Occlusion-Aware Ground Target Tracking by a Dubins Vehicle Using Visibility Volumes

2025-06-03

AI Summary

In dense urban environments, building occlusions frequently interrupt line-of-sight (LOS) tracking of ground-moving targets by unmanned aerial vehicles (UAVs). Method: This paper proposes a dynamic evasion strategy based on time-varying visibility volumes (VVs). It introduces a novel dynamic VV modeling framework and an adaptive static-VV sequence approximation method. For the first time, it theoretically proves and implements stable tracking of time-varying circular loiter orbits by Dubins-constrained UAVs under dynamic-VV constraints. The approach integrates VV modeling, Dubins-path feasibility analysis, piecewise linearized circular orbit interpolation, and nonlinear feedback control subject to LOS constraints. Contribution/Results: Theoretical analysis guarantees orbit feasibility within the dynamic-VV constraint. Simulation and real-flight experiments demonstrate significant improvements in target retention rate and trajectory robustness, effectively mitigating perception interruptions caused by occlusions in high-density urban areas.

Technology Category

• Computer Vision: Motion & Tracking
• Intelligent Robots: State Estimation
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Security and Privacy: Tracking, profiling, and countermeasures against them
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

This paper considers the problem of tracking a point of interest (POI) moving along a known trajectory on the ground with an uncrewed aerial vehicle (UAV) modeled as a Dubins vehicle using a line-of-sight (LOS) sensor through an urban environment that may occlude the POI. A visibility volume (VV) encodes a time-varying, three-dimensional representation of the sensing constraints for a particular POI position. A constant-altitude, translating, and radially time-varying circular standoff orbit is then inscribed within the dynamically changing VV centered at the POI position. The time-varying VV is approximated by placing static VVs along the POI's trajectory using an adaptive metric that restricts the volume change of consecutive visibility volumes to below a specified rate. The time-varying circular standoff orbit is proven to be feasible for a Dubins vehicle and is approximated with a piecewise set of linearly interpolated circular orbits inside the static VVs. A steering controller is derived that drives the UAV to converge to the time-varying standoff orbit. Numerical simulations and a flight test illustrate the proposed approach.

Problem

• Tracking ground POI with UAV in occluded urban environments
• Modeling sensing constraints via dynamic visibility volumes (VV)
• Ensuring Dubins vehicle feasibility for time-varying standoff orbits

Innovation

• Uses visibility volumes for occlusion-aware tracking
• Implements time-varying circular standoff orbits
• Derives Dubins vehicle steering controller