Sampling-based planning algorithm is a powerful tool for solving planning problems in highdimensional state spaces.In this article,we present a novel approach to sampling in the most promising regions,which significan...Sampling-based planning algorithm is a powerful tool for solving planning problems in highdimensional state spaces.In this article,we present a novel approach to sampling in the most promising regions,which significantly reduces planning time-consumption.The RRT#algorithm defines the Relevant Region based on the cost-to-come provided by the optimal forward-searching tree.However,it uses the cumulative cost of a direct connection between the current state and the goal state as the cost-to-go.To improve the path planning efficiency,we propose a batch sampling method that samples in a refined Relevant Region with a direct sampling strategy,which is defined according to the optimal cost-to-come and the adaptive cost-to-go,taking advantage of various sources of heuristic information.The proposed sampling approach allows the algorithm to build the search tree in the direction of the most promising area,resulting in a superior initial solution quality and reducing the overall computation time compared to related work.To validate the effectiveness of our method,we conducted several simulations in both SE(2)and SE(3)state spaces.And the simulation results demonstrate the superiorities of proposed algorithm.展开更多
Most existing biped robots can only walk with their feet or move by wheels.To combine the best of both worlds,this paper introduces the dynamic wheeled control including wheeled locomotion and in-situ wheel-to-foot(Wt...Most existing biped robots can only walk with their feet or move by wheels.To combine the best of both worlds,this paper introduces the dynamic wheeled control including wheeled locomotion and in-situ wheel-to-foot(WtF)transformation of a full-sized wheel-biped transformable robot SR600-II.It can traverse on flat surfaces by wheels and transform to footed stance through its switching modules when facing obstacles.For wheeled locomotion,the kinematics considering upper-body lumped center-of-mass(CoM)constraint is first derived.Then,the dynamics of wheeled locomotion is modeled as a wheeled inverted pendulum(WIP)with variables related to the pose of upper body.After that,a parameter-varying linear quadratic regulator(LQR)controller is utilized to enable dynamic wheeled locomotion.For WtF transformation,the WtF balance constraints are first revealed.Then,a WtF transformation strategy is proposed to tackle the problem when robot transforms from wheeled balance state to in-situ biped stance state.It enables the robot to pass by the transition stages in which both wheels and feet touch the ground and to maintain its balance at the same time.Simulations and experiments on the SR600-II prototype have validated the efficacy of proposed dynamic wheeled control strategies for both wheeled locomotion and in-situ WtF transformation.展开更多
Nonprehensile multiobject rearrangement is the robotic task of planning feasible paths and transferring multiple objects to their predefined target poses without grasping.It must consider how each object reaches the t...Nonprehensile multiobject rearrangement is the robotic task of planning feasible paths and transferring multiple objects to their predefined target poses without grasping.It must consider how each object reaches the target and the order in which objects move,considerably increasing the complexity of the problem.Thus,we propose a hierarchical policy for nonprehensile multiobject rearrangement based on deep-reinforcement learning.We use imitation learning and reinforcement learning to train a rollout policy.In a high-level policy,the policy network directs the Monte Carlo tree search algorithm to efficiently seek the ideal rearrangement sequence for several items.In a low-level policy,the robot plans the paths according to the order of path primitives and manipulates the objects to approach the target poses one by one.Our experiments show that the proposed method has a higher success rate,fewer steps,and shorter path length than the state-of-the-art methods.展开更多
Biomimetics is the development of novel theories and technologies by emulating the models and systems of nature.The transfer of function from biological science into engineering promotes emerging research areas across...Biomimetics is the development of novel theories and technologies by emulating the models and systems of nature.The transfer of function from biological science into engineering promotes emerging research areas across many disparate disciplines.Recently,advances in biomimetic intelligence and robotics have gained great popularity.Biomimetic robotics are designed with biological characteristics and functions to be applied in different scenarios,such as humanoid robot in the home environment,quadruped robot in the field,and bird-like flying robot in the sky.Biomimetic intelligence aims to solve many complex problems by studying the principles of biological systems,resulting in a series of efficient algorithms,such as the genetic algorithm and neural network.Biomimetic intelligence further facilitates the performance of biomimetic robotics,making it possible to be deployed in more and more practical applications.This survey introduces the development of biomimetic intelligence and biomimetic robotics.We survey different biomimetic robots,biomimetic sensors and sensing technologies,and popular biomimetic intelligence algorithms.The conclusion is drawn by discussing current challenges and future research directions.展开更多
基金supported by Shenzhen Key Laboratory of Robotics Perception and Intelligence(ZDSYS20200810171800001)the Hong Kong RGC GRF(14200618)awarded to Max Q.-H.Meng.
文摘Sampling-based planning algorithm is a powerful tool for solving planning problems in highdimensional state spaces.In this article,we present a novel approach to sampling in the most promising regions,which significantly reduces planning time-consumption.The RRT#algorithm defines the Relevant Region based on the cost-to-come provided by the optimal forward-searching tree.However,it uses the cumulative cost of a direct connection between the current state and the goal state as the cost-to-go.To improve the path planning efficiency,we propose a batch sampling method that samples in a refined Relevant Region with a direct sampling strategy,which is defined according to the optimal cost-to-come and the adaptive cost-to-go,taking advantage of various sources of heuristic information.The proposed sampling approach allows the algorithm to build the search tree in the direction of the most promising area,resulting in a superior initial solution quality and reducing the overall computation time compared to related work.To validate the effectiveness of our method,we conducted several simulations in both SE(2)and SE(3)state spaces.And the simulation results demonstrate the superiorities of proposed algorithm.
文摘Most existing biped robots can only walk with their feet or move by wheels.To combine the best of both worlds,this paper introduces the dynamic wheeled control including wheeled locomotion and in-situ wheel-to-foot(WtF)transformation of a full-sized wheel-biped transformable robot SR600-II.It can traverse on flat surfaces by wheels and transform to footed stance through its switching modules when facing obstacles.For wheeled locomotion,the kinematics considering upper-body lumped center-of-mass(CoM)constraint is first derived.Then,the dynamics of wheeled locomotion is modeled as a wheeled inverted pendulum(WIP)with variables related to the pose of upper body.After that,a parameter-varying linear quadratic regulator(LQR)controller is utilized to enable dynamic wheeled locomotion.For WtF transformation,the WtF balance constraints are first revealed.Then,a WtF transformation strategy is proposed to tackle the problem when robot transforms from wheeled balance state to in-situ biped stance state.It enables the robot to pass by the transition stages in which both wheels and feet touch the ground and to maintain its balance at the same time.Simulations and experiments on the SR600-II prototype have validated the efficacy of proposed dynamic wheeled control strategies for both wheeled locomotion and in-situ WtF transformation.
基金This project is financially supported by Shenzhen Key Laboratory of Robotics Perception and Intelligence,China(ZDSYS202008-10171800001)Hong Kong RGC CRF grant C4063-18G+2 种基金Hong Kong RGC GRF grant#14211420Hong Kong RGC GRF grant#14200618which are awarded to Max Q.-H.Meng.
文摘Nonprehensile multiobject rearrangement is the robotic task of planning feasible paths and transferring multiple objects to their predefined target poses without grasping.It must consider how each object reaches the target and the order in which objects move,considerably increasing the complexity of the problem.Thus,we propose a hierarchical policy for nonprehensile multiobject rearrangement based on deep-reinforcement learning.We use imitation learning and reinforcement learning to train a rollout policy.In a high-level policy,the policy network directs the Monte Carlo tree search algorithm to efficiently seek the ideal rearrangement sequence for several items.In a low-level policy,the robot plans the paths according to the order of path primitives and manipulates the objects to approach the target poses one by one.Our experiments show that the proposed method has a higher success rate,fewer steps,and shorter path length than the state-of-the-art methods.
基金supported by Shenzhen Key Laboratory of Robotics Perception and Intelligence(ZDSYS20200810171800001)Southern University of Science and Technology,Shenzhen 518055,China,Hong Kong RGC CRF grant C4063-18GHong Kong RGC GRF grant#14200618。
文摘Biomimetics is the development of novel theories and technologies by emulating the models and systems of nature.The transfer of function from biological science into engineering promotes emerging research areas across many disparate disciplines.Recently,advances in biomimetic intelligence and robotics have gained great popularity.Biomimetic robotics are designed with biological characteristics and functions to be applied in different scenarios,such as humanoid robot in the home environment,quadruped robot in the field,and bird-like flying robot in the sky.Biomimetic intelligence aims to solve many complex problems by studying the principles of biological systems,resulting in a series of efficient algorithms,such as the genetic algorithm and neural network.Biomimetic intelligence further facilitates the performance of biomimetic robotics,making it possible to be deployed in more and more practical applications.This survey introduces the development of biomimetic intelligence and biomimetic robotics.We survey different biomimetic robots,biomimetic sensors and sensing technologies,and popular biomimetic intelligence algorithms.The conclusion is drawn by discussing current challenges and future research directions.