3D Real–time Path Planning of UAVs in dynamic environments in the presence of uncertainty

Abstract

Unmanned Aerial Vehicles (UAVs) are being integrated into all spheres of life varying in a wide range of applications from military to civil applications. In such applications, UAVs are expected to operate safely in the presence of uncertainties present in the dynamic environment and the UAV itself. Based on literature different uncertainty sources are identified, quantified and modelled using bounding shapes. The UAV model, path planner parameters and four different complexity scenarios with different moving shapes are defined in view of real UAVs. For analysis uncertainty is varied between 2% and 20% for UAV position and obstacle position and orientation. Results show that both types of uncertainty deteriorate path planning performance of both A* and RRT algorithms for all scenarios considered especially for RRT. RRT results in faster and shorter paths with approximately the same success rate (>95%) as A* for simple scenarios. For complex scenarios A* performs better. This work shows that 3D real—time path planning in different obstacle density, moving obstacle environments in the presence of uncertainty is possible.