COAA* — An Optimized Obstacle Avoidance and Navigational Algorithm for UAVs Operating in Partially Observable 2D Environments

Abstract

Obstacle avoidance and navigation (OAN) algorithms typically employ offline or online methods. The former is fast but requires knowledge of a global map, while the latter is usually more computationally heavy in explicit solution methods, or is lacking in configurability in the form of artificial intelligence (AI) enabled agents. In order for OAN algorithms to be brought to mass produced robots, more specifically for multirotor unmanned aerial vehicles (UAVs), the computational requirement of these algorithms must be brought low enough such that its computation can be done entirely onboard a companion computer, while being flexible enough to function without a prior map, as is the case of most real life scenarios. In this paper, a highly configurable algorithm, dubbed Closest Obstacle Avoidance and A* (COAA*), that is lightweight enough to run on the companion computer of the UAV is proposed. This algorithm frees up from the conventional drawbacks of offline and online OAN algorithms, while having guaranteed convergence to a global minimum. The algorithms have been successfully implemented on the Heavy Lift Experimental (HLX) UAV of the Autonomous Robots Research Cluster in Taylor’s University, and the simulated results match the real results sufficiently to show that the algorithm has potential for widespread implementation.