Development of control architectures for multi-robot agricultural field production systems

Abstract

The evolution of agricultural equipment from simple mechanical devices to state-of-the-art autonomous machines capable of performing tasks with unlimited autonomy is on the horizon. Modern autonomous machines, although equipped with numerous sensors and controls, require human supervision to warrant safe operation. In contrast, the next generation autonomous machines, in addition to the control hardware, must include intelligence to learn and react to the unknown environment in the absence of an operator. Intelligence to an Agricultural Robot (Ag-Robot) is possible through a robust control methodology/architecture that effectively interprets, arbitrates and prioritizes the information obtained from sensors to produce desired actions. In this dissertation, an Individual Robot Control Architecture (IRCA) and Multi Robot System Control Architecture (MRSCA) were developed to provide intelligence to an individual and a group of coordinated agricultural robots, respectively. The architectures were simulated using MATLAB tools SIMULINK and STATEFLOW CHART, and were deployed on three Autonomous Vehicle Platforms (AVPs) for validation. The successful testing of the AVP in two test scenarios; autonomous single vehicle navigation and autonomous single vehicle obstacle detection and avoidance, validated the Deliberative and Reactive features of the IRCA. The successful simulation of the planting, baling-bale retrieving and the harvest operations using the AVPs in three test scenarios validated the Standalone, Modest and Absolute levels of cooperation offered by the MRSCA. KEYWORDS: Agricultural Robot, Autonomous, Finite State Machine, Sensor Array, Wireless Communication