ROS-Based Approach for Unmanned Vehicles in Civil Applications

Abstract

Unmanned vehicle is the term that describes any platform without a human operator on-board. These vehicles can be either tele-operated remotely through a control station, or autonomously driven using on-board sensors and controllers. With the advances in micro and nano electronics, the increase in computing efficiency, and the ability to work in dull, dirty and dangerous environments, modern unmanned vehicles aim at higher levels of autonomy. This is through development of accurate control systems and a high-level environment understanding, in order to perform complex tasks. The main part of autonomous vehicles is the navigation system, along with the supporting subsystems. The navigation system utilizes information from various sensors, in order to estimate the position and orientation of the vehicle, sense the surrounding environment and perform the correct maneuver to achieve its assigned task. Accordingly, this chapter presents a ROS-based architecture for two different unmanned vehicles to be used in civil applications, which are constrained by Size, Weight and Power (SWap). This architecture includes the algorithms for control, localization, perception, planning, communication and cooperation tasks. In addition, in order to validate the robustness of the presented vehicles, different experiments have been carried out in real world applications with two different types of Unmanned Aerial Vehicle (UAV). The experiments cover applications in various fields; for instance, search and rescue missions, environment exploration, transportation and inspection. The obtained results demonstrates the effectiveness of the proposed architecture and validates its functionality on actual platforms.