Autonomous navigation framework for an MAV/Drone

One of the greatest advantages of the RotorS simulator is that it comes with a fully integrated and functional trajectory tracking controller. This motivates us to implement higher-level tasks such as collision avoidance and path planning without having to implement state estimation and the controller first. Additionally, in simulation, we also have access to perfect ground truth, which is usually hard to get on real systems.

Getting ready

Overall, we have the required setup and resources to experiment and develop a comprehensive autonomous navigation framework for any MAV. Once we have a working solution in a simulation, transitioning to real world systems is typically an engineering task. In this section, we would like to discuss some ideas for how to implement collision avoidance and path planning in the RotorS simulator.

Collision avoidance

A commonly used approach to solve collision avoidance on MAVs is to project the environment onto a 2D...