Neural Configuration-Space Barriers for Manipulation Planning and Control

2025-03-06

AI Summary

Motion planning and control for high-dimensional robotic arms in cluttered, dynamic environments remain challenging due to real-time safety-critical requirements and uncertainties from sensor noise and modeling errors. Method: This paper proposes a neural-network-learned configuration-space distance function (CDF) barrier method that explicitly encodes safety constraints as differentiable, real-time-evaluable differential barrier functions. We introduce the first distributionally robust CDF barrier framework—requiring no prior assumptions on noise distributions—and jointly model both model mismatch and LiDAR/point-cloud sensor noise. The approach integrates online point-cloud perception, distributionally robust optimization, and real-time nonlinear model predictive control (MPC). Results: Validated on the xArm-6 platform, the method achieves a 3.2× improvement in planning efficiency, guarantees 100% collision avoidance within the control cycle, and enables millisecond-scale safe responses using only onboard point-cloud sensing.

Technology Category

• Intelligent Robots: Motion and Path Planning
• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Natural Language Processing: Safety and Robustness

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Large-scale security measurements
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Planning and control for high-dimensional robot manipulators in cluttered, dynamic environments require both computational efficiency and robust safety guarantees. Inspired by recent advances in learning configuration-space distance functions (CDFs) as robot body representations, we propose a unified framework for motion planning and control that formulates safety constraints as CDF barriers. A CDF barrier approximates the local free configuration space, substantially reducing the number of collision-checking operations during motion planning. However, learning a CDF barrier with a neural network and relying on online sensor observations introduce uncertainties that must be considered during control synthesis. To address this, we develop a distributionally robust CDF barrier formulation for control that explicitly accounts for modeling errors and sensor noise without assuming a known underlying distribution. Simulations and hardware experiments on a 6-DoF xArm manipulator show that our neural CDF barrier formulation enables efficient planning and robust real-time safe control in cluttered and dynamic environments, relying only on onboard point-cloud observations.

Problem

• Efficient planning for high-dimensional robot manipulators
• Robust safety guarantees in dynamic environments
• Handling uncertainties in neural network-based CDF barriers

Innovation

• Neural CDF barriers for safe robot control
• Distributionally robust CDF barrier formulation
• Efficient planning with onboard point-cloud observations