Human-Robot Navigation using Event-based Cameras and Reinforcement Learning

2025-06-12

AI Summary

Traditional frame-based robotic navigation suffers from motion blur and high latency due to fixed-frame-rate sampling. To address this, we propose a real-time, human-centered navigation framework that fuses asynchronous event-camera sensing with multimodal perception (event streams + distance sensors). Methodologically, we adopt a two-stage reinforcement learning strategy: first pretraining the policy network via imitation learning, then refining it end-to-end using Deep Deterministic Policy Gradient (DDPG). This work is the first to integrate the event camera’s ultra-low latency and high dynamic range with a staged policy learning paradigm, significantly improving response speed, environmental adaptability, and sample efficiency. Simulation results demonstrate robust real-time pedestrian following and dynamic obstacle avoidance, validating the framework’s effectiveness and generalizability in highly agile scenarios.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Search and Optimization: Learning to Search

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: The sharing economy

Abstract

This work introduces a robot navigation controller that combines event cameras and other sensors with reinforcement learning to enable real-time human-centered navigation and obstacle avoidance. Unlike conventional image-based controllers, which operate at fixed rates and suffer from motion blur and latency, this approach leverages the asynchronous nature of event cameras to process visual information over flexible time intervals, enabling adaptive inference and control. The framework integrates event-based perception, additional range sensing, and policy optimization via Deep Deterministic Policy Gradient, with an initial imitation learning phase to improve sample efficiency. Promising results are achieved in simulated environments, demonstrating robust navigation, pedestrian following, and obstacle avoidance. A demo video is available at the project website.

Problem

• Develop real-time human-centered robot navigation using event cameras
• Overcome motion blur and latency in conventional image-based controllers
• Integrate event-based perception and reinforcement learning for obstacle avoidance

Innovation

• Combines event cameras with reinforcement learning
• Uses asynchronous event data for adaptive control
• Integrates imitation learning for efficiency