A fish evasion behavior-based vector field histogram method for obstacle avoidance of multi-UAVs

Focusing on the obstacle avoidance problem of multiple unmanned aerial vehicles (multi-UAVs) in three-dimensional dynamic environment, a fish evasion behavior-based vector field histogram (FEB-VFH) method is proposed in this paper. First, an obstacle avoidance decision function inspired by fish evasion behavior is designed, which considers the size and movement of the obstacle. Second, the threat of each candidate direction is evaluated using this decision function, and a dynamic memory mechanism is introduced to reproject threat histograms from past moments to the current moment. Then the safety degrees are calculated and the optimal obstacle avoidance direction is determined while considering UAV dynamic constraints. In addition, a speed adjustment strategy based on cloud model is developed to ensure adequate reaction time when approaching obstacles. A formation control protocol based on consensus is also included to generate a desired formation after multi-UAVs avoid obstacles. Simulation results demonstrate that the FEB-VFH method can effectively avoid both static and dynamic obstacles while exhibiting a certain level of robustness against moderate disturbances. Moreover, the dynamic memory mechanism enables UAVs to escape local minima during obstacle avoidance. The proposed method also shows adaptability to different formations and complex scenarios. Compared to existing methods, the FEB-VFH method achieves a balance between obstacle avoidance performance and computational overhead. Furthermore, a parameter analysis is included to investigate the influence of key parameters on computational efficiency and obstacle avoidance performance.