NAMR-RRT: Neural Adaptive Motion Planning for Mobile Robots in Dynamic Environments

TL;DR

The Neural Adaptive Multi-directional Risk-based Rapidly-exploring Random Tree (NAMR-RRT) algorithm is introduced, integrating neural network-generated heuristic regions to guide the search process by integrating neural network-generated heuristic regions to guide the search process.

Abstract

Robots are increasingly deployed in dynamic and crowded environments, such as urban areas and shopping malls, where efficient and robust navigation is crucial. Traditional risk-based motion planning algorithms face challenges in such scenarios due to the lack of a well-defined search region, leading to inefficient exploration in irrelevant areas. While bi-directional and multi-directional search strategies can improve efficiency, they still result in significant unnecessary exploration. This article introduces the Neural Adaptive Multi-directional Risk-based Rapidly-exploring Random Tree (NAMR-RRT) to address these limitations. NAMR-RRT integrates neural network-generated heuristic regions to dynamically guide the exploration process, continuously refining the heuristic region and sampling rates during the planning process. This adaptive feature significantly enhances performance compared to neural-based methods with fixed heuristic regions and sampling rates. NAMR-RRT improves planning efficiency, reduces trajectory length, and ensures higher success by focusing the search on promising areas and continuously adjusting to environments. The experiment results from both simulations and real-world applications demonstrate the robustness and effectiveness of our proposed method in navigating dynamic environments. A website about this work is available at https://sites.google.com/view/namr-rrt.

Figures (12)

• Figure 1: The diagram of the robot's navigation in a dynamic environment. The robot starts at the red flag and moves towards the green flag. Static and moving pedestrians are shown as blue and red icons. The Heuristic Region (yellow) guides the robot's search. The Neural-based Trajectory (red) highlights the efficient trajectory guided by this region, while the Random Trajectory (green) represents an inefficient trajectory from random sampling.
• Figure 2: The structure of the time-based tree.
• Figure 3: A diagram of the architecture of the NAMR-RRT algorithm. The Neural Adaptive Guiding Function first generates the heuristic region based on the input map, followed by a BFS to identify waypoints. The Multi-directional Searching Function then creates subTrees within the heuristic region, while the Risk-aware Growing Function grows the rootTree. The heuristic region and subTrees guide the growth of the rootTree, ultimately generating control instructions for the robot. Additionally, as the robot moves, its position relative to the waypoints is continually assessed, prompting updates to the heuristic region.
• Figure 4: The search process of NAMR-RRT at different key stages. (a) t = 1.3 s, (b) t = 3.9 s, (c) t = 11.9 s, (d) t = 15.1 s, (e) t = 19.2 s, and (f) t = 20.8 s.
• Figure 5: Three simulation maps. In each map, the robot is represented by the blue rectangle from the start point (red dot) to the goal point(green dot). The environment contains static obstacles, illustrated as gray blocks, while dynamic obstacles, simulating crowd movement, are represented by orange icons.