Towards Fault-Tolerant AGV Task Scheduling in Flexible Manufacturing Systems Using a Tree-Based Max-Plus Predictive...

Area of Science:

• Robotics
• Intralogistics Systems
• Cyber-Physical Systems

Background:

• Efficient task assignment for mobile robots is critical for modern intralogistics operations.
• Real-world disturbances like non-stationary transport delays pose significant challenges to existing scheduling systems.

Purpose of the Study:

• To develop an integrated cyber-physical framework for enhanced mobile robot task assignment.
• To implement a fault-tolerant control (FTC) mechanism for adaptive scheduling under disturbances.
• To validate the framework's effectiveness using simulations and IoT integration.

Main Methods:

• Modeling the scheduling problem as a discrete event system using switching max-plus linear systems.
• Employing a predictive tree search algorithm with a quadratic cost function.
• Integrating KIS.BOX IoT devices for a physical dispatch layer and human-in-the-loop task injection.
• Utilizing a fault-tolerant control (FTC) mechanism to adapt to non-stationary transport delays.

Main Results:

• The proposed FTC predictive strategy significantly reduces the variance of task completion times under fault conditions.
• Comparison with a First-In-First-Out (FIFO) approach demonstrates superior performance in reducing completion time variance.
• IoT integration successfully simulated and validated the feasibility of human-in-the-loop task injection in a stochastic scenario.

Conclusions:

• The integrated cyber-physical framework offers an effective solution for efficient and robust mobile robot task assignment.
• The fault-tolerant control mechanism enhances system resilience against real-world operational disturbances.
• The study validates the practical feasibility of human-in-the-loop interventions within automated intralogistics systems.