Adapt Swarm Path Planning for UAV Based on Artificial Potential Field with Birds Intelligence Extensions

Artificial potential field (APF), a concept from physic field, has been successfully adopted for path planning of unmanned aerial vehicle (UAV). The cooperation between the expulsion against obstacle and the gravity from target, ensures a global planning optimization considering obstacle avoidance. Unfortunately, under different UAV-to-UAV distance conditions, the APF also has a weakness unable to support swarm path planning for multiple UAVs due to the highly dynamic shift between the repulsion and the gravity. To utilize APF to realize robust swarm planning, we redesign the APF and embed it into the swarm avoidance mechanism from bird intelligence involving group collision avoidance (GCA) and individual collision avoidance (ICA), forming two kinds of a APF-based swarm planning respectively: GCA-APF and ICA-APF. We then propose an adapt switch mechanism for dynamically choosing GCA-APF or ICA-APF in contexts of different obstacle environment. Experiments show the effectiveness of our approach, 15.25% higher planning efficiency than that of original GCA and avoiding certain polynomial cost increase from original ICA.