Modeling and simulation of autonomous quadrotor systems and their onboard sensors

Abstract

Simulators are valuable tools to evaluate autonomous systems and methods of their control in a cost-effective and safe fashion. There are several implementations of simulators for autonomous systems in the literature. However, they provide limited capabilities in terms of the simulation of specific autonomous methods, for example control, flight, or sensor dynamics. In this paper, we present our design of a simulator for an environment that handles multiple quadrotors with their own set of sensors. We simulate three main sensors to get spatial and environmental information: an Inertial Measurement System that contains a three-axis gyroscope and an accelerometer; a pressure sensor; and ultrasonic sensors. To provide a realistic environment, we implement the thrust force, hub force, total torque, rolling moments, and equations of motion. We provide several attitude controllers, altitude controllers, obstacle detection methods, positioning methods, and flocking methods. We then present a comparison of their performance and behavior. Until the writing of this paper, there were no other simulators that provide this number of various autonomous methods. Moreover, the object-oriented design of our simulator makes it easy to support the addition of any other controller or method. Our simulator provides a comprehensive testbed for current and future methods for autonomous quadrotors.