In recent years, obstacle avoidance in unknown environments is an increasingly important component in the numerous autonomous vehicle applications. This paper presents a novel low-cost obstacle avoidance system for multi-rotor unmanned aerial vehicles. The proposed system employs several body-fixed laser range-finders as its primary sensors. A least squares linear regression is developed for measurement filtering and obstacle identification. The Batch Informed Trees algorithm is then utilized for trajectory planning. The feasibility and performance of the resulting system is evaluated using a numerical simulation. The simulation results clearly demonstrate the feasibility of the suggested obstacle avoidance approach as well as examine the effect of various design parameters on the system performance.
展开▼
