Simulation of Autonomous UAV Navigation with Collision Avoidance and Space Awareness

This research developed a safe navigation system of an autonomous UAV within a comprehensive simulation framework. The navigation system can find a collision-free trajectory to a randomly assigned 3D target position without any prior map information. It contains four main components: mapping, localization, cognition, and control. The cognition system makes execution command based on the perceived position information about obstacles and the UAV from mapping and localization system. The control system is responsible for executing the input command made by the cognition system. Three case studies for real-life scenarios, such as space awareness, static obstacle avoidance, and dynamic obstacle avoidance, are conducted. The experiments demonstrate that the UAV can determine a collision-free trajectory under all three cases of environments. All simulated components are designed to match their real-world counterparts' dynamics and properties. Ideally, the simulated navigation framework can be transferred to a real UAV without any changes. As the navigation system is implemented modularly, it is easier to test and validate to ensure its performance. Moreover, the system has excellent readability, maintainability, and extensibility. Hence, the simulation framework provides an excellent platform for future robotic research.