FlashNav: Ultra-Fast Policy Training for Robot Navigation within 20 Seconds

2026-06-14

AI Summary

This work addresses the challenges of prolonged training times and deployment difficulties in deep reinforcement learning (DRL) for robotic navigation. The authors propose FlashNav, a GPU-centric ultra-fast training framework that constructs a lightweight bitmap-based simulation environment, retaining only the essential components of the navigation Markov decision process. By integrating a fully GPU-resident pipeline parallelism mechanism with a novel FastDSAC algorithm, FlashNav enables end-to-end efficient policy training. Remarkably, it achieves deployable navigation policies within 20 seconds on an RTX 5090 GPU with 100% success rate, without compromising obstacle avoidance performance. The trained policies demonstrate strong generalization capabilities, successfully transferring to both wheeled and legged physical robots in diverse static and dynamic indoor environments.

Abstract

Deep reinforcement learning has shown strong potential for robot navigation, but its practical deployment is still limited by the long wall-clock cost of policy training. This paper presents FlashNav, a GPU-first framework for ultra-fast range-based robot navigation training. To the best of our knowledge, FlashNav is the first DRL-based robot navigation framework that reaches seconds-level policy training, with the fastest deployable policy trained in less than 20 seconds. The key idea is to align simulation with the navigation MDP: FlashNav preserves the essential components for velocity-level navigation, including occupancy geometry, range sensing, goal-conditioned control, robot motion dynamics, collision handling, termination, and reset, while removing unnecessary rendering and high-fidelity physical details from the training loop. Built on a batched bitmap simulator and a fully GPU-resident training pipeline with our FastDSAC learner, FlashNav generates massive parallel navigation transitions entirely on GPU. Experiments on TurtleBot2 and Unitree Go2 show that FlashNav achieves a 100\% success-rate below 20 seconds on an RTX 5090 and remains within tens of seconds across desktop GPUs. The learned policies further transfer to physical wheeled and legged robots in static and dynamic indoor scenes, demonstrating that DRL-based navigation can be trained at seconds-level speed while preserving deployable obstacle-avoidance behavior.

Problem

• robot navigation
• deep reinforcement learning
• policy training
• training efficiency
• wall-clock time

Innovation

• ultra-fast training
• GPU-first framework
• range-based navigation
• MDP-aligned simulation
• policy transfer