PA-MPPI: Perception-Aware Model Predictive Path Integral Control for Quadrotor Navigation in Unknown Environments

2025-09-18

AI Summary

This work addresses three core challenges in quadrotor navigation within unknown environments: non-convex free-space geometry, high-dimensional dynamical constraints, and the need for active exploration. To this end, we propose an autonomous navigation framework integrating model predictive control (MPC) with a perception-aware mechanism. Our method introduces an online trajectory replanning strategy that employs perception gain to prioritize observation of unknown regions, thereby dynamically expanding the traversability map. Path optimization is performed in real time (50 Hz) using a sampling-based approach grounded in full-state dynamical modeling, jointly minimizing perception cost and motion feasibility. Hardware experiments demonstrate that our approach doubles navigation success rate (100% improvement) over baseline methods under severe occlusion, while also enabling safe, robust maneuver generation even when target poses are kinematically unreachable. The framework serves as a foundational navigation policy for general-purpose autonomous systems.

Technology Category

• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Intelligent Robots: Motion and Path Planning
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

Quadrotor navigation in unknown environments is critical for practical missions such as search-and-rescue. Solving it requires addressing three key challenges: the non-convexity of free space due to obstacles, quadrotor-specific dynamics and objectives, and the need for exploration of unknown regions to find a path to the goal. Recently, the Model Predictive Path Integral (MPPI) method has emerged as a promising solution that solves the first two challenges. By leveraging sampling-based optimization, it can effectively handle non-convex free space while directly optimizing over the full quadrotor dynamics, enabling the inclusion of quadrotor-specific costs such as energy consumption. However, its performance in unknown environments is limited, as it lacks the ability to explore unknown regions when blocked by large obstacles. To solve this issue, we introduce Perception-Aware MPPI (PA-MPPI). Here, perception-awareness is defined as adapting the trajectory online based on perception objectives. Specifically, when the goal is occluded, PA-MPPI's perception cost biases trajectories that can perceive unknown regions. This expands the mapped traversable space and increases the likelihood of finding alternative paths to the goal. Through hardware experiments, we demonstrate that PA-MPPI, running at 50 Hz with our efficient perception and mapping module, performs up to 100% better than the baseline in our challenging settings where the state-of-the-art MPPI fails. In addition, we demonstrate that PA-MPPI can be used as a safe and robust action policy for navigation foundation models, which often provide goal poses that are not directly reachable.

Problem

• Addresses quadrotor navigation in unknown environments
• Overcomes exploration limitations in occluded goal scenarios
• Integrates perception-awareness for adaptive trajectory optimization

Innovation

• Perception-aware MPPI for quadrotor navigation
• Online trajectory adaptation using perception objectives
• Biases trajectories to explore unknown occluded regions