Search

The process-aware source seeking (PASS) problem in flow fields aims to find an informative trajectory to reach an unknown source location while taking the energy consumption in the flow fields into consideration. Taking advantage of the dynamic flow field partition technique, this paper formulates this problem as a task and motion planning (TAMP) problem and proposes a bi-level hierarchical planning framework to decouple the planning of inter-region transition and inner-region trajectory by introducing inter-region junctions. An integrated strategy is developed to enable efficient upper-level planning by investigating the optimal solution of the lower-level planner. In order to leverage the information acquisition and computational burden, a dynamic event-triggered mechanism is introduced to enable asynchronized estimation, region partitioning and re-plans. The proposed algorithm provides guaranteed convergence of the trajectory, and achieves automatic trade-offs of both exploration-exploitation and accuracy-efficiency. Simulations in a highly complicated and realistic ocean surface flow field validate the merits of the proposed algorithm, which demonstrates a significant reduction in computational burden without compromising planning optimality. [ABSTRACT FROM AUTHOR]

This paper explores the problem of automated robot program generation from limited historical data when neither accurate geometric environmental models nor online vision feedback are available. The Program Reuse Framework (PRF) is developed, which uses expert-defined motion classes, a novel data structure introduced in this work, to learn affordances, workspaces, and skills from historical data. Historical data comprise raw robot joint trajectories and descriptions of the robot task being completed. Given new tasks, motion classes are then used again to formulate an optimization problem capable of generating new open-loop, skill-based programs to complete the tasks. To cope with a lack of geometric models, a technique to learn safe workspaces from demonstrations is developed, allowing the risk of new programs to be estimated before execution. A new learnable motion primitive for redundant manipulators is introduced, called a redundancy dynamical movement primitive, which enables new end-effector goals to be reached while mimicking the whole-arm behavior of a demonstration. A mobile manipulator part transportation task is used throughout to illustrate each step of the framework. [ABSTRACT FROM AUTHOR]

This study focuses on addressing the problem of motion planning within workspaces cluttered with obstacles while considering temporal and input constraints. These specifications can encapsulate intricate high-level objectives involving both temporal and spatial constraints. The existing literature lacks the ability to fulfill time specifications while simultaneously managing input-saturation constraints. The proposed approach introduces a hybrid three-component control algorithm designed to learn the safe execution of a high-level specification expressed as a timed temporal logic formula across predefined regions of interest in the workspace. The first component encompasses a motion controller enabling secure navigation within the minimum allowable time interval dictated by input constraints, facilitating the abstraction of the robot's motion as a timed transition system between regions of interest. The second component utilizes formal verification and convex optimization techniques to derive an optimal high-level timed plan over the mentioned transition system, ensuring adherence to the agent's specification. However, the necessary navigation times and associated costs among regions are initially unknown. Consequently, the algorithm's third component iteratively adjusts the transition system and computes new plans as the agent navigates, acquiring updated information about required time intervals and associated navigation costs. The effectiveness of the proposed scheme is demonstrated through both simulation and experimental studies. [ABSTRACT FROM AUTHOR]

This paper presents an autonomous multi-agent drone system utilized for an artificial pollination task to address the declining bee population and its impact on natural pollination. The main goal is to visit all the targets (flowers) by drones. The system receives the position of the targets located on a vertical fruiting wall from a single sensor mounted on an unmanned ground vehicle (UGV). Then, it determines the sequence of target allocation to minimize the task time and to minimize the accumulated distance each drone travels. The limited payload capacity of each drone and the need to refill stock carried by the UGV are also considered. After allocating the targets, a safe trajectory is generated for each drone to avoid collision with other drones, considering uncertainty in the drones’ positions. When a drone reaches a target, the system allocates a new target, and the cycle repeats until the task is completed. The system is reactive to the evolution of the task and adjusts accordingly. An additional tested factor is the impact of the positioning error on the task time. The system was evaluated through dynamic simulations using real datasets of peach and pear orchards and finally in a laboratory proof-of-concept experiment using a micro-drone system.

Autonomous robots are increasingly served in real-world unstructured human environments with complex long-horizon tasks, such as restaurant serving and office delivery. Task and motion planning (TAMP) is a recent research method in Artificial Intelligence Planning for these applications. TAMP integrates high-level abstract reasoning with the low-level geometric feasibility check and thus is more comprehensive than traditional task planning methods. While regular TAMP approaches are challenged by different types of uncertainties and the generalization of various applications when implemented in real-world scenarios. This article systematically reviews the most relevant approaches to TAMP and classifies them according to their features and emphasis; it categorizes the challenges and presents online TAMP and machine learning-based TAMP approaches for addressing them. [ABSTRACT FROM AUTHOR]

Many tasks require robots to manipulate objects while satisfying a complex interplay of spatial and temporal constraints. For instance, a table setting robot first needs to place a mug and then fill it with coffee, while satisfying spatial relations such as forks need to placed left of plates. We propose the spatio-temporal framework SpaTiaL that unifies the specification, monitoring, and planning of object-oriented robotic tasks in a robot-agnostic fashion. SpaTiaL is able to specify diverse spatial relations between objects and temporal task patterns. Our experiments with recorded data, simulations, and real robots demonstrate how SpaTiaL provides real-time monitoring and facilitates online planning. SpaTiaL is open source and easily expandable to new object relations and robotic applications. [ABSTRACT FROM AUTHOR]

This paper explores the problem of automated robot program generation from limited historical data when neither accurate geometric environmental models nor online vision feedback are available. The Program Reuse Framework (PRF) is developed, which uses expert-defined motion classes, a novel data structure introduced in this work, to learn affordances, workspaces, and skills from historical data. Historical data comprise raw robot joint trajectories and descriptions of the robot task being completed. Given new tasks, motion classes are then used again to formulate an optimization problem capable of generating new open-loop, skill-based programs to complete the tasks. To cope with a lack of geometric models, a technique to learn safe workspaces from demonstrations is developed, allowing the risk of new programs to be estimated before execution. A new learnable motion primitive for redundant manipulators is introduced, called a redundancy dynamical movement primitive, which enables new end-effector goals to be reached while mimicking the whole-arm behavior of a demonstration. A mobile manipulator part transportation task is used throughout to illustrate each step of the framework.

The current paper proposes a hierarchical reinforcement learning (HRL) method to decompose a complex task into simpler sub-tasks and leverage those to improve the training of an autonomous agent in a simulated environment. For practical reasons (i.e., illustrating purposes, easy implementation, user-friendly interface, and useful functionalities), we employ two Python frameworks called TextWorld and MiniGrid. MiniGrid functions as a 2D simulated representation of the real environment, while TextWorld functions as a high-level abstraction of this simulated environment. Training on this abstraction disentangles manipulation from navigation actions and allows us to design a dense reward function instead of a sparse reward function for the lower-level environment, which, as we show, improves the performance of training. Formal methods are utilized throughout the paper to establish that our algorithm is not prevented from deriving solutions.

This work addresses task planning under uncertainty for precision agricultureapplications whereby task costs are uncertain and the gain of completing a taskis proportional to resource consumption (such as water consumption in precisionirrigation). The goal is to complete all tasks while prioritizing those thatare more urgent, and subject to diverse budget thresholds and stochastic costsfor tasks. To describe agriculture-related environments that incorporatestochastic costs to complete tasks, a new Stochastic-Vertex-Cost Aisle Graph(SAG) is introduced. Then, a task allocation algorithm, termed Next-Best-ActionPlanning (NBA-P), is proposed. NBA-P utilizes the underlying structure enabledby SAG, and tackles the task planning problem by simultaneously determining theoptimal tasks to perform and an optimal time to exit (i.e. return to a basestation), at run-time. The proposed approach is tested with both simulated dataand real-world experimental datasets collected in a commercial vineyard, inboth single- and multi-robot scenarios. In all cases, NBA-P outperforms otherevaluated methods in terms of return per visited vertex, wasted resourcesresulting from aborted tasks (i.e. when a budget threshold is exceeded), andtotal visited vertices.

This study focuses on addressing the problem of motion planning within workspaces cluttered with obstacles while considering temporal and input constraints. These specifications can encapsulate intricate high-level objectives involving both temporal and spatial constraints. The existing literature lacks the ability to fulfill time specifications while simultaneously managing input-saturation constraints. The proposed approach introduces a hybrid three-component control algorithm designed to learn the safe execution of a high-level specification expressed as a timed temporal logic formula across predefined regions of interest in the workspace. The first component encompasses a motion controller enabling secure navigation within the minimum allowable time interval dictated by input constraints, facilitating the abstraction of the robot’s motion as a timed transition system between regions of interest. The second component utilizes formal verification and convex optimization techniques to derive an optimal high-level timed plan over the mentioned transition system, ensuring adherence to the agent’s specification. However, the necessary navigation times and associated costs among regions are initially unknown. Consequently, the algorithm’s third component iteratively adjusts the transition system and computes new plans as the agent navigates, acquiring updated information about required time intervals and associated navigation costs. The effectiveness of the proposed scheme is demonstrated through both simulation and experimental studies.

This paper addresses the problem of task assignment and trajectory generation for installing bird diverters using a fleet of multi-rotors. The proposed solution extends our previous motion planner to compute feasible and constrained trajectories, considering payload capacity limitations and recharging constraints. Signal Temporal Logic (STL) specifications are employed to encode the mission objectives and temporal requirements. Additionally, an event-based replanning strategy is introduced to handle unforeseen failures. An energy minimization term is also employed to implicitly save multi-rotor flight time during installation operations. The effectiveness and validity of the approach are demonstrated through simulations in MATLAB and Gazebo, as well as field experiments carried out in a mock-up scenario.

We propose two algorithms based on local and global searches for a Task and Motion Planning (TAMP) problem, which considers a robotic manipulator to rearrange obstacles and grasp a target in clutter. In the problem, no collision-free path for a robotic manipulator is available unless some obstacles blocking a target are relocated. The two algorithms determine the sequence of obstacles to be removed for grasping a target without collisions. The local search algorithm determines the sequence quickly in an online manner but could be suboptimal in the number of removed obstacles. The global search algorithm based on tree search needs upfront computation but runs in polynomial time. In numerical simulation settings, we consider objects in various shapes, which could make reachable directions bounded. From the simulations the planning time of local search algorithm is faster than that of global search, whereas the global search algorithm removes the less number of obstacles than the local search. In addition, we show that the global search algorithm with a heuristic cost is faster than without the cost but the minimum number of removed obstacles is still obtained from the global search algorithm without the heuristic cost. Practical experiments show the applicability of our algorithms in real environments

Robots tasked with object assembly by manipulation of parts require not only a high-level plan for order of placement of parts but also detailed low-level information on how to place and pick the part based on its state. This is a complex multi-level problem prone to failures at various levels. This paper employs meta reasoning architecture along with robotics principles and proposes dual encoding of state expectations during the progression of task to ground nominal scenarios. We present our results on table-top scenario using perceptual expectations based in the concept of occupancy grids and key point representations. Our results in a constrained manipulation setting suggest using low-level information or high-level expectations alone the system performs worse than if the architecture uses them both. We then outline a complete architecture and system which tackles this problem for repairing more generic assembly plans with objects moving in spaces with 6 degrees of freedom.

We propose a Task and Motion Planning algorithm for a robot manipulator to rearrange objects in a narrow and highly-dense workspace. In such a workspace, there may not be enough space to rearrange the objects that obstructs the way for a robot manipulator to access a target object. The present work aims to relocate as few objects (so called obstacles) as possible for the robot manipulator to grasp the target object successfully. The proposed algorithm seeks both the sequence of obstacles to be rearranged and the corresponding places to move them. A heuristic search is employed to find a rearrangement sequence and places. Especially, in the proposed algorithm, stacking an object on the top of other objects is allowed, while searching a place to rearrange obstacles. Virtual simulations and real robot experiments show that the proposed algorithm works reasonably for dense real-world environments. [ABSTRACT FROM AUTHOR]

A service robot needs integration of symbolic reasoning for high-level task planning and geometric computation for low-level motion planning to provide services in an everyday human-living environment. For this purpose, one may develop individual modules for object recognition, knowledge inference, task planning, and motion planning, which requires a system that integrates them to provide services autonomously. In this paper, we propose a combined task-motion planning system implemented using existing open-source libraries. We implemented an action library module that specifies the relationship between the compound actions that are modeled in an AI planning language to enable abstract reasoning and the primitive actions that can be geometrically verified its feasibility. This serves as an interface between the two levels by providing a rule in which a compound action sequence obtained from task planning at the symbolic level is decomposed into a primitive action sequence capable of motion planning at the geometric level. In addition, we defined the relationship between the two types of actions in a hierarchical structure and added conditional clauses according to the task states, so that primitive actions that need to be additionally performed are automatically added to the action sequence during decomposition. This procedure enables the robot to successfully perform the task in response to an unintended change in the environment. We establish a task domain where a robot delivers an object to a human user in unexpected situations and verify the proposed method under dynamic simulation environments. [ABSTRACT FROM AUTHOR]

Although collective robotic construction systems are beginning to showcase how multi‐robot systems can contribute to building construction by efficiently building low‐cost, sustainable structures, the majority of research utilizes non‐structural or highly customized materials. A modular collective robotic construction system based on a robotic actuator, which leverages timber struts for the assembly of architectural artifacts as well as part of the robot body for locomotion is presented. The system is co‐designed for in‐plane assembly from an architectural, robotic, and computer science perspective in order to integrate the various hardware and software constraints into a single workflow. The system is tested using five representative physical scenarios. These proof‐of‐concept demonstrations showcase three tasks required for construction assembly: the ability of the system to locomote, dynamically change the topology of connecting robotic actuators and timber struts, and collaborate to transport timber struts. As such, the groundwork for a future autonomous collective robotic construction system that could address collective construction assembly and even further increase the flexibility of on‐site construction robots through its modularity is laid. [ABSTRACT FROM AUTHOR]

Contact-based decision and planning methods are becoming increasingly important to endow higher levels of autonomy for legged robots. Formal synthesis methods derived from symbolic systems have great potential for reasoning about high-level locomotion decisions and achieving complex maneuvering behaviors with correctness guarantees. This study takes a first step toward formally devising an architecture composed of task planning and control of whole-body dynamic locomotion behaviors in constrained and dynamically changing environments. At the high level, we formulate a two-player temporal logic game between the multi-limb locomotion planner and its dynamic environment to synthesize a winning strategy that delivers symbolic locomotion actions. These locomotion actions satisfy the desired high-level task specifications expressed in a fragment of temporal logic. Those actions are sent to a robust finite transition system that synthesizes a locomotion controller that fulfills state reachability constraints. This controller is further executed via a low-level motion planner that generates feasible locomotion trajectories. We construct a set of dynamic locomotion models for legged robots to serve as a template library for handling diverse environmental events. We devise a replanning strategy that takes into consideration sudden environmental changes or large state disturbances to increase the robustness of the resulting locomotion behaviors. We formally prove the correctness of the layered locomotion framework guaranteeing a robust implementation by the motion planning layer. Simulations of reactive locomotion behaviors in diverse environments indicate that our framework has the potential to serve as a theoretical foundation for intelligent locomotion behaviors. [ABSTRACT FROM AUTHOR]

Abstract Although collective robotic construction systems are beginning to showcase how multi‐robot systems can contribute to building construction by efficiently building low‐cost, sustainable structures, the majority of research utilizes non‐structural or highly customized materials. A modular collective robotic construction system based on a robotic actuator, which leverages timber struts for the assembly of architectural artifacts as well as part of the robot body for locomotion is presented. The system is co‐designed for in‐plane assembly from an architectural, robotic, and computer science perspective in order to integrate the various hardware and software constraints into a single workflow. The system is tested using five representative physical scenarios. These proof‐of‐concept demonstrations showcase three tasks required for construction assembly: the ability of the system to locomote, dynamically change the topology of connecting robotic actuators and timber struts, and collaborate to transport timber struts. As such, the groundwork for a future autonomous collective robotic construction system that could address collective construction assembly and even further increase the flexibility of on‐site construction robots through its modularity is laid.

This study proposes a Task and Motion Planning (TAMP) method with symbolic decisions embedded in a bilevel optimization. This TAMP method exploits the discrete structure of sequential manipulation for long-horizon and versatile tasks in dynamic environments. At the symbolic planning level, we propose a scalable decision-making method for long-horizon manipulation tasks using the Planning Domain Definition Language (PDDL) with causal graph decomposition. At the motion planning level, we devise a trajectory optimization (TO) approach based on the Differential Dynamic Programming (DDP) and Alternating Direction Method of Multipliers (ADMM), suitable for solving constrained, large-scale nonlinear optimization in a distributed manner. Distinct from conventional geometric motion planners, our approach generates highly dynamic manipulation motions by incorporating the full robot and object dynamics. Furthermore, in lieu of a hierarchical planning approach, we solve a holistically integrated bilevel optimization problem involving costs from both the low-level TO and the high-level search. Simulation and experimental results demonstrate dynamic manipulation for long-horizon object sorting tasks in clutter and on a moving conveyor belt.

To achieve the tasks provided by a user, it is necessary for robots to have a plan that fully exploits their functionalities in an environment. The objective of this study is to realize robot task planning in real space for effectively use of the robot’s functions The plan is formed by deriving a feasible action sequence by interpreting the instructions within the scope of the action possibilities of the robots and the changes in them. In this paper, we first propose an action graph as a novel environmental representation approach to facilitate the understanding of the robot’s action possibility in real space. In the action graph, the action possibility is represented by nodes, which describe the spatial position to perform each feasible action, and edges, which describe the feasible actions, based on the subsystem-level affordance and the arrangement of objects in the environment. We also propose an action-based spatio-temporal robot navigation (ASTRON), which focuses on robot navigation tasks. ASTRON enables the robots to determine a feasible action sequence that utilizes their functions by interpreting the instructions based on the action graph. The effectiveness of the proposed method was evaluated through simulations and actual machine experiments in a coffee shop environment. In the actual machine experiments, the proposed method was applied to robots with different subsystem configurations. The experimental results demonstrated that the proposed method could plan the feasible action sequence to complete the tasks by considering the environmental state and the subsystem configurations of the robot.

In this article, we propose deep visual reasoning, which is a convolutional recurrent neural network that predicts discrete action sequences from an initial scene image for sequential manipulation problems that arise, for example, in task and motion planning (TAMP). Typical TAMP problems are formalized by combining reasoning on a symbolic, discrete level (e.g., first-order logic) with continuous motion planning such as nonlinear trajectory optimization. The action sequences represent the discrete decisions on a symbolic level, which, in turn, parameterize a nonlinear trajectory optimization problem. Owing to the great combinatorial complexity of possible discrete action sequences, a large number of optimization/motion planning problems have to be solved to find a solution, which limits the scalability of these approaches. To circumvent this combinatorial complexity, we introduce deep visual reasoning: based on a segmented initial image of the scene, a neural network directly predicts promising discrete action sequences such that ideally only one motion planning problem has to be solved to find a solution to the overall TAMP problem. Our method generalizes to scenes with many and varying numbers of objects, although being trained on only two objects at a time. This is possible by encoding the objects of the scene and the goal in (segmented) images as input to the neural network, instead of a fixed feature vector. We show that the framework can not only handle kinematic problems such as pick-and-place (as typical in TAMP), but also tool-use scenarios for planar pushing under quasi-static dynamic models. Here, the image-based representation enables generalization to other shapes than during training. Results show runtime improvements of several orders of magnitudes by, in many cases, removing the need to search over the discrete action sequences. [ABSTRACT FROM AUTHOR]

The objective of this work is to augment the basic abilities of a robot by learning to use sensorimotor primitives to solve complex long-horizon manipulation problems. This requires flexible generative planning that can combine primitive abilities in novel combinations and, thus, generalize across a wide variety of problems. In order to plan with primitive actions, we must have models of the actions: under what circumstances will executing this primitive successfully achieve some particular effect in the world? We use, and develop novel improvements to, state-of-the-art methods for active learning and sampling. We use Gaussian process methods for learning the constraints on skill effectiveness from small numbers of expensive-to-collect training examples. In addition, we develop efficient adaptive sampling methods for generating a comprehensive and diverse sequence of continuous candidate control parameter values (such as pouring waypoints for a cup) during planning. These values become end-effector goals for traditional motion planners that then solve for a full robot motion that performs the skill. By using learning and planning methods in conjunction, we take advantage of the strengths of each and plan for a wide variety of complex dynamic manipulation tasks. We demonstrate our approach in an integrated system, combining traditional robotics primitives with our newly learned models using an efficient robot task and motion planner. We evaluate our approach both in simulation and in the real world through measuring the quality of the selected primitive actions. Finally, we apply our integrated system to a variety of long-horizon simulated and real-world manipulation problems. [ABSTRACT FROM AUTHOR]

Combined Task and Motion Planning (TAMP) is an area where no one-fits-all solution can exist. Many aspects of the domain, as well as operational requirements, have an effect on how algorithms and representations are designed. Frequently, trade-offs have to be made to build a system that is effective. We propose five research questions that we believe need to be answered to solve real-world problems that involve combined TAMP. We show which decisions and trade-offs should be made with respect to these research questions, and illustrate these on examples of existing application domains. By doing so, this article aims to provide a guideline for designing combined TAMP solutions that are adequate and effective in the target scenario.

Manipulation planning under incomplete information is a highly challenging task for mobile manipulators. Uncertainty can be resolved by robot perception modules or using human knowledge in the execution process. Human operators can also collaborate with robots for the execution of some difficult actions or as helpers in sharing the task knowledge. In this scope, a contingent-based task and motion planning is proposed taking into account robot uncertainty and human–robot interactions, resulting a tree-shaped set of geometrically feasible plans. Different sorts of geometric reasoning processes are embedded inside the planner to cope with task constraints like detecting occluding objects when a robot needs to grasp an object. The proposal has been evaluated with different challenging scenarios in simulation and a real environment. [ABSTRACT FROM AUTHOR]

The current paper proposes a hierarchical reinforcement learning (HRL) method to decompose a complex task into simpler sub-tasks and leverage those to improve the training of an autonomous agent in a simulated environment. For practical reasons (i.e., illustrating purposes, easy implementation, user-friendly interface, and useful functionalities), we employ two Python frameworks called TextWorld and MiniGrid. MiniGrid functions as a 2D simulated representation of the real environment, while TextWorld functions as a high-level abstraction of this simulated environment. Training on this abstraction disentangles manipulation from navigation actions and allows us to design a dense reward function instead of a sparse reward function for the lower-level environment, which, as we show, improves the performance of training. Formal methods are utilized throughout the paper to establish that our algorithm is not prevented from deriving solutions., 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. The author(s) declared that they were an editorial board member of Frontiers, at the time of submission. This had no impact on the peer review process and the final decision., (Copyright © 2023 Petsanis, Keroglou, Kapoutsis, Kosmatopoulos and Sirakoulis.)

In this paper, we propose a learning algorithm that speeds up the search in task and motion planning problems. Our algorithm proposes solutions to three different challenges that arise in learning to improve planning efficiency: what to predict, how to represent a planning problem instance, and how to transfer knowledge from one problem instance to another. We propose a method that predicts constraints on the search space based on a generic representation of a planning problem instance, called score-space, where we represent a problem instance in terms of the performance of a set of solutions attempted so far. Using this representation, we transfer knowledge, in the form of constraints, from previous problems based on the similarity in score-space. We design a sequential algorithm that efficiently predicts these constraints, and evaluate it in three different challenging task and motion planning problems. Results indicate that our approach performs orders of magnitudes faster than an unguided planner. [ABSTRACT FROM AUTHOR]

Robotic systems composed of several mobile robots moving in human environments pose several problems at perception, planning and control levels. In these environments, there may be obstacles obstructing the paths, which robots can remove by pushing or pulling them. At planning level, therefore, an efficient combination of task and motion planning is required. Even more if we assume a cooperative system in which robots can collaborate with each other by e.g. pushing together a heavy obstacle or by one robot clearing the way to another one. In this paper, we cope with this problem by proposing -TMP, a smart combination of an heuristic task planner based on the Fast Forward method, a physics-based motion planner, and reasoning processes over the ontologies that code the knowledge on the problem. The significance of the proposal relies on how geometric and physics information is used within the computation of the heuristics in order to guide the symbolic search, i.e. how an artificial intelligence planning method is combined with low-level motion planning to achieve a feasible sequence of actions (composed of collision-free motions plus physically-feasible push/pull actions). The proposal has been validated with several simulated scenarios (using up to five robots that need to collaborate with each other to reach the goal state), showing how the method is able to solve challenging situations and also find an efficient solution in terms of power. [ABSTRACT FROM AUTHOR]

Manipulation planning under incomplete information is a highly challenging task for mobile manipulators. Uncertainty can be resolved by robot perception modules or using human knowledge in the execution process. Human operators can also collaborate with robots for the execution of some difficult actions or as helpers in sharing the task knowledge. In this scope, a contingent-based task and motion planning is proposed taking into account robot uncertainty and human−robot interactions, resulting a tree-shaped set of geometrically feasible plans. Different sorts of geometric reasoning processes are embedded inside the planner to cope with task constraints like detecting occluding objects when a robot needs to grasp an object. The proposal has been evaluated with different challenging scenarios in simulation and a real environment.

We present a new constraint-based framework for task and motion planning (TMP). Our approach is extensible, probabilistically complete, and offers improved performance and generality compared with a similar, state-of-the-art planner. The key idea is to leverage incremental constraint solving to efficiently incorporate geometric information at the task level. Using motion feasibility information to guide task planning improves scalability of the overall planner. Our key abstractions address the requirements of manipulation and object rearrangement. We validate our approach on a physical manipulator and evaluate scalability on scenarios with many objects and long plans, showing order-of-magnitude gains compared with the benchmark planner and improved scalability from additional geometric guidance. Finally, in addition to describing a new method for TMP and its implementation on a physical robot, we also put forward requirements and abstractions for the development of similar planners in the future. [ABSTRACT FROM AUTHOR]

Mobile manipulation problems involving many objects are challenging to solve due to the high dimensionality and multi-modality of their hybrid configuration spaces. Planners that perform a purely geometric search are prohibitively slow for solving these problems because they are unable to factor the configuration space. Symbolic task planners can efficiently construct plans involving many variables but cannot represent the geometric and kinematic constraints required in manipulation. We present the FFRob algorithm for solving task and motion planning problems. First, we introduce extended action specification (EAS) as a general purpose planning representation that supports arbitrary predicates as conditions. We adapt existing heuristic search ideas for solving strips planning problems, particularly delete-relaxations, to solve EAS problem instances. We then apply the EAS representation and planners to manipulation problems resulting in FFRob. FFRob iteratively discretizes task and motion planning problems using batch sampling of manipulation primitives and a multi-query roadmap structure that can be conditionalized to evaluate reachability under different placements of movable objects. This structure enables the EAS planner to efficiently compute heuristics that incorporate geometric and kinematic planning constraints to give a tight estimate of the distance to the goal. Additionally, we show FFRob is probabilistically complete and has a finite expected runtime. Finally, we empirically demonstrate FFRob’s effectiveness on complex and diverse task and motion planning tasks including rearrangement planning and navigation among movable objects. [ABSTRACT FROM AUTHOR]

Symbolic planning can provide an intuitive interface for non-expert users to operate autonomous robots by abstracting away much of the low-level programming. However, symbolic planners assume that the initially provided abstract domain and problem descriptions are closed and complete. This means that they are fundamentally unable to adapt to changes in the environment or tasks that are not captured by the initial description. We propose a method that allows an agent to automatically extend the abstract description of its skill set upon encountering such a situation. We introduce strategies for generalizing from previous experience, completing sequences of key actions and discovering preconditions to ensure computational efficiency. The resulting system is evaluated on a symbolic planning benchmark task and on object rearrangement tasks in simulation. Compared to a Monte Carlo Tree Search baseline, our strategies for efficient search have on average a 25% higher success rate at a 67% faster runtime. Code is available at https://github.com/ethz-asl/high_level_planning. • Discovery of open-world symbolic action descriptions from interactions. • Generate descriptions used by task and motion planning solvers. • Strategy for completing action sequences to speed up exploration. • Utilize previous experiences to generalize to novel scenarios. • Integrate non-expert user demonstrations to speed up exploration. [ABSTRACT FROM AUTHOR]