| Popis: |
This chapter focuses on path planning techniques for autonomous navigation of small unmanned aerial vehicles (UAVs) in complex urban environments. It is difficult for UAVs to navigate at low altitudes in dynamic or complex urban environments because they encounter large structures, unpredictable or unrecognized environments, or other UAVs or obstacles flying in their path. Path planning is the core capability of an autonomous UAV to adapt dynamically in changing environments. This chapter will study and implement different path planning algorithms, slowly increasing complexity from optimal A ⁎ search algorithms to voxmap, probabilistic roadmap, and Rapidly Exploring Random Tree 2D grids to 3D motion planning using random sampling, heuristic, collinear, and path pruning methods. Differences in performance characteristics such as time of execution, cost, and waypoint count of different path planning algorithms are also discussed. Udacity Flying Car Nanodegree is used to design, build, and check motion planning methods for UAVs in a complex and dynamic environment. All of the path planning algorithms used in this chapter are written in a different way of logic based on the events taking place during the UAV flight, which is called event-driven programming. This chapter provides a number of effective route planning methodologies that ensure secure paths for UAVs by considering potential contingencies and taking into account various uncertainties. The simulation platform and software are open-sourced for community use. |
