UAV: Trajectory Generation and Simulation

Abstract

Unmanned Aerial Vehicles (UAVs) have gained mass popularity in fields such as Defense, Agriculture, Sport and Photography etc. Their wide adoption in such research projects have opened new horizons for opportunities. However, UAVs operate in a rather dynamic environment since various factors such as airspeed, temperature, pressure, turbulence etc. are capable of hampering the flight operation. There is a need for rather robust trajectory generation algorithm and system such that it provides an easy user interface to specify the path of UAV through space, alongside making the UAV capable enough to maneuver itself from the initial to final position. This project focuses on two aspects of UAV control, the first part focuses on trajectory generation using various mathematically modeled techniques for the position, velocity and acceleration expressed as a function of time. Their comparison is done based on the constraint-based optimization techniques followed by some discussions. The second part focuses on UAV simulators and lays out the foundations and requirements for them, followed by some proposed simulators and their comparison, which could be used not only for flight control simulation, but also have potential to integrate Reinforcement Learning techniques for Autonomous flight control and Autopilot systems. Finally, in-depth analysis of AirSim simulator is carried out along with its environment setup process which may be used as an Interfacing, Installation and set-up documentation for future work.