DPNet: Doppler LiDAR Motion Planning for Highly-Dynamic Environments

2025-11-29

AI Summary

To address motion planning failures caused by rapidly moving obstacles in highly dynamic environments, this paper proposes a real-time planning framework integrating Doppler LiDAR. Methodologically, we introduce the Doppler Kalman Neural Network (D-KalmanNet), the first architecture to jointly model range and instantaneous velocity; it is coupled with a tunable, learning-based Model Predictive Control (MPC) scheme and a partially observable Gaussian state-space representation, enabling adaptive parameter optimization under low-data regimes. Experimental evaluation on high-fidelity simulations and real-world datasets demonstrates that our framework significantly outperforms state-of-the-art baselines: it achieves higher trajectory tracking accuracy, sustains planning frequencies exceeding 50 Hz, and reduces obstacle-avoidance response latency by 37%. The proposed solution delivers an efficient and robust perception-decision integration for autonomous navigation in dynamic environments.

Technology Category

• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Intelligent Robots: Motion and Path Planning
• Search and Optimization: Learning to Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Existing motion planning methods often struggle with rapid-motion obstacles due to an insufficient understanding of environmental changes. To address this limitation, we propose integrating motion planners with Doppler LiDARs which provide not only ranging measurements but also instantaneous point velocities. However, this integration is nontrivial due to the dual requirements of high accuracy and high frequency. To this end, we introduce Doppler Planning Network (DPNet), which tracks and reacts to rapid obstacles using Doppler model-based learning. Particularly, we first propose a Doppler Kalman neural network (D-KalmanNet) to track the future states of obstacles under partially observable Gaussian state space model. We then leverage the estimated motions to construct a Doppler-tuned model predictive control (DT-MPC) framework for ego-motion planning, enabling runtime auto-tuning of the controller parameters. These two model-based learners allow DPNet to maintain lightweight while learning fast environmental changes using minimum data, and achieve both high frequency and high accuracy in tracking and planning. Experiments on both high-fidelity simulator and real-world datasets demonstrate the superiority of DPNet over extensive benchmark schemes.

Problem

• Integrates Doppler LiDAR with motion planners for dynamic environments
• Tracks rapid obstacles using a Doppler Kalman neural network
• Enables auto-tuning controller parameters for high-frequency accurate planning

Innovation

• Doppler LiDAR integration for motion planning
• Doppler Kalman neural network for obstacle tracking
• Doppler-tuned MPC for auto-tuning controller parameters