A survey of learning‐based robot motion planning

Wang, Jiankun; Zhang, Tianyi; Ma, Nachuan; Li, Zhaoting; Ma, Han; Meng, Fei; Meng, Max Q.-H.

doi:10.1049/csy2.12020

Cited by 27 publications

(9 citation statements)

References 129 publications

(219 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Multiple approaches have used statistical learning to boost motion planning. Wang et al [35] present a comprehensive survey of methods that utilize a variety of learning methods to improve the efficiency of SBMPs. Multiple approaches discussed by Wang et al [35] use end-to-end deep learning to learn low-level reactive policies.…”

Section: Related Workmentioning

confidence: 99%

“…Wang et al [35] present a comprehensive survey of methods that utilize a variety of learning methods to improve the efficiency of SBMPs. Multiple approaches discussed by Wang et al [35] use end-to-end deep learning to learn low-level reactive policies. End-to-end approaches are attractive given if they succeed, they can compute solutions much faster than traditional approaches, but it is not exactly clear under which conditions these algorithms would succeed.…”

Section: Related Workmentioning

confidence: 99%

“…Formally, these end-to-end deep learning-based approaches lack the guarantees about completeness and soundness that our approach provides. Wang et al [35] also discuss approaches that use learning to aid sampling-based motion planning. We discuss a few of these approaches that are relevant to this work.…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Using Deep Learning to Bootstrap Abstractions for Hierarchical Robot Planning

Shah¹,

Srivastava²

2022

Preprint

View full text Add to dashboard Cite

This paper addresses the problem of learning abstractions that boost robot planning performance while providing strong guarantees of reliability. Although state-of-the-art hierarchical robot planning algorithms allow robots to efficiently compute long-horizon motion plans for achieving user desired tasks, these methods typically rely upon environment-dependent state and action abstractions that need to be hand-designed by experts.We present a new approach for bootstrapping the entire hierarchical planning process. This allows us to compute abstract states and actions for new environments automatically using the critical regions predicted by a deep neural network with an auto-generated robot-specific architecture. We show that the learned abstractions can be used with a novel multi-source bi-directional hierarchical robot planning algorithm that is sound and probabilistically complete. An extensive empirical evaluation on twenty different settings using holonomic and non-holonomic robots shows that (a) our learned abstractions provide the information necessary for efficient multi-source hierarchical planning; and that (b) this approach of learning, abstractions, and planning outperforms state-of-the-art baselines by nearly a factor of ten in terms of planning time on test environments not seen during training.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Using Deep Learning to Bootstrap Abstractions for Hierarchical Robot Planning

Shah¹,

Srivastava²

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…With the advancement of Deep Learning, there has been a significant amount of work focusing on creating autonomous navigation systems based on neural networks [ 4 , 5 ]. While many of these works focus on high-level planning and leave the execution of control and motion to low-level controllers [ 6 , 7 ], various methods of direct action execution from sensor inputs have also been considered [ 8 ]. A popular DL-enabled motion policy training method is performed with neural networks based on the Q-learning approach [ 9 , 10 , 11 , 12 ].…”

Section: Related Workmentioning

confidence: 99%

Leveraging Expert Demonstration Features for Deep Reinforcement Learning in Floor Cleaning Robot Navigation

Cimurs¹,

Merchán-Cruz

2022

Sensors

View full text Add to dashboard Cite

In this paper, a Deep Reinforcement Learning (DRL)-based approach for learning mobile cleaning robot navigation commands that leverage experience from expert demonstrations is presented. First, expert demonstrations of robot motion trajectories in simulation in the cleaning robot domain are collected. The relevant motion features with regard to the distance to obstacles and the heading difference towards the navigation goal are extracted. Each feature weight is optimized with respect to the collected data, and the obtained values are assumed as representing the optimal motion of the expert navigation. A reward function is created based on the feature values to train a policy with semi-supervised DRL, where an immediate reward is calculated based on the closeness to the expert navigation. The presented results show the viability of this approach with regard to robot navigation as well as the reduced training time.

show abstract

“…The advancement of formation control for autonomous vehicles has accelerated in recent years, see Ref. [3][4][5][6][7][8][9][10][11][12] and references therein. Nevertheless, handling a team in an efficient and robust manner faces new challenges compared to handling a single robot.…”

Section: Introductionmentioning

confidence: 99%

Distributed non‐ideal leader estimation and formation control for multiple non‐holonomic mobile robots

Ren

et al. 2022

IET Cyber-Syst and Robotics

View full text Add to dashboard Cite

This paper studies a distributed formation problem for non-holonomic mobile robots. Consideration of the leader dynamics of the robots as non-ideal, that is, subject to disturbances/unmodelled variables, is the distinguishing feature of this work. The issue is resolved by a distributed combined disturbance-and-leader estimator, allowing for the distributed reconstruction of the leader's signals. The estimator needs to detect the leader's information and disturbance. In order to reject such disturbance and achieve the formation asymptotically, the control law incorporates the smooth estimator's estimate of the leader disturbance. Furthermore, the stability of the total distributed formation control algorithm is also examined using the Lyapunov technique. Finally, to show the viability of the proposed theoretical results, simulations and actual experiments are carried out.

show abstract

A survey of learning‐based robot motion planning

Cited by 27 publications

References 129 publications

Using Deep Learning to Bootstrap Abstractions for Hierarchical Robot Planning

Using Deep Learning to Bootstrap Abstractions for Hierarchical Robot Planning

Leveraging Expert Demonstration Features for Deep Reinforcement Learning in Floor Cleaning Robot Navigation

Distributed non‐ideal leader estimation and formation control for multiple non‐holonomic mobile robots

Contact Info

Product

Resources

About