Dynamic path re-planning under emergency in very low level (VLL) complex urban airspace

Abstract

Unmanned Aerial Vehicles (UAVs) have been developed for a myriad of applications across several industries and has potential for integration into urban airspace. To operate efficiently and safely in a complex urban airspace, such drones require robust path finding, dynamic re-planning, and obstacle avoidance capabilities. We recommend the use of MATLAB software for development and simulation, where Rapidly exploring Random Tree* Algorithm (RRT*) was selected as a global planner to generate obstacle free paths, and a host drone was equipped with lidar sensors and Vector Field Histogram (VFH) controller for obstacle detection and avoidance. While it was unable to resolve a conflict with temporary obstacles generated along the global path, a dynamic path re-planner was created with A* algorithm to perform real time path re-planning away from the obstacle. Finally, a path optimizer was used to smooth out global path to reduce large inflections in the global route and allow for a more flyable path. In MATLAB, simulations were conducted to validate the desired capabilities and achieved a fully obstacle-free re-planned flight.