Autonomous Robot Orchestration Solution for OHT with Machine Learning and Digital Twin

Autonomous Robot Orchestration Solution (AROS) is revolutionizing robot fleet management. AROS identifies the state and environment of each robot and enables them to achieve common goals collaboratively. In this paper, we introduce AROS and its application in controlling massive Overhead Hoist Transport (OHT) vehicles in semiconductor fabs. Key technologies in AROS include reinforcement learning algorithms and deep auto-encoder models. Digital Twin (DT) replicates the real system in a virtual environment with real-time communication to enhance decision-making for OHTs. We demonstrate the positive impact of AROS on OHT system performance, reducing average delivery times and increasing delivery capacity. AROS anomaly detection algorithm and health monitoring solution handle unexpected errors. This eliminates the need for human intervention. Our findings are supported by real use cases in large-scale semiconductor fabs.