Principles of construction of control systems for unmanned vehicles

Abstract

The work considers the principles of construction and organization of the control system of unmanned vehicles. Unmanned vehicles (UAVs) are a type of vehicles that move without a human pilot on board using a special autonomous control system. The key components of control systems include perception sensors, navigation subsystems and algorithms, data processing algorithms, localization algorithms, dynamic route planning algorithms, communication tools, etc. In the available sources, there are different visions of architectural solutions for UAV control systems. As a rule, such solutions depend on a specific set of sensors. The purpose of the work is to develop the principles of building control systems for unmanned vehicles. Solved tasks: analysis of means of collecting information about the state and position of unmanned vehicles; development of a generalized architecture of the control system of unmanned vehicles. A classification of sensor scanning tools is proposed. The following groups of sensor scanning modules have been defined: computer vision, odometry, coordination, range, detection. The generalized structure of the control system is developed, its separate functional subsystems are described. The presented architecture involves software implementation and includes seven subsystems: perception, localization, motion prediction, navigation, decision-making, remote control, control of executive mechanisms. In order to minimize the structure of the control system, it is possible to combine some subsystems that are similar in purpose and interaction with each other. So, localization can be included in the perception subsystem, and the navigation subsystem is integrated into the motion prediction unit. The main hardware components of the system are high-level and low-level controllers. The presented architecture can be used in the creation of unmanned vehicles.