A hierarchical path planning approach based on A ⁎ and least-squares policy iteration for mobile robots

Zuo, Lei; Guo, Qi; Xu, Xin; Fu, Hao

doi:10.1016/j.neucom.2014.09.092

Cited by 73 publications

(46 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Given that it is difficult for an individual DRL method to solve the navigation problem, hierarchical approaches are widely researched in the literature [24][25][26]. Lei et al [26] combined A* and least-squares policy iteration for mobile robot navigation in complex environments. Aleksandra et al [25] integrated sampling-based path planning with reinforcement learning (RL) agents for indoor navigation and aerial cargo delivery.…”

Section: Deep Reinforcement Learningmentioning

confidence: 99%

“…To sum up, JPS+ (P)-which we borrow from [32]-has obvious advantages over other high-level planners of hierarchical methods in the literature [24][25][26], in terms of its low precomputation costs and outstanding online performance. Moreover, it serves as an ideal for the location distribution of jump points, providing a DRL-based controller with meaningful subgoals that can completely throw the problem of local minima out of consideration.…”

Section: Global Path Planner Based On Jps+ (P)mentioning

confidence: 99%

“…In [26], the problem of local navigation is decomposed into two subproblems (i.e., approaching targets and avoiding obstacles), which easily leads to local optimal policies due to simplifying the problem by adding prior experiences. In JPS-IA3C, to acquire optimal navigation policies, the motion controller directly builds models for the entire navigation problem.…”

Section: Construction Of Navigation Pomdpmentioning

confidence: 99%

See 2 more Smart Citations

Integrating a Path Planner and an Adaptive Motion Controller for Navigation in Dynamic Environments

Zeng

Qin

et al. 2019

Applied Sciences

View full text Add to dashboard Cite

Since an individual approach can hardly navigate robots through complex environments, we present a novel two-level hierarchical framework called JPS-IA3C (Jump Point Search improved Asynchronous Advantage Actor-Critic) in this paper for robot navigation in dynamic environments through continuous controlling signals. Its global planner JPS+ (P) is a variant of JPS (Jump Point Search), which efficiently computes an abstract path of neighboring jump points. These nodes, which are seen as subgoals, completely rid Deep Reinforcement Learning (DRL)-based controllers of notorious local minima. To satisfy the kinetic constraints and be adaptive to changing environments, we propose an improved A3C (IA3C) algorithm to learn the control policies of the robots’ local motion. Moreover, the combination of modified curriculum learning and reward shaping helps IA3C build a novel reward function framework to avoid learning inefficiency because of sparse reward. We additionally strengthen the robots’ temporal reasoning of the environments by a memory-based network. These improvements make the IA3C controller converge faster and become more adaptive to incomplete, noisy information caused by partial observability. Simulated experiments show that compared with existing methods, this JPS-IA3C hierarchy successfully outputs continuous commands to accomplish large-range navigation tasks at shorter paths and less time through reasonable subgoal selection and rational motions.

show abstract

Section: Deep Reinforcement Learningmentioning

confidence: 99%

Section: Global Path Planner Based On Jps+ (P)mentioning

confidence: 99%

Section: Construction Of Navigation Pomdpmentioning

confidence: 99%

See 1 more Smart Citation

Integrating a Path Planner and an Adaptive Motion Controller for Navigation in Dynamic Environments

Zeng

Qin

et al. 2019

Applied Sciences

View full text Add to dashboard Cite

show abstract

“…In the first level, the algorithm employs grids to find a geometric path quickly, and several path points are selected as subgoals for the next level. In the second level, an approximate policy iteration algorithm denoted as leastsquares policy iteration (LSPI) is used to learn a near-optimal local planning policy that can generate smooth trajectories under the kinematic constraints of the robots [19]. Further improvements in the A * algorithm have been obtained by dividing the nodes generated by the algorithm into smaller steps, eliminating redundant nodes, and reducing the cost of step and handover [20].…”

Section: Introductionmentioning

confidence: 99%

Mobile Robot Path Planning Based on a Generalized Wavefront Algorithm

Zhang

2020

Mathematical Problems in Engineering

View full text Add to dashboard Cite

is study develops a generalized wavefront algorithm for conducting mobile robot path planning. e algorithm combines multiple target point sets, multilevel grid costs, logarithmic expansion around obstacles, and subsequent path optimization. e planning performances obtained with the proposed algorithm, the A * algorithm, and the rapidly exploring random tree (RRT) algorithm optimized using a Bézier curve are compared using simulations with different grid map environments comprising different numbers of obstacles with varying shapes. e results demonstrate that the generalized wavefront algorithm generates smooth and safe paths around obstacles that meet the required kinematic conditions associated with the actual maneuverability of mobile robots and significantly reduces the planned path length compared with the results obtained with the A * algorithm and the optimized RRT algorithm with a computation time acceptable for real-time applications. erefore, the generated path is not only smooth and effective but also conforms to actual robot maneuverability in practical applications.

show abstract

“…11 In the RL framework, the robot can learn an optimal behavior policy through trialand-error by interacting with the working environment. Zuo et al 11 propose an improved A* algorithm which uses the least squares policy iteration to learn a near-optimal local planning policy that can generate smooth trajectories under kinematic constraints of the robot. Plaza et al 12 combine the RL strategies with cell-mapping techniques to solve the optimal-control problem for the car-like robot.…”

Section: Introductionmentioning

confidence: 99%

Phototropism rapidly exploring random tree: An efficient rapidly exploring random tree approach based on the phototropism of plants

Zhuge

Liu

Guo

2020

International Journal of Advanced Robotic Systems

View full text Add to dashboard Cite

Inspired by the phototropism of plants, a novel variant of the rapidly exploring random tree algorithm as called phototropism rapidly exploring random tree is proposed. The phototropism rapidly exploring random tree algorithm expands less tree nodes during the exploration period and has shorter path length than the original rapidly exploring random tree algorithm. In the algorithm, a virtual light source is set up at the goal point, and a light beam propagation method is adopted on the map to generate a map of light intensity distribution. The phototropism rapidly exploring random tree expands its node toward the space where the light intensity is higher, while the original rapidly exploring random tree expands its node based on the uniform sampling strategy. The performance of the phototropism rapidly exploring random tree is tested in three scenarios which include the simulation environment and the real-world environment. The experimental results show that the proposed phototropism rapidly exploring random tree algorithm has a higher sampling efficiency than the original rapidly exploring random tree, and the path length is close to the optimal solution of the rapidly exploring random tree* algorithm without considering the non-holonomic constraint.

show abstract

A hierarchical path planning approach based on A ⁎ and least-squares policy iteration for mobile robots

Cited by 73 publications

References 42 publications

Integrating a Path Planner and an Adaptive Motion Controller for Navigation in Dynamic Environments

Integrating a Path Planner and an Adaptive Motion Controller for Navigation in Dynamic Environments

Mobile Robot Path Planning Based on a Generalized Wavefront Algorithm

Phototropism rapidly exploring random tree: An efficient rapidly exploring random tree approach based on the phototropism of plants

Contact Info

Product

Resources

About