Search papers, labs, and topics across Lattice.
This paper presents a novel cooperative obstacle avoidance algorithm for multi-UAV formations that integrates a goal-biased Bidirectional Artificial Potential Field method with Rapidly-exploring Random Trees (BI-APF-RRT) and affine transformations. By replacing traditional centroid path planning with BI-APF-RRT, the approach ensures global collision-free path planning while maintaining smoothness and rapid convergence through cubic B-spline interpolation. The method allows UAV formations to adaptively deform based on obstacle proximity, enabling safe navigation through complex environments without losing formation integrity.
Multi-UAV formations can now navigate complex environments safely while dynamically adapting their shape to avoid obstacles, all without losing cohesion.
Aiming at the problem that obstacle avoidance flexibility and formation integrity are difficult to coexist in multi-UAV formation motion in complex obstacle environments , and that the traditional artificial potential field (APF) method easily falls into local optima, a cooperative obstacle avoidance algorithm for multi-UAV formations integrating BI-APF-RRT and affine transformation is proposed. First, abandoning the traditional APF centroid path planning method , a goal-biased Bidirectional Artificial Potential Field method RRT (BI-APF-RRT) algorithm is adopted to conduct global collision-free path planning for the centroid of the leader formation. By introducing an improved artificial potential field and cubic B-spline interpolation, the smoothness and rapid convergence of the global path are ensured. Secondly, using the generated global path as the guiding trajectory for the formation's centroid , combined with an affine transformation matrix (including non-uniform scaling and rotation) , the formation can adaptively deform based on the distance to obstacles while moving along the optimal path. Finally, the followers track the leaders through a distributed control law , enabling the entire formation to safely cross complex obstacle areas without disassembling.
