GundamQ: Multi-Scale Spatio-Temporal Representation Learning for Robust Robot Path Planning

2025/09/12 by Yutong Shen, Shen, Yutong, Bokai Yan +8
Computer Science · #FOS: Computer and information sciences #Human Pose and Action Recognition #Multimodal Machine Learning Applications #Robotic Path Planning Algorithms #Robotics (cs.RO)

paper · pdf · doi:10.48550/arxiv.2509.10305

openalex publication_date 2025/09/12 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

In dynamic and uncertain environments, robotic path planning demands accurate spatiotemporal environment understanding combined with robust decision-making under partial observability. However, current deep reinforcement learning-based path planning methods face two fundamental limitations: (1) insufficient modeling of multi-scale temporal dependencies, resulting in suboptimal adaptability in dynamic scenarios, and (2) inefficient exploration-exploitation balance, leading to degraded path quality. To address these challenges, we propose GundamQ: A Multi-Scale Spatiotemporal Q-Network for Robotic Path Planning. The framework comprises two key modules: (i) the Spatiotemporal Perception module, which hierarchically extracts multi-granularity spatial features and multi-scale temporal dependencies ranging from instantaneous to extended time horizons, thereby improving perception accuracy in dynamic environments; and (ii) the Adaptive Policy Optimization module, which balances exploration and exploitation during training while optimizing for smoothness and collision probability through constrained policy updates. Experiments in dynamic environments demonstrate that GundamQ achieves a 15.3% improvement in success rate and a 21.7% increase in overall path quality, significantly outperforming existing state-of-the-art methods.

Citations

• Hierarchical Deep Deterministic Policy Gradient for Autonomous Maze Navigation of Mobile Robots
• Multi-Robot Motion Planning with Diffusion Models
• Deep Reinforcement Learning with Dynamic Graphs for Adaptive Informative Path Planning
• Universal Optimality of Dijkstra via Beyond-Worst-Case Heaps
• Weakly Coupled Deep Q-Networks
• Breadcrumbs to the Goal: Goal-Conditioned Exploration from Human-in-the-Loop Feedback
• Direct Preference Optimization: Your Language Model is Secretly a Reward Model
• ARiADNE: A Reinforcement learning approach using Attention-based Deep Networks for Exploration
• Bang-Bang Boosting of RRTs
• Conservative Q-Learning for Offline Reinforcement Learning
• Deep Reinforcement Learning in a Handful of Trials using Probabilistic Dynamics Models
• Hierarchical Multiscale Recurrent Neural Networks
• Incremental Sampling-based Algorithms for Optimal Motion Planning

Related

• HDPlanner: Advancing Autonomous Deployments in Unknown Environments through Hierarchical Decision Networks
• Mobile Robot Path Planning in Dynamic Environments through Globally Guided Reinforcement Learning
• Efficient Environment Design for Multi-Robot Navigation via Continuous Control
• Research on Autonomous Robots Navigation based on Reinforcement Learning
• GRATE: a Graph transformer-based deep Reinforcement learning Approach for Time-efficient autonomous robot Exploration