Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Multi-Goal Motion Memory

Yuanjie Lu, Dibyendu Das, Erion Plaku, Xuesu Xiao

TL;DR

This work addresses the computational burden of multi-goal motion planning in obstacle-rich, dynamically changing environments. It introduces Multi-Goal Motion Memory, a memory-augmented framework that predicts collision-free, dynamically-feasible trajectories and goal-to-goal distances to guide sampling, build a memory-informed motion map, and compute a memory-guided TSP tour, which accelerates motion-tree growth. Experiments with car- and snake-like robots demonstrate up to 90% faster planning while maintaining plan quality comparable to or better than state-of-the-art baselines. The approach is planner-agnostic and shows promise for generalization across robot models and potential multi-robot extensions.

Abstract

Autonomous mobile robots (e.g., warehouse logistics robots) often need to traverse complex, obstacle-rich, and changing environments to reach multiple fixed goals (e.g., warehouse shelves). Traditional motion planners need to calculate the entire multi-goal path from scratch in response to changes in the environment, which result in a large consumption of computing resources. This process is not only time-consuming but also may not meet real-time requirements in application scenarios that require rapid response to environmental changes. In this paper, we provide a novel Multi-Goal Motion Memory technique that allows robots to use previous planning experiences to accelerate future multi-goal planning in changing environments. Specifically, our technique predicts collision-free and dynamically-feasible trajectories and distances between goal pairs to guide the sampling process to build a roadmap, to inform a Traveling Salesman Problem (TSP) solver to compute a tour, and to efficiently produce motion plans. Experiments conducted with a vehicle and a snake-like robot in obstacle-rich environments show that the proposed Motion Memory technique can substantially accelerate planning speed by up to 90\%. Furthermore, the solution quality is comparable to state-of-the-art algorithms and even better in some environments.

Multi-Goal Motion Memory

TL;DR

This work addresses the computational burden of multi-goal motion planning in obstacle-rich, dynamically changing environments. It introduces Multi-Goal Motion Memory, a memory-augmented framework that predicts collision-free, dynamically-feasible trajectories and goal-to-goal distances to guide sampling, build a memory-informed motion map, and compute a memory-guided TSP tour, which accelerates motion-tree growth. Experiments with car- and snake-like robots demonstrate up to 90% faster planning while maintaining plan quality comparable to or better than state-of-the-art baselines. The approach is planner-agnostic and shows promise for generalization across robot models and potential multi-robot extensions.

Abstract

Autonomous mobile robots (e.g., warehouse logistics robots) often need to traverse complex, obstacle-rich, and changing environments to reach multiple fixed goals (e.g., warehouse shelves). Traditional motion planners need to calculate the entire multi-goal path from scratch in response to changes in the environment, which result in a large consumption of computing resources. This process is not only time-consuming but also may not meet real-time requirements in application scenarios that require rapid response to environmental changes. In this paper, we provide a novel Multi-Goal Motion Memory technique that allows robots to use previous planning experiences to accelerate future multi-goal planning in changing environments. Specifically, our technique predicts collision-free and dynamically-feasible trajectories and distances between goal pairs to guide the sampling process to build a roadmap, to inform a Traveling Salesman Problem (TSP) solver to compute a tour, and to efficiently produce motion plans. Experiments conducted with a vehicle and a snake-like robot in obstacle-rich environments show that the proposed Motion Memory technique can substantially accelerate planning speed by up to 90\%. Furthermore, the solution quality is comparable to state-of-the-art algorithms and even better in some environments.
Paper Structure (24 sections, 10 equations, 5 figures, 2 algorithms)

This paper contains 24 sections, 10 equations, 5 figures, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. Multi-Goal Motion Planning
  4. Machine Learning for Motion Planning
  5. Preliminaries
  6. Problem Definition
  7. Classical Sampling-Based Multi-Goal Motion Planning
  8. Single-Start-Goal Motion Memory
  9. Methodology
  10. Multi-Goal Motion Memory
  11. Multi-Goal Motion Planner Integration
  12. Implementation
  13. Guided Motion Map Construction
  14. Multi-Goal Motion Planning Augmented with Motion Memory
  15. Experiments and Results
  16. ...and 9 more sections

Figures (5)

  • Figure 1: In multi-goal motion planning, the robot must visit each goal ($G_i, \forall i$) while avoiding collisions. Classical multi-goal motion planners require extensive computation when facing a new planning problem (shown in green), which can be significantly reduced by Multi-Goal Motion Memory.
  • Figure 2: Different Planning Problem Classes: random, curve, maze, and storage (Left to Right) and Car and Snake Robot.
  • Figure 3: Example Goal-to-Goal Path Predictions by Multi-Goal Motion Memory in random, curve, maze, and storage (Left to Right).
  • Figure 4: Runtime Results with Different Numbers of Goals: (a) Multi-Goal Motion Memory [car]; (b) DROMOS [car]; (c) SequentialRRT [car]; (d) Multi-Goal Motion Memory [snake]; (e) DROMOS [snake]; and (f) SequentialRRT [snake]. Entries marked with $X$ indicate failure.
  • Figure 5: Solution Distance with Different number of Goals: (a) Multi-Goal Motion Memory (MM); (b) DROMOS.