Algorithm for Adaptive Robot Control in Case of Device Failures During Agricultural Tasks

The paper presents an adaptive control algorithm for a ground agricultural robot, designed to ensure task execution in crisis situations, such as the failure of one or more onboard devices. The algorithm considers the functional roles of the devices and their importance to the task at hand. The primary objective of this work is to enhance the robot’s autonomy by enabling dynamic adaptation in the event of device failures. А key component is a knowledge base that stores information about tasks, device purposes, their operational status, and the number of available onboard devices. The robot’s tasks are represented by predicates that account for both scenarios: operation with a fully functional set of devices and operation with a minimal set of devices required for task execution. Simulation modeling was conducted to evaluate the decision-making time under various conditions for three types of tasks. Each task type was analyzed in three scenarios: normal operation with all devices functional, partial device failure, and a crisis involving significant device failures. The results indicate that the shortest average decision-making time in crisis situations is 0.0072 μs, while for handling device failures it is 0.0083 μs. The longest decision-making time, 0.0112 μs, occurs during partial failures due to the need to search for solutions to enable task completion. The most time-consuming scenario involves enumerating all devices to identify those available for task execution. The simulation results confirm the algorithm's functionality and provide an estimate of decision-making times under various onboard device failure scenarios.