Search

This paper proposes a neural network model that estimates the rotation angle of unknown objects from RGB images using an approach inspired by biological neural circuits. The proposed model embeds the understanding of rotational transformations into its architecture, in a way inspired by how rotation is represented in the ellipsoid body of Drosophila. To effectively capture the cyclic nature of rotation, the network’s latent space is structured in a circular manner. The rotation operator acts as a shift in the circular latent space’s units, establishing a direct correspondence between shifts in the latent space and angular rotations of the object in the world space. Our model accurately estimates the difference in rotation between two views of an object, even for categories of objects that it has never seen before. In addition, our model outperforms three state-of-the-art convolutional networks commonly used as the backbone for vision-based-models in robotics. [ABSTRACT FROM AUTHOR]

This paper studies neural network models of vehicle dynamics. We consider both models with a generic layer architecture and models with specialised topologies that hard-wire physics principles. Network pre-wiring is limited to universal laws; hence it does not limit the network modelling abilities on one side but allows more robust and interpretable models on the other side. Four different network types (with and without pre-wired structure, recursive and non-recursive) are compared for the longitudinal dynamics of a car with gears and two controls (brake and engine). Results show that pre-wiring effectively improves the performance. Non-recursive networks also look to be preferable for several reasons. [ABSTRACT FROM AUTHOR]

This paper deals with (deep) neural network implementations of sensorimotor control for automated driving. We show how to construct complex behaviors by re-using elementary neural network building blocks that can be trained and tested extensively; one of our goals is to mitigate the “black box” and verifiability issues that affect end-to-end trained networks. By structuring complex behaviors within a subsumption architecture, we retain the ability to learn (mostly at motor primitives level) with the ability to create complex behaviors by subsuming the (well-known) learned elementary perception-action loops. The learning process itself is simplified, since the agent needs only to learn elementary behaviors. At the same time, the structure imposed with the subsumption architecture ensures that the agent behaves in predictable ways (e.g., treating all obstacles uniformly). We demonstrate these ideas for longitudinal obstacle avoidance behavior, but the proposed approach can also be adapted to other situations. [ABSTRACT FROM AUTHOR]

This paper studies a control architecture for vehicle lateral dynamics based on the execution of optimal trajectories via feedforward inverse model control. The focus here is on assessing the robustness of this arrangement when the vehicle real forward dynamics is not identically cancelled by the inverse model due to model approximations and parameter uncertainties. The trajectories that are considered are analytic solutions of the minimum square jerk optimal control problem for a simplified kinematic vehicle model in curvilinear coordinates. Various hypotheses are made concerning the mismatch between the inverse model and the actual forward dynamics of the vehicle. Closed-loop stability analysis shows that the studied control scheme guarantees asymptotic stability of both the nominal and the perturbed kinematic model with significant robustness margins. In addition to the theoretical robustness analysis, the same control scheme is validated in simulation using the industry-standard software for virtual vehicle testing IPG CarMaker. [ABSTRACT FROM AUTHOR]

This paper presents a principled approach to the modeling of human drivers—applied to stop behavior—by uniting recent ideas in cognitive science and optimal control. With respect to the former, we invoke the affordance competition hypothesis, according to which human behavior is produced by resolving the competition between action affordances that are simultaneously instantiated in response to the environment. From the theory of optimal control, we deploy motor primitives based on minimum jerk as the potential suite of actions. Furthermore, we invoke a layered control architecture, which carries out action priming and action selection sequentially, to model the biological affordance competition process. Motor output may be directed to distinct motor channels, which may be partially inhibited, e.g., to model gas pedal release saturation. Within this architecture, two types of motor units—“deceleration” acting on a gas pedal channel and “brake” acting on a brake pedal channel—are sufficient to model, with remarkable accuracy, the various phases that can be observed in human maneuvers in stopping a car, namely: gas release, gas chocked, brake, and final brake release at stop. The model is validated using experimental data collected in 16 different stop locations, from roundabouts to traffic lights. We also carry out a comparison with the well-known Intelligent Driver Model, discuss the scaling of this framework to more general driving scenarios and finally give an example application where the driver model is used, within a mirroring process, to infer the human driver intentions. [ABSTRACT FROM AUTHOR]

This paper presents a cooperative intersection support system implemented with an artificial cognitive system enacted by an agent that replicates human driver longitudinal sensorimotor control in the application domain. The engineering of the agent exploits recent ideas from Cognitive Science, among which the notion of mirroring (the agent discovering driver intentions by simulating possible human actions). The paper introduces the design principles for the agent and the following implementation. The system is evaluated with experimental data and compared to state of the art implementations that use different approaches. [ABSTRACT FROM PUBLISHER]

This paper aims at estimating the longitudinal jerk of the vehicle as it is acted by a human driver, in the context of preventive safety. A reliable estimate is needed to infer the current driver intention in an advanced driving assistance system developed by the authors. The derived intention-oriented model for the longitudinal dynamics is embedded into an enhanced Kalman filter that provides the user with a knob to trade off between responsiveness of the estimate and noise rejection. The scheme is fit for on-line usage, relies on signals commonly available on the CAN bus of modern vehicles, and requires a very limited number of parameters. Its effectiveness is validated on experimental data and compared with alternative approaches. [ABSTRACT FROM PUBLISHER]

The study of action selection in humans can present challenges of task design since our actions are usually defined by many degrees of freedom and therefore occupy a large action-space. While saccadic eye-movement offers a more constrained paradigm for investigating action selection, the study of reach-and-grasp in upper limbs has often been defined by more complex scenarios, not easily interpretable in terms of such selection. Here we present a novel motor behaviour task which addresses this by limiting the action space to a single degree of freedom in which subjects have to track (using a stylus) a vertical coloured target line displayed on a tablet computer, whilst ignoring a similarly oriented distractor line in a different colour. We ran this task with 55 subjects and showed that, in agreement with previous studies, the presence of the distractor generally increases the movement latency and directional error rate. Further, we used two distractor conditions according to whether the distractor’s location changes asynchronously or synchronously with the location of the target. We found that the asynchronous distractor yielded poorer performance than its synchronous counterpart, with significantly higher movement latencies and higher error rates. We interpret these results in an action selection framework with two actions (move left or right) and competing ‘action requests’ offered by the target and distractor. As such, the results provide insights into action selection performance in humans and supply data for directly constraining future computational models therein. [ABSTRACT FROM AUTHOR]

This paper describes a curve negotiation “behavior” that can be used—within subsumption architectures—to produce artificial agents with the ability to negotiate curves in a humanlike way. This may be used to implement functions spanning different levels of automation, from assistance (curve warning) to automated (curve speed control). This paper gives the following: 1) a summary of related works and of the subsumption architecture conceptual framework; 2) a detailed description of the function within this framework; 3) experimental data for validation and tuning derived from user tests; 4) guidelines on integration of the function within advanced driver assistance systems with different automation levels, with examples; and 5) a comparison with experimental data of the human curve speed choice models in the state of the art. [ABSTRACT FROM PUBLISHER]

This position paper introduces the concept of artificial “co-drivers” as an enabling technology for future intelligent transportation systems. In Sections I and II, the design principles of co-drivers are introduced and framed within general human–robot interactions. Several contributing theories and technologies are reviewed, specifically those relating to relevant cognitive architectures, human-like sensory-motor strategies, and the emulation theory of cognition. In Sections III and IV, we present the co-driver developed for the EU project interactIVe as an example instantiation of this notion, demonstrating how it conforms to the given guidelines. We also present substantive experimental results and clarify the limitations and performance of the current implementation. In Sections IV and V, we analyze the impact of the co-driver technology. In particular, we identify a range of application fields, showing how it constitutes a universal enabling technology for both smart vehicles and cooperative systems, and naturally sets out a program for future research. [ABSTRACT FROM PUBLISHER]

In this work, we present the results of collaboration between an engineering department and a rehabilitation hospital in using innovative touch interfaces properly designed for both neurocognitive and physical rehabilitation. The novel touch interface also measures force, thereby enabling dexterity training through 'direct' manipulation of virtual objects in 3D. Two dimensions are recorded via touch screen, the third by the force channel. We believe that this tool could increase the degree of effectiveness of traditional rehabilitation treatments thanks to its capability to merge physical and cognitive rehabilitation. Furthermore, the exergames implemented allow an easy personalization of the exercise structure and difficulty level. The effectiveness of the FP technology compared with more traditional methods of rehabilitation is measured according to specific parameters observed in an experimental group in comparison with a control group. [ABSTRACT FROM AUTHOR]

Going beyond standard lane-departure-avoidance systems, this paper addresses the development of a system that is able to deal with a large set of different traffic situations. Its foundation lies on a thoroughly constituted environment detection through which a decision system is built. From the output of the decision module, the driver is warned or corrected through suited actuators that are coupled to control strategies. The input to the system comes from cameras, which are supplemented by active sensors (such as radar and laser scanners) and vehicle dynamic data, digital road maps, and precise vehicle-positioning data. In this paper, the presented system design is divided into three layers: the perception layer, which is responsible for the environment perception, and the decision and action layers, which are responsible for evaluating and executing actions, respectively. [ABSTRACT FROM PUBLISHER]

This paper deals with the integration of multiple advanced driver-assistance systems (ADAS) and in-vehicle information systems (IVIS) in a holistic driver-support system. The paper presents the results of a project named Integrated Safety Systems (INSAFES), which was part of PReVENT: an integrating project carried out under the European Framework Programme 6. Integration in INSAFES is tackled at three different levels in the framework of a “cognitive car” perspective: 1) at the perception level, to represent the world around the vehicle, including object-tracking between sensor fields and the detection of driver intentions; 2) at the decision level, to reproduce humanlike holistic motion plans, which serve as “reference maneuvers” to evaluate the motion alternatives that a driver faces; and 3) at the level of interaction with the driver and vehicle control ( action level), to arbitrate between the requests of functions competing for driver attention. A function that provides simultaneous longitudinal and lateral support has been developed. It gives support for safe speed, safe distance, lane change, and all-around collision avoidance all at the same time. At its core, there is a tool (evasive/reference maneuver) that constantly evaluates two possible alternatives (in lane and evasive/lane change) and compares them with the driver input to detect which one applies, which dictates warnings and driver interactions, and whether there is a better alternative. In addition, a “warning manager” has been developed, acting like a referee who lets the ADAS applications work standalone and then combines the requests of each application, prioritizes them, and manages the interaction with the user. The warning manager can be particularly useful in the case of integration of pre-existing standalone functions, which can be quickly reused. If a holistic ADAS is developed, the warning manager can still be used to combine it with IVIS functions. In fact, depending on the kind of ADAS and IVIS considered, the most suitable approach can be either to combine functions in a unified multifunctional driver-support application or to arbitrate between them through the warning manager. [ABSTRACT FROM PUBLISHER]

This paper describes a novel driver-support system that helps to maintain the correct speed and headway (distance) with respect to lane curvature and other vehicles ahead. The system has been developed as part of the Integrating Project PReVENT under the European Framework Programme 6, which is named SAfe SPEed and safe distaNCE (SASPENCE). The application uses a detailed description of the situation ahead of the vehicle. Many sensors [radar, video camera, Global Positioning System (GPS) and accelerometers, digital maps, and vehicle-to-vehicle wireless local area network (WLAN) connections] are used, and state-of-the-art data fusion provides a model of the environment. The system then computes a feasible maneuver and compares it with the driver's behavior to detect possible mistakes. The warning strategies are based on this comparison. The system “talks” to the driver mainly via a haptic pedal or seat belt and “listens” to the driver mainly via the vehicle acceleration. This kind of operation, i.e., the comparison between what the system thinks is possible and what the driver appears to be doing, and the consequent dialog can be regarded as simple implementations of the rider–horse metaphor (H-metaphor). The system has been tested in several situations (driving simulator, hardware in the loop, and real road tests). Objective and subjective data have been collected, revealing good acceptance and effectiveness, particularly in awakening distracted drivers. The system intervenes only when a problem is actually detected in the headway and/or speed (approaching curves or objects) and has been shown to cause prompt reactions and significant speed correction before getting into really dangerous situations. [ABSTRACT FROM PUBLISHER]

This paper focuses on the collaboration between human drivers and intelligent vehicles. We propose a collaboration mechanism grounded on the concept of distributed cognition. With distributed cognition, intelligence does not lie just in the single entity but also in the interaction with the other cognitive components in a system. We apply this idea to vehicle intelligence, proposing a system distributed into two cognitive entities-the human and the autonomous agent-that together contribute to drive the vehicle. This account of vehicle intelligence differs from the mainstream research effort on highly autonomous cars. The proposed mechanism follows one of the paradigm derived from distributed cognition, the rider-horse metaphor: just like the rider communicates their intention to the horse through the reins, the human influences the agent using the pedals and the steering wheel. We use a driving simulator to demonstrate the collaboration in action, showing how the human can communicate and interact with the agent in various ways with safe outcomes., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2022 Plebe, Rosati Papini, Cherubini and Da Lio.)

Beta power suppression in the basal ganglia is stronger during movements that require high force levels and high movement effort but it has been difficult to dissociate the two. We recorded scalp EEG and basal ganglia local field potentials in Parkinson's disease patients (11 STN, 7 GPi) ON and OFF dopaminergic medication while they performed a visually-guided force matching task using a pen on a force-sensitive graphics tablet. Force adjustments were accompanied by beta power suppression irrespective of whether the force was increased or reduced. Before the adjustment was completed, beta activity returned. High beta power was specifically associated with slowing of the force adjustment. ON medication, the peak force rate was faster and cortico-basal ganglia beta phase coupling was more readily modulated. In particular, phase decoupling was stronger during faster adjustments. The results suggest that beta power in the basal ganglia does not covary with force per se, but rather with a related factor, the absolute force rate, or a more general concept of movement effort. The results also highlight that beta activity reappears during stabilization of isometric contractions, and that dopamine-related suppression of cortico-basal ganglia beta coupling is linked to faster force adjustments., (Copyright © 2019. Published by Elsevier Inc.)

Computerized interfaces are able to represent 3D immersive simulations. Most of them make use of joystick, mouse, gloves, or grasp pressure transducers. Those have the drawback of 'filtering' the user interaction and/or de-locate the touch with respect to the visual stimulus. To overcome this we developed dexterity rehabilitation games on a novel touch interface that measures also force. The system allows dexterity training through 'direct' manipulation of virtual objects in 3D. Two dimensions via the touch screen, the third by the force channel. Tactile feedback is provided with a vibration device mounted on the screen back.