PILOT: Privileged Imitation Learning for End-to-End Motion Planning of Autonomous UAVs under Partial Observability

2026-08-14

AI Summary

This study addresses the limitations imposed by dynamic constraints and insufficient perception in end-to-end UAV planning under partial observability by proposing a constraint-aware privileged imitation learning framework. The approach integrates a mapless spatiotemporal perception module with a trajectory parameterization layer and employs a dual-objective loss regularization strategy to effectively leverage expert privileged information. Experimental results demonstrate that the proposed framework achieves performance comparable to expert policies while reducing computational overhead by over 80%. Furthermore, it enables successful zero-shot cross-domain deployment in both indoor and outdoor environments, significantly enhancing the safety and generalization capabilities of visual navigation systems.

Abstract

Autonomous navigation in cluttered environments is hampered by partial observability and dynamic constraints. This paper presents PILOT, a constraint-aware privileged imitation learning framework for vision-based end-to-end UAV motion planning under partial observability. The framework distills planning strategies from a computationally intensive optimal control expert into a student policy regularized toward safety and dynamic requirements via a dual-objective loss function. To mitigate partial observability, a spatiotemporal perception fusion module using a Temporal Convolutional Network (TCN) is developed to integrate historical depth images and odometry. This module infers task-relevant latent context from historical observations, enhancing spatial awareness beyond the instantaneous FOV without maintaining persistent map memory. A trajectory parameterization layer mapping network outputs to a structured trajectory, while enabling explicit continuity, dynamic-consistency, and obstacle soft penalties during training, encouraging constraint satisfaction for unseen observations without formal guarantees. Simulations on quadrotor and fixed-wing aircraft demonstrate that PILOT achieves performance comparable to the privileged expert while reducing computational overhead by over 80\%. Successful indoor and outdoor zero-shot deployment confirms the practical feasibility and cross-domain generalization of the planner.

Problem

• Autonomous UAVs
• End-to-End Motion Planning
• Partial Observability
• Dynamic Constraints
• Cluttered Environments

Innovation

• Privileged Imitation Learning
• Spatiotemporal Perception Fusion
• Trajectory Parameterization
• Partial Observability
• Constraint-aware Policy Distillation