Accelerated-Gradient-Based Flexible-Object Transport With Decentralized Robot Teams

Gombo, Yoshua; Tiwari, Anuj; Devasia, Santosh

doi:10.1109/lra.2020.3036569

Cited by 8 publications

(3 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…First, in the grasping method, a single or more robots grasp an object with a robotic arm and transport it toward a goal in 2D [12,15,16] or 3D environment [13,14,30]. Robots do not have to consider the motion of an object during transportation because the object is tightly coupled with the robotic arm; the object can be easily transported if it is successfully grasped once by a gripper.…”

Section: A Object Transportation Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Automatic Curriculum Design for Object Transportation Based on Deep Reinforcement Learning

Eoh¹,

Park

2021

IEEE Access

View full text Add to dashboard Cite

This paper presents an automatic curriculum learning (ACL) method for object transportation based on deep reinforcement learning (DRL). Previous studies on object transportation using DRL have a sparse reward problem that an agent receives a rare reward for only the transportation completion of an object. Generally, curriculum learning (CL) has been used to solve the sparse reward problem. However, the conventional CL methods should be manually designed by users, which is difficult and tedious work. Moreover, there were no standard CL methods for object transportation. Therefore, we propose an ACL method for object transportation in which human intervention is unnecessary at the training step. A robot automatically designs curricula itself and iteratively trains according to the curricula. First, we define the difficult level of object transportation using a map, which is determined by the predicted travelling distance of an object and the existence of obstacles and walls. In the beginning, a robot learns the object transportation at an easy level (i.e., travelling distance is short and there are less obstacles around), then learns a difficult task (i.e., the long travelling distance of an object is required and there are many obstacles around). Second, training time also affects the performance of object transportation, and thus, we suggest an adaptive determining method of the number of training episodes. The number of episodes for training is adaptively determined based on the current success rate of object transportation. We verified the proposed method in simulation environments, and the success rate of the proposed method was 14% higher than nocurriculum. Also, the proposed method showed 63% (maximum) and 14% (minimum) higher success rates compared with the manual curriculum methods. Additionally, we conducted real experiments to verify the gap between simulation and practical results.

show abstract

Section: A Object Transportation Methodsmentioning

confidence: 99%

“…[ [12][13][14][15][16] Pushing A single or more robots push an object toward a goal using their bodies.…”

Section: Graspingmentioning

confidence: 99%

Automatic Curriculum Design for Object Transportation Based on Deep Reinforcement Learning

Eoh¹,

Park

2021

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Formation control, a major research topic in the field of swarm intelligence, aims at exploiting local interaction rules to regulate the overall state of the system towards desired coordinated behaviours, which provides the system with significant advantages in dealing with complex tasks. 1,2 In the last decade, the generation strategy, motion planning, and topology condition of formation, 3 as well as the formation control of many autonomous unmanned systems have been investigated, for instance, the satellite, 4,5 the intelligent robots, [6][7][8] surface vessels, 9,10 underwater vehicles 11 and unmanned aerial vehicles. 12,13 The process of formation control can be divided into two tasks.…”

Section: Introductionmentioning

confidence: 99%

Enhanced affine formation maneuver control using historical acceleration command

Yu,

Yi,

Zhu

et al. 2023

Intl J Robust & Nonlinear

View full text Add to dashboard Cite

SummaryHow a network of multiple vehicles gets to a goal formation from an initial formation is as important as reaching the goal formation. Despite that stress matrix has been well applied to construct distributed affine formation control, most works have focused on the stabilization or stability issues and paid little attention to the performance regulation or the control parameterization issues. In this paper, we aim to design an error‐model based distributed control for the affine formation maneuver of second‐order multi‐agent systems (MASs), and to demonstrate the advantages of the control approach in performance tuning. The basic idea behind the control design is to force the actual error‐trajectory dynamics of each controlled follower to approximate that of an ideal second‐order error model. The designed controller possesses a notable feature that the historical acceleration command (HAC) is used to enhance system performance during the formation maneuvers. By a theory tool of neutral functional differential equations, we develop a delay‐independent condition on the design parameters, under which the system stability is robust to the actively introduced HAC. Furthermore, we derive an inequality describe the relationship between the ultimate bounds of tracking errors and the delay parameter. This paper presents a clear mechanism to tune the comprehensive performance of affine formation control. Simulation results show the effectiveness of the stability condition and the performance‐tuning mechanism, as well as the possible fragility of the system stability to the delay when the condition is not satisfied.

show abstract

A Two-Stage Trajectory Prediction Algorithm for Mobile Robots Combining the Bayesian and the DMOC Frameworks

Luo

Eberhard²

2023

Lecture Notes in Electrical Engineering

View full text Add to dashboard Cite

Accelerated-Gradient-Based Flexible-Object Transport With Decentralized Robot Teams

Cited by 8 publications

References 26 publications

Automatic Curriculum Design for Object Transportation Based on Deep Reinforcement Learning

Automatic Curriculum Design for Object Transportation Based on Deep Reinforcement Learning

Enhanced affine formation maneuver control using historical acceleration command

A Two-Stage Trajectory Prediction Algorithm for Mobile Robots Combining the Bayesian and the DMOC Frameworks

Contact Info

Product

Resources

About