Research on the Dynamic Path Planning of Manipulators Based on a Grid-Local Probability Road Map Method

Abstract

A Grid-Local Probability Road Map (PRM) method was proposed for the path planning of manipulators in dynamic environments. Based on the idea of boundary discretization, a double-grid model was built to obtain a mapping from dynamic obstacles to configuration space. The collision detection was simplified as a data indexing process to improve its efficiency. Times of collision detections were reduced by employing local programming strategies and the stratified sampling method. Moreover, the validity of sampling was increased. Taking the PUMA560 manipulator as a research object, the simulation experiments show that the time consumption of the proposed simplified collision-detection algorithm is about 14% of that of the standard one, and the stratified sampling is beneficial to the generation of probability maps compared with simple random sampling method. The simulation experiment of the static path planning shows that the proposed algorithm consumes an average of 10ms, which is superior to the comparison algorithm and has high efficiency and real-time performance. The simulation experiment of the dynamic path planning shows that the proposed algorithm consumes an average of 7ms per step, which is better than the comparison algorithm. The proposed algorithm can adjust the global path in real time to avoid obstacles as the environment changes. The algorithm mentioned has been proved to be efficient.INDEX TERMS Dynamic environment, double grid model, probability road map, path planning.

Performance Analysis of Probabilistic Roadmaps in Simulation for Mobile Robot Path Planning

“…Since the image represents the condensing medium, the time complexity of the algorithm will cause the plan to take a long time. For mobile robots operating in dynamic environments with dynamic paths [10] may vary, making quick decisions to avoid collisions and adapt to changes is crucial.…”

A Path Planning Method with a Bidirectional Potential Field Probabilistic Step Size RRT for a Dual Manipulator

“…However, the roadmap construction rate is unstable in three-dimensional environments. Liu et al [15] proposed a grid-local PRM method, which has high efficiency and real-time performance. However, this type of algorithm has weak scalability and a low roadmap reuse rate.…”

Deep Reinforcement Learning Based Trajectory Planning Under Uncertain Constraints

“…The other approach is PRM, which takes random samples from the configuration space of the robot and finds a collision-free path between the start and goal nodes. Liu et al ( 2021 ) proposed a Grid-Local PRM that combined a mapping model, sampling strategies, lazy collision detection, and a single local detection method. This proposed method can implement for dynamic path planning for static/dynamic obstacle avoidance.…”

“…However, it is a challenge for traditional motion planning algorithms to define a safe, collision-free HRC system, since all its parameters are established based on a specific environment which makes it difficult to adapt new workspace. Probability Road Map (PRM) and Rapidly-exploring Random Tree (RRT) for instance, are not suitable for dynamics environments, since they require higher real-time performance of algorithms to deal with dynamic obstacles, i.e., they need to construct a real-time mapping of obstacles in the configuration space so as to plan a collision-free path, which is very computationally expensive (Adiyatov and Varol, 2017 ; Kurosu et al, 2017 ; Wei and Ren, 2018 ; Wittmann et al, 2020 ; Jiang et al, 2021 ; Liu et al, 2021 ). Another common approach, potential field (PF), has less computation and better real-time control compared to PRM and RRT, however, it often gets stuck in the local minimum, and has limited performance when the obstacles are in the vicinity of the target (Flacco et al, 2012 ; Lu et al, 2018 , 2021 ; Xu et al, 2018 ; Melchiorre et al, 2019 ; Zhou et al, 2019 ).…”