matlab Simulink and Adams View provides the implementation of joint simulation and allows to perform kinematic, dynamic and force analysis of the mechanism, taking into account its design features through the use of a virtual prototype. Simulink provides the ability to import of physical and mechanical parameters of a solid model from Adams. The proposed implementation of the Python-Adams interface in the form of special procedures and functions automates the execution of a computational experiment and allows solving the problem of optimizing structural elements and finding the optimal geometric design of the platform in accordance with the selected optimality criteria. This is implemented through conducting a series of experiments consisting in sequential multiple changes in geometric parameters, followed by simulation model and analysis of the results in order to find design options that meet the specified criteria. The technique was tested on the example of one of the design versions of the platform when moving along several characteristic types of a given trajectory. The results of a computational experiment are presented and an analysis of the force characteristics during movement along a given trajectory is carried out. [ABSTRACT FROM AUTHOR]

This paper presents a formalization of the "following the leader" task and an implementation of the environment emulator based on that formalization for training a neural network model that solves this task. The emulator is based on the path planning algorithm that creates the leader route used in the training process of the neural network model. The paper presents a comparative evaluation of the speed of a number of path planning algorithms with the choice of the best one. In the created emulator, it is possible to set up an environment with a different number of obstacles, routes of different lengths and complexity, as well as set up the desired behavior of the agent following the leader. Due to the speed of the developed emulator, it allows training leader-following models based on reinforcement learning technology, the tuning process of which requires a large number of training iterations in various environments. [ABSTRACT FROM AUTHOR]

Employing a continuous-time control algorithm to control the practical system based on discrete-time digital computer will lead to the cost of performance degeneration. To address this issue, this paper proposes a discrete-time barrier Lyapunov function based controller for human–robot interaction in constrained task space to guarantee control performance. The Euler discrete-time stability of closed-loop system controlled by the proposed method is proved, and a feasible difference scheme to support the stability analysis is uncovered based on monotonic scaling. The parameter dependence of this study is well discussed, which involves sample interval and preset boundary of state constraints, and based on the architecture of barrier Lyapunov function, the dependence relationship is demonstrated by using analytical synthesis technique. With a certain sample interval, the proposal of controller parameters is qualified to guarantee that end-effector states are constrained with preset boundary. The discrete-time neural network estimation is designed to approximate the human being's behavior to rebuild the reference trajectory from the desired trajectory and impedance for smoothing the human–robot interaction. Controlled discrete-time states and estimated force are uniformly ultimately bounded, and the convergence vicinity around the origin is proven to be determined by sample interval, lumped uncertainty and preset boundary of state constraints. Numerical simulation and experimental results verify the effectiveness of proposed discrete-time barrier Lyapunov function based methods. • The proposed controller is capable of guaranteeing that the discrete-time constrained states are uniformly ultimately bounded. • The tolerance of error between the state and the boundary is expressed quantitatively. • The operation behavior is estimated by the one-step discrete-time NN technique. • The superior performance of the proposed method is confirmed by experimental results. [ABSTRACT FROM AUTHOR]

Accuracy and robust trajectory tracking for electro-hydraulic servo systems in the presence of load disturbances and model uncertainties are of great importance in many fields. In this work, a new adaptive sliding mode control method based on the RBF neural networks (SMC–RBF) is proposed to improve the performances of a robotic excavator. Model uncertainties and load disturbances of the electro-hydraulic servo system are approximated and compensated using the RBF neural networks. Adaptive mechanisms are designed to adjust the connection weights of the RBF neural networks in real time to guarantee the stability. A nonlinear term is introduced into the sliding mode to design an adaptive terminal sliding mode control structure to improve dynamic performances and the convergence speed. Moreover, a sliding mode chattering reduction method is proposed to suppress the chattering phenomenon. Three types of step, ramp and sine signals are used as the simulation reference trajectories to compare different controllers on a co-simulation platform. Experiments with leveling and triangle conditions are presented on a robotic excavator. Results show that the proposed SMC–RBF controller is superior to existing proportional integral derivative (PID) and sliding mode controller (SMC) in terms of tracking accuracy and disturbance rejection. • A new adaptive sliding mode controller based on the RBF neural network is proposed for improving trajectory tracking performances. • System models and nonlinear factors of the electro-hydraulic servo system are described in detail. • A nonlinear term and a chattering reduction method are synthesized into the control law. • A co-simulation platform is built to compare different controllers. • Effectiveness of the proposed controller is validated by simulation and experiment. [ABSTRACT FROM AUTHOR]

Nonlinear time-varying contact state is a crucial factor to prevent the traditional robotic belt grinding method from precision machining of blade. In this case, a novel selected force controlling method (SFC) with consideration of regional division (RD) based on machining allowance is proposed for improving robotic grinding accuracy of complex curved blade, on basis of the self-developed adaptive impedance controller. Ideal normal grinding force at each cutter-contact (CC) point is calculated by principal curvature radius and regional allowance of blade surface. Then, the CC points with similar ideal normal grinding force are divided into one region along grinding path based on the force threshold. Furthermore, an adaptive impedance controller with neural network online compensation algorithm (AICNN) is developed, and the verification test results of grinding four profile areas of intake side, exhaust side, convex and concave, indicate that the force control accuracy with AICNN has increased by 80.33%, 50.58%, 82.65% and 69.01% than that without the controller, respectively. Based on this, the grinding experiment of typical turbine blade is conducted with SFC, and the surface profile accuracy values at the four profile areas have evidently improved by 48.79%, 35.67%, 59.54%, and 66.90% than that with conventional grinding (CG), respectively. • The developed SFC method used for improving robotic grinding accuracy is essentially consisted of the regional division and force controller. • Regional division (RD) is mainly determined by similar ideal normal grinding force values of blade surface profile, which are calculated based radius of principal curvature and residual distribution at corresponding CC points. • AICNN force controller is achieved by developed adaptive impedance controller and proposed neural network online compensation algorithm, which is used to compensate force control errors in real time and control the normal grinding force. [ABSTRACT FROM AUTHOR]

We address the problem of designing a distributed algorithm for two robots that sketches the boundary of an unknown shape. Critically, we assume a certain amount of delay in how quickly our robots can react to external feedback. In particular, when a robot moves, it commits to move along path of length at least λ , or turn an amount of radians at least λ for some positive λ ≤ 1 / 2 6 , that is normalized based on a unit diameter shape. Then, our algorithm outputs a polygon that is an ϵ -sketch, for ϵ = 8 λ , in the sense that every point on the shape boundary is within distance ϵ of the output polygon. Moreover, our costs are asymptotically optimal in two key criteria for the robots: total distance traveled and total amount of rotation. Additionally, we implement our algorithm, and illustrate its output on some specific shapes. [ABSTRACT FROM AUTHOR]

• Flexible robotic arms composed of X-tensegrity inspired modules are proposed. • Each planar module comprises two sets of triangular bars linked by three strings. • Adjustable angles between modules facilitate the robotic arm's 3D expansions. • Our robotic arm can achieve various bending, V-, S-, and J-shaped deformations. • Theoretical model, simulations, and experiments confirm the arm's superior capability. Robotic arms have remarkable applications in diverse fields such as medical rehabilitation, disaster relief, and space exploration. Enhancing their rigidity, load-bearing capacity, and motion simplicity is key to broadening their usage. Utilizing the admirable flexibility and strength of tensegrity structures, made of rigid bars and elastic strings, we introduce a new type of flexible robotic arm. This arm is constructed using a sequence of two-dimensional X-tensegrity inspired modules. Each module comprises two sets of triangular bars linked by three strings, enhancing the arm's ability to deform and resist impact forces. The joints between modules are stiff, allowing for angular adjustments to create three-dimensional configurations with adjustable stiffness and curvature. Through theoretical analysis, simulations, and experiments, we have shown that this tensegrity-based robotic arm exhibits superior stability, flexibility, and scalability. [ABSTRACT FROM AUTHOR]

Butterfly wings possess distinct micro/nanostructures that contribute to their vibrant coloration, light-trapping capabilities, and sensitivity to various stimuli. These complex features have inspired the creation of diverse devices and systems, such as sensors, photovoltaics, photocatalysis, and robotics. Specifically, the wing scales of the Emerald Swallowtail (Papilio palinurus) display iridescent, polarization-sensitive, and retroreflective colors due to their hierarchical structures. However, current technologies fail to mimic these natural designs fully, limiting their practical application in everyday life. In this study, we introduce a groundbreaking method for fabricating artificial wing scales that emulate the biological structure's functionality with a much simpler geometry. By integrating self-graded lossy media into metallic micro-concavity arrays, we achieve pronounced iridescent effects in both coaxial and non-coaxial arrangements, while preserving retroreflective properties. In particular, the simplified design allows for switchable color patterns based on the viewing angle. Demonstrating the concept, we successfully employ these conspicuous retroreflectors in hydrogen gas detection and the bi-directional/switchable recognition of patterned signals. [ABSTRACT FROM AUTHOR]

The computer vision-based analysis of railway superstructure has gained significant attention in railway engineering. This approach utilises advanced image processing and machine learning techniques to extract valuable information from visual data captured in the railway track environment. By analysing images from various sources such as cameras, drones, or sensors, computer vision algorithms can accurately detect and classify different components of the ballast superstructure, including the catenary system support, rail surface and profile, fastening system, sleeper, and ballast layer. This enables the automated assessment of the railway track's condition, stability, and maintenance needs. This paper comprehensively reviews the recent advancements, challenges, and potential applications of computer vision techniques in analysing railway superstructure. It discusses various vision-based methodologies and machine-learning approaches utilised in this context. Furthermore, it examines the benefits and limitations of computer vision-based analysis and presents future research directions for improving its applicability in railway track engineering. • Explores advanced sensing technologies and robotics. • Discusses image preprocessing, segmentation, and detection methods. • Examines track components: ballast, sleeper, fasteners, rail, and catenary system. [ABSTRACT FROM AUTHOR]

Flexible pressure sensors are receiving a lot of attention due to their use in wearable electronic devices, human-machine interactive systems, and physiological signal monitoring. In this paper, we present a novel nanofiller embedded thin film based piezoelectric pressure sensor for finger tapping analysis. The active material is fabricated using poly (vinylidene fluoride) (PVDF) having zirconium oxide (ZrO 2) nanofillers. It is observed that the incorporation of ZrO 2 nanofillers has a substantial impact on the structural characteristics of PVDF polymer, leading to enhancements in its piezoelectric properties. Moreover, the PVDF/2 wt% ZrO 2 composite film sensor showed a notable maximum piezoelectric output voltage of ∼210 millivolts at a load of 20 g (0.88 kPa applied pressure). A four-fold enhancement in response is observed in case of PVDF/2 wt% ZrO 2 composite in comparison to pristine PVDF sensor. Further, a sensitivity of 0.255 ± 0.083 kPa−1 has been recorded for the composite based sensor. Additionally, maximum of 115 mV is obtained when the composite sensor is tapped manually. The recorded maximum values for impulsive pressure (38 g), high pressure (108 g), and low pressure (45 g) are 76 mV, 150 mV, and 90 mV, respectively. The sensor is efficiently able to detect robotic tapping and stress on the human fingers. An efficiency of around 90% was achieved in PVDF/ 2 wt% ZrO 2 at 5 kV/cm using PE hysteresis loop data. The findings indicate that PVDF/ZrO 2 composite based piezoelectric pressure sensor has promising capabilities for applications in wearable electronic devices and medical diagnostics. [Display omitted] • Nanofiller embedded thin film based piezoelectric pressure sensor pad for finger tapping analysis. • Active material is fabricated using poly (vinylidene fluoride) (PVDF) having zirconium oxide (ZrO 2) nanofillers. • A four-fold enhancement in response is observed in case of PVDF/2 wt% ZrO 2 composite in comparison to pristine PVDF sensor. • The sensor is efficiently able to detect the robotic finger tapping, human finger tapping and stress on fingers. [ABSTRACT FROM AUTHOR]

This work introduces a Graphene fabric-like groove pattern for flexible piezoresistive sensors. This unique strcture allows the sensor to maintain a high linearity and high sensitivity over a wide pressure range. The GNWs groove piezoresistive sensor exhibits unprecedented sensitivity (S= 2297.47 kPa−1) over an extensive pressure range (0.2 Pa − 425 kPa). Base on the GNWs groove sensor we built a high-resolution tactile imaging system which can reconstruct the texture of a leaf and reconstruct the shape of letters under soft elastomer. [Display omitted] • This work introduces a conformal GNWs fabric-like groove structure for flexible piezoresistive sensors. • The GNWs groove sensor exhibits high sensitivity (S = 2297.47 kPa−1) over an extensive pressure range (0.2 Pa − 425 kPa). • A high-resolution tactile imaging system was built. • 3D reconstruction of leaf surface texture owns a surface height resolution of 20 μm. • Tactile imaging system is capable of distinguishing elastic films with a thickness as small as 0.1 mm. Flexible pressure sensors hold significant promise for applications in fields such as prosthetics, electronic skin, robotics, and healthcare. Nevertheless, the challenge lies in preserving the linear sensitivity range of these sensors across a broad pressure range. In this study, we introduce a structure featuring Graphene Nano-walls (GNWs) arranged in a fabric-like groove pattern for flexible piezoresistive sensors. This design effectively accommodates material deformations, enabling the sensor to maintain exceptional sensitivity across a wide pressure range. The GNWs groove piezoresistive sensor exhibits ultra-high sensitivity (S = 2297.47 kPa−1) and outstanding linearity (R2 = 0.986) over an extensive pressure range (0.2 Pa − 425 kPa), while maintaining remarkable mechanical stability over 10,000 cycles. It boasts rapid response times of 9 ms and a recovery time of 7 ms. Owing to the fast response and high sensitivity of this sensor, we constructed a high-resolution tactile imaging system, with a surface height resolution of 20 μm and a elastic film thickness resolution of 0.1 mm. Noteworthy, we reconstructed the 3D texture of a leaf and reconstructed the shape of letters under soft elastomer. These applications underscore the potential of GNWs groove pressure sensors in various fields, including artificial intelligence, robotics, prosthetics, and so on. [ABSTRACT FROM AUTHOR]

• Design three path planning methods for mobile robots. • Combining the ACO algorithm and the improved artificial potential field method. • Improve the ant colony algorithm through firefly algorithm and heuristic function. • Provide technical support for path planning of mobile robots. The rapid development of robotics technology has driven the growth of robot types and the development of related technologies. As an important aspect of robot research, path planning technology plays an irreplaceable role in practical production and application. Ant colony algorithm has a wide range of applications in robot path planning, but there is also a problem of performance overly relying on initial parameter selection. In order to solve this problem and improve the performance of mobile robot path planning, an improved ant colony algorithm based on firefly algorithm was studied and designed in a two-dimensional environment. In order to further explore the performance of ant colony algorithm in solving robot coordinated path planning problems, an improved ant colony algorithm based on heuristic function was also designed. In a three-dimensional environment, an improved ant colony algorithm based on the improved artificial potential field method was designed. The research results show that the maximum running time of the improved ant colony algorithm based on the firefly algorithm in different grid environments is 819.36 s, 847.01 s, and 811.54 s, respectively. The average running time of the improved ant colony algorithm based on heuristic function in different grid environments is 5.19 s, 5.97 s, and 9.09 s, with average path lengths of 29.90 cm, 31.08 cm, and 37.01 cm, and path length variances of 0.35, 0.87, and 2.21, respectively. The ant colony algorithm based on the improved artificial potential field method has a running time of 1.930 s, 3.182 s, and 4.662 s in different grid environments, and a path length of 29.275 cm, 49.447 cm, and 67.057 cm, respectively. The ant colony algorithm for research and design optimization has good performance. The contribution of the research lies in the design of three path planning methods for mobile robots, including two-dimensional path planning and three-dimensional path planning, which improves the time of path planning and shortens the average path length. The novelty of the research is reflected in the design of a path planning method for mobile robots in two-dimensional and three-dimensional environments, which improves the ant colony algorithm through firefly algorithm and heuristic function, and combines the ant colony algorithm with the improved artificial potential field method. The method designed by the research institute can provide technical support for path planning of mobile robots. [ABSTRACT FROM AUTHOR]

Adaptive informative path planning (AIPP) is important to many robotics applications, enabling mobile robots to efficiently collect useful data about initially unknown environments. In addition, learning-based methods are increasingly used in robotics to enhance adaptability, versatility, and robustness across diverse and complex tasks. Our survey explores research on applying robotic learning to AIPP, bridging the gap between these two research fields. We begin by providing a unified mathematical problem definition for general AIPP problems. Next, we establish two complementary taxonomies of current work from the perspectives of (i) learning algorithms and (ii) robotic applications. We explore synergies, recent trends, and highlight the benefits of learning-based methods in AIPP frameworks. Finally, we discuss key challenges and promising future directions to enable more generally applicable and robust robotic data-gathering systems through learning. We provide a comprehensive catalog of papers reviewed in our survey, including publicly available repositories, to facilitate future studies in the field. • We survey learning-based methods for robotic adaptive informative path planning. • We introduce a mathematical framework for general informative path planning problems. • We establish taxonomies to overview the state-of-the-art in a structured way. • We discuss limitations and directions for future research. • We provide a comprehensive paper catalog with publicly available repositories. [ABSTRACT FROM AUTHOR]

This article presents the self-balancing slosh-free control (SBSFC) scheme, a notable advancement for stable navigation in food-serving robots. The uniqueness of SBSFC is that it does not require direct modeling of slosh dynamics. Utilizing just two inertial measurement units (IMUs), the proposed scheme offers an online solution, obviating the need for complex dynamics or high-cost supplementary systems. Central to this work is the design of a control strategy favorable for sloshing suppression, achieved through feedforward reference shaping and disturbance compensation. This means the SBSFC indirectly alleviates and compensates for sloshing effects, rather than directly controlling them as a state variable by relying on pixel-based measurements of sloshing. Key contributions include rapid slosh damping via reference shaping, robust posture stabilization through optimal control, and enhanced disturbance handling with a disturbance observer. These strategies synergistically ensure immediate vibration reduction and long-term stability under real-world conditions. This study is expected to lead to a significant leap forward in commercial food-serving robotics. [Display omitted] • Stabilizes the body of the mobile robot to reduce the possibility of sloshing. • Utilizes feedforward to suppress sloshing in response to step velocity commands. • Requires only two IMUs and a simple intuitive control logic with a small number of tuning parameters. [ABSTRACT FROM AUTHOR]

It is shown how nonlinear joint stiffness in industrial robots can be determined quickly and accurately through a combination of statistical linearization and optimized data acquisition configurations. The statistical linearization is carried out using the histogram of the measured motor torques. The result of this linearization is used in a criterion that is minimized to determine optimal configurations for data collection. The proposed approach is validated using data from both simulations and experiments with a medium-size industrial robot. In both cases, there is a significant improvement in accuracy compared to both using conventional linearization and collecting data in a larger but random set of configurations. [Display omitted] • Higher robot model accuracy using nonlinear transmission stiffnesses. • Experiment design algorithm improved by including statistical linearization. • Informative robot configurations are optimally selected from a set of candidates. • Benefits of experiment design and statistical linearization are shown by experiments. [ABSTRACT FROM AUTHOR]

The Human decision-making process works by recollecting past sequences of observations and using them to decide the best possible action in the present. These past sequences of observations are stored in a derived form which only includes important information the brain thinks might be useful in the future, while forgetting the rest. we propose an architecture that tries to mimic the human brain and improve the memory efficiency of transformers by using a modified TransformerXL architecture which uses Automatic Chunking which only attends to the relevant chunks in the transformer block. On top of this, we use ForgetSpan which is technique to remove memories that do not contribute to learning. We also theorize the technique of Similarity based forgetting to remove repetitive memories. We test our model in various tasks that test the abilities required to perform well in a human–robot collaboration scenario. [ABSTRACT FROM AUTHOR]

Robotic inspection of complex geometry components using immersion ultrasonic techniques is relevant for many ultrasonic Non-Destructive Testing (NDT) applications in different industries. To image large component, ultrasonic probe is manipulated around the component immersed in water using a robotic guided system to create a tiled image of the entire component. Due to the mechanical vibrations of the high-speed robotic arm, the probe position coordinates provided by the robotic system are not precise enough to ensure an accurate reconstruction (stitching) of a composite image from all individual images. In this work, we propose a stitching method specifically designed to create a single 2D image of the surface of an inspected component from a stack of Total Focusing Method (TFM) images. This stitching method uses Iterative Closest Point (ICP) registration to estimate a 2D rigid transformation between two consecutive 2D images, by maximizing the overlap between the surface geometries extracted from each image. The capabilities of this imaging technique are illustrated by various simulated and experimental results carried out in a water tank. Significant improvements in surface image quality, leading to accurate surface reconstruction, are shown for vertical vibrations with displacements of more than two operating wavelengths and for rotational vibrations with deviation angles of less than 1°. The results also show that the resolving power of the ICP algorithm decreases for strong rotational vibrations. [ABSTRACT FROM AUTHOR]

Learning from Demonstration (LfD) can be used to make it easier for robots to learn to perform tasks such as apple harvesting. The required apple harvesting motions can be split into four steps: 1. Approaching, 2. Grasping, 3. Detaching, 4. Placing. So far, only parts of the apple harvesting task have been learned using LfD. In our work, we learned an abstracted version of the complete apple harvesting task using a cube. We used two methods and tested the sensitivity to model parameters and the number of demonstrations. We selected (1) the Gaussian Mixture Model (GMM) with Gaussian Mixture Regression (GMR) and (2) the Hidden Markov Model (HMM) with Linear Quadratic Regression (LQR). Both methods make use of Task Parameterization (TP), which allows models to combine multiple phases from the apple harvesting task. Combining multiple harvesting phases is important for effective LfD. We analysed the performance of these algorithms on four performance metrics, namely the grasp pose, the detachment motion, the place pose, and the success rate when using the real robot. A high grasp pose accuracy was achieved, specifically a position error of 0.004 and an orientation error of 0.004 for the GMM with GMR, trained with 100 demonstrations and 9 Gaussian components. The detachment motion was performed but with a reduced range. The model that performed the best was the GMM with GMR, trained with 60 demonstrations and 5 Gaussian components. A high place pose accuracy was achieved, specifically a position error of 0.003 and an orientation error of 0.003 for the GMM with GMR, trained with 100 demonstrations and 15 Gaussian components. The task was executed by a real robot successfully, resulting in a success rate of 67% for HMM with LQR, trained with 40 demonstrations, 5 states, and a scaling factor of 1.0. For the number of components, no clear relation was found. The same was true for the number of states in the HMM with LQR. For the scaling factor, a clear relation could be found. For this task, a value in the range of -1.0 to 1.0 should be used. For the number of demonstrations, there was an optimum of 40 demonstrations. Further improvements are needed to deal with challenges such as apple shift and detecting detachment of fruits. • Learning from Demonstration was used to learn the apple harvesting task. • Methods were trained on at most 100 demonstrations. • High accuracy at grasp pose and place pose. • Effect of training with fewer demonstrations investigated. • Sensitivity analysis of the model parameters was performed. [ABSTRACT FROM AUTHOR]

Accurate detection and localization of fruits within the canopy are crucial for various tasks, such as perception for robotic harvesters, vision-based yield estimation, and early disease detection. However, complex canopy structures render fruit detection challenging due to obstructions from leaves, branches, and other fruits. This study assesses the efficacy of utilizing multiple cameras at varying elevations and azimuthal angles to enhance fruit visibility. We used a V-trellised apple orchard as a case study, where four cameras were employed; the first was oriented normally to the fruit canopy, while three others maintained a fixed position relative to the first but with variable orientation. Adjusting elevation and azimuthal angles revealed that, in the region of interest of the canopy, a single camera whose optical axis was perpendicular to the canopy could detect up to 88.3% of the fruits detected by four cameras. Adding one more camera – for a total of two cameras – increased the detection rate up to 97.5% of the four-camera detection rate. Also, the experimental results showed that three cameras could detect between 96.0% to 99.7% of fruits compared to four cameras under the proposed camera configuration. Hence, the utility of adding a third or fourth camera would need to be considered carefully, given the added cost and complexity. The results of this study are a step toward exploring and optimizing multiple-camera perception systems for orchard operations, in particular, robotic harvesters. • Multi-camera systems enhance visibility in V-shaped canopies. • A single camera detected up to 88.3% of apples compared to four cameras. • Two cameras detected up to 97.5% of apples compared to four cameras. • Four cameras show marginal 0.3% to 1.2% improvement compared to three cameras. [ABSTRACT FROM AUTHOR]

To boost productivity and efficiency, construction robots are anticipated to become increasingly prevalent in future construction workplaces. In such multi-robot environments, forming a well-composed robotic team is crucial for efficient task allocation and execution. However, the complexity of multi-robot dynamics, arising from the heterogeneity and varying scales of robotic teams, coupled with long-horizon tasks, presents challenges in estimating team utility and optimizing team composition. To tackle these challenges, a time- and utilization-guided simulation approach is proposed. Specifically, heterogeneous multi-robot dynamics is modeled as a cooperative multi-robot board game, and team utility is estimated using Monte Carlo tree search-based task scheduling. The optimal team composition minimizes completion time and maximizes robot utilization. Simulation results demonstrate the effectiveness of identifying the optimal team composition in three robot collaboration scenarios. This paper contributes to efficient and robust team utility estimation for robotic team formation in long-horizon construction tasks, addressing gaps in existing research. • Multi-robot team composition optimization for long-horizon tasks is presented. • Heterogeneous multi-robot construction is modeled as a cooperative board game. • Long-horizon task scheduling is approached based on Monte Carlo tree search. • Optimal team composition is determined by completion time and robot utilization. • Simulations of scaffolding assembly tasks demonstrate the robustness of the method. [ABSTRACT FROM AUTHOR]

Natural wood, renowned for its low carbon emissions, renewability, and impressive durability, is gaining prominence in the construction industry as a sustainable solution. However, the irregularities in size and shape of natural wood pose challenges in precise measurement and fabrication. Although innovative methods have emerged for materials like curved logs and small-diameter wood, tree forks have not been extensively investigated due to their complex shapes. This research introduces cooperative robotic fabrication to address the challenges associated with tree fork construction. An integrated workflow has been established, encompassing material scanning, co-design of fabrication tools and joint systems, and cooperative robotic fabrication. The proposed workflow was evaluated in the construction of a tree fork-based tower structure. This study provides a technological solution for the sustainable utilization of natural tree forks and holds potential for extending into the construction of other non-standard materials with geometric complexity. • This research presents an integrated workflow for natural tree fork construction. • Fabrication tools and connecting system are developed collaboratively. • A dual-robot cooperative fabrication system is established. • The workflow is evaluated through the construction of a case-study structure. [ABSTRACT FROM AUTHOR]

In this work, we present a general control architecture for robotic systems dedicated to the remote handling of in-vessel components in fusion machines. This control architecture will be tested for the inspection and maintenance of the first wall components of the RFX-mod2 experiment, in the scope of the New Equipment For the Experimental Research and Technological Advancement for the Rfx Infrastructure (NEFERTARI) project. The architecture is split into low-level and high-level layers. The former is used for the low-level control of the robotics systems and to manage safety; the latter is used for the high-level planning and implements several sub-modules, such as a Virtual Reality (VR) environment for the visualisation of the robot's digital twin. Moreover, an Application Programming Interface (API) is intended to connect the two layers, and the communication between modules and layers is provided by the ROS2 framework. A typical usage of such a framework involves a human operator who is teleoperating the real robot, data from motor encoders are used as inputs for the dynamic model module. This module is used to compute the forward dynamics in real-time, providing an accurate simulation of the robot (digital twin) in the virtual environment. • Generic control architecture for control of robotic manipulators used in remote handling tasks. • Description of the software control architecture, which is based on ROS2 Framework, dynamic simulation engine and VR for accurate digital twin. • Description of the NEFERTARI project and its robotic arm where the architecture will be implemented. [ABSTRACT FROM AUTHOR]

Recently, several movement control algorithms have been proposed to move robots to the desired locations and complete various collaborative tasks. These algorithms usually require information exchange among robots, which in turn relies on efficient neighbor discovery. However, neighbor discovery of robots is a challenging problem due to limited communication range and beacon collision. Here, how to set the time slot length and contention window to alleviate the beacon collisions is a major issue: too large or too small slot length and contention window values cannot achieve good performance in the robotic network. Therefore, we propose a neighbor discovery protocol with an adaptive collision alleviation mechanism, i.e. ND-ACA. ND-ACA adopts a simple mechanism to detect and avoid potential collisions before sending beacon, and adaptively adjusts both slot length and backoff window according to the number of historical beacon collections and beacon collisions. Extensive simulation results show the efficiency of our proposed approach. To the best of our knowledge, this is the first work in the literature that explore the joint implementation effects that simultaneously consider the robot mobility model and neighbor discovery. [ABSTRACT FROM AUTHOR]

Vision-Based Tactile Sensors (VBTS) play a key role in enhancing the accuracy and efficiency of machining operations in robotic-assisted precision machining systems. Equipped with VBTS, these systems offer contact-based measurements, which are essential in machining accurate components for industries such as aerospace, automotive, medical devices, and electronics. This paper presents a novel approach to virtual prototyping of VBTS, specifically in perpendicularity measurements using Computer-Aided Design (CAD) generation of VBTS designs, Finite Element Analysis (FEA) simulations, and Sim2Real deep learning to achieve VBTS with high precision measurements. The virtual prototyping approach enables an understanding of the contact between VBTS with different designs and machined surfaces in terms of contact module shape, thickness, markers' density. Additive manufacturing was employed to fabricate the molds of VBTS contact module, followed by experimental validation of the robotic arm to confirm the effectiveness of the optimized VBTS design. The results show that deviation from the hemispherical shape reduces the data quality captured by the camera, hence increasing the prediction errors. Additionally, reducing the thickness of the contact module enhances the precision of perpendicularity measurements. Importantly increasing markers' distribution density significantly enhances the accuracy of up to 92 markers at which above it the rate of improvement becomes less pronounced. An VBTS with height of 20 mm, thickness of 2 mm, and 169 markers was found to be within the stringent perpendicularity standards of the aerospace manufacturing industry of 0. 5 8 ∘ as a root mean square error, and 1. 64 ° as a max absolute error around the roll and pitch axis of rotation. The established virtual prototyping methodology can be transferred to a wide variety of elastomer-based sensors. • Establishment of parametric virtual prototyping for vision-based tactile sensors for perpendicularity measurements. • Methods, namely: CAD generation, FEA simulation, Sim2Real gap bridging, and practical implementation. • Constructed the relationship between sensor design and accuracy, achieving optimization. • Practical robotic implementation and validation of the optimized sensor show only 0.58 degrees error. [Display omitted] [ABSTRACT FROM AUTHOR]

• The model decipts spontaneous helical formation mechanism in linear structures. • Energy barrier prevents spontaneous curling between initial and final states. • Critical scale parameter is determined by nonlinear characteristic length. • A novel model is developed for flexible electronics and soft robotics. Tendrils initially exhibit straight morphologies before gradually transforming into helical conformations during longitudinal growth. Tendril-inspired helical structures possess advantageous physical properties and functions including versatile morphologies, adaptability, enhanced strength, and scalability. With advancements in nanoscience and nanotechnology, helical architectures across various length scales have been discovered or artificially synthesized. However, the fundamental mechanisms underlying the formation and evolution of helical structures remain elusive. This lack of mechanistic insight has impeded the controlled design and fabrication of helical structures, especially three-dimensional functional devices with integrated helical motifs. Here a mechanical model is systematically proposed for the spontaneous formation of the helical structures, based on the hypothesis of decoupled curling and warping phases. Applying nonlinear elasticity theory, a detailed analysis of the energetics and thermodynamics is presented, accounting for the large deformation that may emerge during planar curling. Additionally, differential relations between internal forces and associated displacements induced by bending and twisting of a slender curved beam under out-of-plane loading are derived, obtaining solutions via Laplace and inverse transforms. The theoretical solutions closely match ex- perimental evidence from shape memory polymers. Furthermore, the theoretical model is utilized to guide the fabrication of the smart helical antenna proposed in our previous work, and discuss the influence of configuration changes on the electromagnetic properties. Elucidating these fundamental mechanisms will facilitate the development of next-generation technologies exploiting helical architectures for diverse applications including flexible electronics, optics, and soft robotics. [ABSTRACT FROM AUTHOR]

Accurately measuring the force in friction stir welding (FSW) is essential for monitoring and controlling the welding process. However, current research primarily focuses on force sensing in gantry FSW equipment, which has limited demand compared to FSW robot. Commercial force sensors are mostly designed for six-axis welding robot and lack for high-force hybrid FSW robot. To address this gap, we developed a wireless multi-dimensional force sensing system for hybrid FSW robot. Our system utilizes elastomer and incorporates 8 pairs of strain gauges to detect tiny deformations. Through calibration, we achieved a wider measurement range than most commercial sensors, ensuring high accuracy with low cross-talk. Extensive testing determined the static and dynamic characteristics of the system. We performed robotic FSW experiments, analysing the force/torque characteristics in each direction and investigating the effects of welding speed and plunge depth on axial force. The results demonstrate the stability and practicality of our system. • Strain-based system measures multi-dimensional forces in RFSW. • 8 strain gauges in elastomer measure axial, traverse force, and tool torque. • Direct, accurate force measurement using spindle front-end form. • Wide range: 0-25 kN axial, 0-10 kN traverse, 0-100 Nm torque, <1% crosstalk. • Experimental validation shows accurate, reliable measurement in RFSW. [ABSTRACT FROM AUTHOR]

A relatively new goods-to-person order picking system (robotic compact storage and retrieval system, RCS/RS) has been implemented within order fulfillment centers. In such a system, robots move in x - and y -dimensions on top of a grid-based storage area to retrieve bins from stacks by "digging" in z -dimension. To transport a requested bin to a port, robots (or a robot) remove(s) all blocking bins above the requested bin and place(s) them around neighboring stacks. Defining the throughput of an RCS/RS as the number of bins processed per unit time, we analyze the time required for bins to be removed from stacks, transported to ports, returned to stacks, and placed in stacks. In making a case for considering skewness among activity levels for bins, we develop closed-form expressions to calculate the expected operation time for the movement of bins, validate the closed-form expressions using simulation, and obtain the system configuration that maximizes throughput performance (or minimizes operation time for bins) based on current practice. Then, with an objective of having an optimal RCS/RS design based on activity levels of bins, we show the significant impact on throughput when different skewness values for activity levels of the bins are incorporated into the design. [ABSTRACT FROM AUTHOR]

As one of fundamental problems in robotics, coverage path planning (CPP) requires the robot path to cover the entire workspace which has been employed in several essential applications such as cleaning robots, land mine detector, lawnmowers and automated harvesters. Unlike most of existing studies considering the CPP problem under a unrealistic assumption of infinity energy, this paper takes the battery limitation of robots into account. This poses a significant challenge for enabling an efficient coverage path while satisfying the limited energy constraint, even in a priori known environment. Handling this challenge, we propose a BWave Framework that guides the robot to move following an improved Boustrophedon-like motion and a special area prioritization and especially, to return a charging station effectively before an exhausted energy. To that end, a weighted map is applied for recognizing the special areas, namely trap regions, and governing the robot to enter these fields in priority. Moreover, a return matrix, which forms the shortest-path tree from the charging station, is pre-computed to not only validate the energy requirement, but also speed up the calculation process of return and advance paths during the robot's operation. We then evaluate BWave Framework extensively in various scenarios in both generated and real-life indoor maps datasets. The results show that compared to typical baseline methods, BWave Framework achieves the CPP solution at a significantly accelerated running time, namely 51.5 to 72.8 times lower for generated maps, and 44.8 to 255 times for real maps, while reducing the total path length by 2.4%–17.6% and by 2.9%–18.5%, respectively. Moreover, the proposed method also outperforms the baselines in terms of overlap rate, number of returns and accounts for a lower number of deadlocks. • This paper tackles coverage path planning with energy constraints. • It introduces BWave, a novel framework for complete coverage. • BWave has two phases: environment knowledge extracting and complete coverage. • BWave is evaluated in various scenarios and compared with state-of-the-art methods. • BWave performs well in environments with many trap areas. [ABSTRACT FROM AUTHOR]

The concept of a nano-laboratory inside an SEM involves performing a sequence of tasks in a continuous pattern. Robotic systems with nanoscale motion resolution facilitate in-situ manipulation and characterization of nanomaterials to assemble nanodevices inside SEMs. Precise motion control of micromanipulators is required at both macro and micro-nano scales to effectively execute multiple sequential experimental tasks in nanofabrication. However, managing the entire nanomanipulation setup is challenging due to the constricted workspace inside an SEM and the lack of proper process feedback information at that scale. This study explores the application of path planning algorithms to generate a collision-free motion path for the micromanipulators operating within the confined space of an SEM. A MATLAB-based computational tool is first developed using PRM and Dijkstra's path planning algorithms. Considering environmental constraints, the program generates an optimal motion path for the micromanipulators, facilitating automatic configuration changes within the SEM chamber. It ensures a seamless workflow and facilitates the smooth integration of additional experimental tools within the existing setup. Manipulation strategies using the nanorobotic setup are established based on the application of the developed path planning module. A pick-and-place 3D nanomanipulation technique of CNTs using cooperative control of dual micromanipulators has been demonstrated for nanodevice construction. Additionally, the electrical response of individually manipulated CNTs is recorded using a two-probe measurement technique. [ABSTRACT FROM AUTHOR]

• Interaction of human and socially assistive robot can have applicability in actual settings, and adoption to function in human environment for Alzheimer's disease. • To achieve this, the overall review revealed that there are still several technical challenges that must be addressed in future advancements before being integrated in real life of individuals with Alzheimer's disease. • The results will help move towards the establishment of best practices for the development of human and socially assistive robot interaction for Alzheimer's disease and, in turn, help enhance the experience of people suffering from this disease. The most common form of dementia, Alzheimer's Disease (AD), is challenging for both those affected as well as for their care providers, and caregivers. Socially assistive robots (SARs) offer promising supportive care to assist in the complex management associated with AD. To conduct a scoping review of published articles that proposed, discussed, developed or tested SAR for interacting with AD patients. We performed a scoping review informed by the methodological framework of Arksey and O'Malley and adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) checklist for reporting the results. At the identification stage, an information specialist performed a comprehensive search of 8 electronic databases from the date of inception until January 2022 in eight bibliographic databases. The inclusion criteria were all populations who recive or provide care for AD, all interventions using SAR for AD and our outcomes of inteerst were any outcome related to AD patients or care providers or caregivers. All study types published in the English language were included. After deduplication, 1251 articles were screened. Titles and abstracts screening resulted to 252 articles. Full-text review retained 125 included articles, with 72 focusing on daily life support, 46 on cognitive therapy, and 7 on cognitive assessment. We conducted a comprehensive scoping review emphasizing on the interaction of SAR with AD patients, with a specific focus on daily life support, cognitive assessment, and cognitive therapy. We discussed our findings' pertinence relative to specific populations, interventions, and outcomes of human-SAR interaction on users and identified current knowledge gaps in SARs for AD patients. [ABSTRACT FROM AUTHOR]

• A deep-learning-based flexible four-finger gripper grasping detection method. • A tool for marking pixel-level labels. • A powerful neural network. • Capable of grasping a wide variety of objects. The flexible four-finger gripper, as a specialized robotic end-effector, is highly valued for its ability to passively adapt to the shape of objects and perform non-destructive grasping. However, the development of grasping detection algorithms for flexible four-finger grippers remains relatively unexplored. This paper addresses the unique characteristics of the flexible four-finger gripper by proposing a grasping detection method based on deep learning. Firstly, the Acute Angle Representation model (AAR-model), which is based on the structure of the flexible four-finger gripper and consists of grasp points and angles, is designed as the grasping representation model that reduces unnecessary rotations of the gripper and improves its versatility in grasping objects. Then, the Flexible Gripper Adaptive Attribute model (FGAA-model) is proposed to represent the grasping attributes of objects, calculate the grasp angles that meet the criteria of the AAR-model, and aggregate the AAR-models on the image data into a unified set, thereby circumventing the time-consuming process of pixel-level annotation. Finally, the Adaptive Grasping Neural Net (AGNN), which is based on Adaptive Feature Fusion and the Grasp Aware Network (AFFGA), is introduced by eliminating redundant network detection headers, fusing color and depth images as inputs, and incorporating a Series Atrous Spatial Pyramid (SASP) structure to produce more accurate grasp poses. Our method not only attains a remarkable accuracy of 97.62% on the Cornell dataset but also swiftly completes grasping detection within 25 ms. In practical robotic arm grasping tests, where a robot is outfitted with a flexible four-finger gripper, it successfully grasps unknown objects with a 96% success rate. These results underscore the reliability and real-time performance of our method, significantly enhancing the gripper's adaptability and precision when handling objects of varying sizes and shapes. This advancement provides a powerful technical solution for robots utilizing flexible four-finger grippers, enabling autonomous, real-time, and highly accurate grasping maneuvers. Moreover, it addresses the persistent challenge of the scarcity of efficient grasping detection techniques tailored for flexible four-finger grippers. [ABSTRACT FROM AUTHOR]

Behavior trees (BTs) are increasingly popular in the robotics community. Yet in the growing body of published work on this topic, there is a lack of consensus on what to measure and how to quantify BTs when reporting results. This is not only due to the lack of standardized measures, but due to the sometimes ambiguous use of definitions to describe BT properties. This work provides a comprehensive overview of BT properties the community is interested in, how they relate to each other, the metrics currently used to measure BTs, and whether the metrics appropriately quantify those properties of interest. Finally, we provide the practitioner with a set of metrics to measure, as well as insights into the properties that can be derived from those metrics. By providing this holistic view of properties and their corresponding evaluation metrics, we hope to improve clarity when using BTs in robotics. This more systematic approach will make reported results more consistent and comparable when evaluating BTs. [ABSTRACT FROM AUTHOR]

We consider a square non linear parametric equations system F(P,X) = 0 which is constituted of n non differential equations in the n unknowns { x 1 , ... , x n } that are the components of X while P = { p 1 , ... , p m } is a set of m parameters that play a role in the definition of the equations F. We assume that P is restricted to lie in a bounded region and we are interested in developing a solver for obtaining all real solutions exactly (a notion that is defined in the paper) for any parameter values within the bounded region. The starting point of the proposed approach is that we assume that a numerical methods has allowed us to determine the real solutions (but not necessarily all of them) for a very limited number of fixed P called the initial solution set. Starting from this set we show that we can create multiple pairs (parameters, solution) and that these pairs may be structured into coherent learning sets that will be used to train multi-layer perceptrons (MLP). The training process is specific: although it still uses a decrease of a loss function its main objective is to maximize the success rate i.e. the number of occurrences, expressed in percentage of number of samples of the training set, for which the Newton scheme, initialized with the MLP prediction, converges toward the expected solution. We then show that for a sufficiently large number of MLPs we may obtain a 100% success rate for all learning sets. The solver is obtained by running a set of local solvers each of which is based on a specific MLP whose prediction may lead to an exact solution of the system. This solver is tested on verification sets i.e. set of samples constituted of parameter values (all different from the samples in the learning set) and all the solutions of the corresponding system. We show that these sets may be automatically generated and that they may also be used in a self-learning process for improving the performance of the solver established from the initial solution set. This approach is illustrated on two engineering problems in robotics and chemistry and it is shown that the solver provides all solutions for any instance of P with a high probability although we cannot guarantee it. The time required to design the final solver is large but the solving time is extremely low so that this approach should be used when the system has to be solved for a sufficient number of occurrences of the P. Furthermore we will show that the computation time required for establishing the solver may be drastically reduced by using a distributed implementation. [ABSTRACT FROM AUTHOR]

We propose a test-time adaptation for 6D object pose tracking that learns to adapt a pre-trained model to track the 6D pose of novel objects. We consider the problem of 6D object pose tracking as a 3D keypoint detection and matching task and present a model that extracts 3D keypoints. Given an RGB-D image and the mask of a target object for each frame, the proposed model consists of the self- and cross-attention modules to produce the features that aggregate the information within and across frames, respectively. By using the keypoints detected from the features for each frame, we estimate the pose changes between two frames, which enables 6D pose tracking when the 6D pose of a target object in the initial frame is given. Our model is first trained in a source domain, a category-level tracking dataset where the ground truth 6D pose of the object is available. To deploy this pre-trained model to track novel objects, we present a test-time adaptation strategy that trains the model to adapt to the target novel object by self-supervised learning. Given an RGB-D video sequence of the novel object, the proposed self-supervised losses encourage the model to estimate the 6D pose changes that can keep the photometric and geometric consistency of the object. We validate our method on the NOCS-REAL275 dataset and our collected dataset, and the results show the advantages of tracking novel objects. The collected dataset and visualisation of tracking results are available: https://qm-ipalab.github.io/TA-6DT/ • Test-time adaptation strategy for tracking novel objects. This strategy reduces the dependency on annotated data by introducing self-supervised loss functions based on photometric and geometric consistency. By leveraging RGB-D videos and initial pose information, our method enables accurate and robust 6D pose tracking for objects not encountered during training. • We propose a new model that incorporates transformers, enabling the model to find dependencies and relationships between keypoints. This design facilitates the accurate matching of 3D keypoints, crucial for robust 6D pose tracking. • We evaluate the performance of our method on a set of novel objects that we collected by using a robotic arm. This evaluation allows us to assess the ability of our method to achieve accurate 6D pose tracking for a wide range of novel objects. [ABSTRACT FROM AUTHOR]

Cross-scale measurement techniques are pivotal for seamless coordination between robotic macro and micromanipulation. While computer microvision offers non-contact and multi-degree-of-freedom capabilities, its limitations in measurement range and narrow field-of-view (FOV) prove unsatisfactory for robotic macro-micro manipulation. This study presents a marker-assisted microvision-based method to minimize errors during measurement range extension and ensure precise cross-scale motion measurement for robotic macro-micro manipulation. The microvision-based system utilizes the microstructural pattern designed for global map construction. Additionally, it employs a learning-based approach to efficiently estimate the FOV position for macro-motion measurement, followed by the integration of micro-motion tracking within the FOV for precise micro-motion measurements. Experimental investigations showcase a measurement accuracy of 9 μ m in an extended range of 3.11 mm for macro-motion estimation and 219 nm for micro-motion measurement, which highlight the microvision-based method in achieving effective cross-scale motion measurement, potentially paving the way for the automation of robotic macro-micro manipulation. • A microstructural marker is designed to minimize errors in cross-scale measurement. • Image stitching errors are converted to controllable marker manufacturing errors. • A practical microvision system is developed for robotic macro-micro manipulation. [ABSTRACT FROM AUTHOR]

We evaluated the effect of haptic coordination on anxiety and arousal. Participants looked at a stressful or calming image and then repeatedly squeezed a vibrating stress ball for 20 s. Using a pre-post paradigm with a control group, we showed that squeezing the vibrating ball reduced anxiety and arousal, as assessed by the State-Trait Anxiety Inventory and electrodermal activity, respectively. The stability of haptic coordination was manipulated by varying the detuning between the spontaneous squeezing frequency and the intrinsic frequency of ball vibration. Coordination stability affected arousal and stress affected stability. The data were discussed in the light of Kahneman's attentional resource-sharing model. [Display omitted] • Users were pleased to interact with our new haptic robot. • Squeezing a vibrating robotic ball decreases anxiety and modulates arousal. • The detuning between the frequency of hand squeezing and of ball vibrations affects arousal. • When the detuning is large, the level of stress affects the stability of haptic coordination. [ABSTRACT FROM AUTHOR]

The relationship between artificial entities' human likeness and aesthetic preference is thought to be best modelled by an N -shaped cubic "uncanny valley" function, which however suffers from conceptual criticisms and lack of parsimony. Here it is argued that uncanniness effects may instead be modelled by a linear function of deviation moderated by perceptual specialization. The two models are compared in an experiment with five incrementally distorted face types (cartoon, CG, drawing, real, robot). Recognition performance for upright and inverted faces were used as a specialization measure. Specialization significantly moderated the linear effect of distortion on uncanniness, and could explain the data better than a conventional uncanny valley. The uncanny valley may thus be better understood as a moderated linear function of specialization sensitizing the uncanniness of deviating stimuli. This simpler yet more accurate model is compatible with neurocognitive theories and can explain uncanniness effects beyond the conventional uncanny valley. • Face realism affects both specialization and deviation sensitivity. • Specialization moderates the effect of deviation on uncanniness in faces. • This moderated linear function explains uncanniness better than a cubic "uncanny valley" model. • Parsimony, accuracy, generalizability, and plausibility of the moderation model over the uncanny valley are discussed. [ABSTRACT FROM AUTHOR]

Little is known about how children perceive, trust and learn from social robots compared to humans. The goal of this study was to compare a robot and a human agent in a selective trust task across different combinations of reliability (both reliable, only human reliable, or only robot reliable). 111 children, aged 3 to 6 years, participated in an online study where they viewed videos of a human and a robot labelling both familiar and novel objects. We found that, although children preferred to endorse a novel object label from the agent who previously labelled familiar objects correctly, when both the human and the robot were reliable they were biased more towards the robot. Their social evaluations also tended much more strongly towards a general robot preference. Children's conceptualisations of the agents making a mistake also differed, such that an unreliable human was selected as doing things on purpose, but not an unreliable robot. These findings suggest that children's perceptions of a robot's reliability are separate from their evaluation of its desirability as a social interaction partner and its perceived agency. Further, they indicate that a robot making a mistake does not necessarily reduce children's desire to interact with it as a social agent. • 3–6 year old children prefer a reliable robot over a reliable human. • Children evaluate robot competency separately to social and agency judgements. • Humans are more likely to be seen as making a mistake on purpose than a robot. [ABSTRACT FROM AUTHOR]

The theory of Enactive Inference was proposed by Karl Friston and his colleagues to explain how the brain infers knowledge about the world through the subject's interactive experiences. Sensorimotor states induce perturbations in neural activity, and the brain infers hypothetical causes in the world that may explain these perturbations. This article aims to reconcile this neuroscience theory with computer science and artificial-intelligence theories wherein artificial agents receive input data derived from the environment's state and infer internal data structures used to guide decisions. Two critical challenges arise in both the agent's active role and the inference algorithm's scalability as the environment's complexity increases. To address these challenges, we formalize artificial enactive inference through a new Spatial Enactive Markov Decision Process (SEMDP) model. This model rests on low-level control loops enacted in a three-dimensional Euclidean space containing objects. Based on the SEMDP, we present a proof-of-concept cognitive architecture and an experiment to demonstrate the transcription of the theory of enactive inference into the domain of artificial intelligence and robotics. [ABSTRACT FROM AUTHOR]

• LegUp robot enables the motion of the lower limb in the sagittal and coronal planes. • The mathematical models for legup are representative for the rehabilitation task. • Numerical optimizations lead to a singularity-free operational workspace. The paper presents the mathematical modeling and analysis of LegUp, a novel parallel robotic system for lower limb rehabilitation of bedridden patients. The operational workspace of the robotic system is defined based on a set of parameters that describe the motion of the lower limb joints, which is natural in the rehabilitation task. To comply with this representation of the operational workspace, the parallel robot kinematic models describe the dependency between the robotic system actuators and the lower limb joints. Furthermore, the singularity analysis is achieved in the joint space, which shows whether the operational workspace is singularity-free. To achieve a feasible mechanical design for the prescribed operational workspace and to ensure safe operation, the design parameters of the parallel robotic system are determined based on a multi-objective optimization problem. Numerical simulations show that the operational workspace is singularity-free for the selected design parameters, which are then used to construct the experimental model of LegUp. [ABSTRACT FROM AUTHOR]

The main objectives of this research are (1) to uniquely design assistive behaviors for socially assistive robots using the principles of persuasion from behavioral psychology, and (2) to investigate caregivers' perspectives and opinions on the use of these behaviors to engage and motivate older adults in cognitive activities. We developed 10 unique robot persuasive assistive behavior strategies for the social robot Pepper using both verbal and nonverbal communication modes. Robot verbal behaviors were designed using Cialdini's principles of persuasion; nonverbal behaviors included expansive movements of the body. Care providers' perceptions of the quality, strength, and persuasiveness of these robot persuasive behaviors were assessed based on the Perceived Argument Strength Likert scale. Eighteen formal and informal care providers caring for older adults including those living with mild cognitive impairments participated. An online survey was designed consisting of short videos of the Pepper robot displaying each behavior. After viewing each video, care providers completed the Perceived Argument Strength Likert scale to evaluate 6 attributes for each behavior. They also provided comments. Results show robot assistive behaviors using praise with emotion, along with emotion with commitment were the most positively rated by care providers. Qualitative responses indicate robot body language and speech quality were influencing factors in how a person perceives assistance in human-robot interactions. Our findings provide new insights into incorporating persuasive strategies into the design of assistive social robot behaviors with the aim of engaging and motivating older adults in an activity. The majority of care providers rated the robot persuasive behaviors positively. In designing a persuasive socially assistive robot for older adults, it is beneficial to display a combination of persuasive strategies, such as praise and commitment with emotion, to address individual users' needs and cognitive levels. [ABSTRACT FROM AUTHOR]