Adaptive Formation Reconfiguration and Artificial Potential Field Avoidance for Multi UAVs in Constrained Spaces

Abstract

Multi-UAV formations operating in constrained environments must balance formation maintenance and local obstacle avoidance. Fixed formation geometry may become unsuitable when obstacles restrict lateral clearance, whereas abrupt formation switching can generate tracking transients. This paper proposes a virtual-leader-based adaptive formation control method that integrates restricted-space recognition, smooth formation reconfiguration, artificial potential field avoidance, inter-UAV repulsion, and speed constraints. The controller selects a V-shaped or single-file reference according to the surrounding obstacle configuration. A smoothing mechanism is introduced to ensure continuous transition between formation geometries, while the control input is continuously updated from formation-tracking, obstacle-repulsion, and collision-avoidance terms. Simulations on a multi-UAV simulation platform verify the proposed method in constrained and obstacle-rich environments. The formation can switch to a compact single-file configuration in restricted regions and recover the V formation after passing the obstacles. The results show that the controller maintains coordinated formation motion, avoids recorded obstacle contact, preserves inter-UAV separation, and keeps UAV speeds within the prescribed range.