Embedded AI framework for adaptive control in automated manufacturing systems
Abstract. The growing need of intelligent and adaptable manufacturing systems requires the design of adaptive control strategies that can work in the dynamism and uncertainties. Conventional forms of control such as fixed-gain PID controllers may not be effective to provide optimal control in a system that incorporates process variations, tool wear and external disturbances. The paper introduces a built-in artificial intelligence (AI) framework to adaptive control of automated manufacturing systems, which allows making decisions in real-time at the edge level. The proposed architecture will integrate sensor-driven data acquisition with a mini-AI unit to execute on an embedded system to dynamically adjust control points. The hybrid control method is developed and an introduction of AI-based corrective actions to the baseline controller target are to minimize the tracking error and increase the stability of the system. The framework is applicable to a representative manufacturing system, and the performance of the system is compared with the traditional control methods. Experimentally it has been demonstrated to possess superior response time, reduced steady-state error and high robustness to various operating conditions. The proposed solution offers a computationally efficient and scalable solution to intelligent manufacturing systems of the next generation.