Adaptive Optimal Multiple Object Tracking Based on Global Cameras for a Decentralized Autonomous Transport System

In order to adapt to the mass customization, a new concept of material flow systems that can handle product varieties is needed. Firstly, this paper analyzes the current problems and future requirements of the structure of a new production system. Then, in response to current limitations, the corresponding concept of a decentralized transport system for low payloads with high flexibility is introduced. For this purpose, automated guided vehicles (AGVs) as an effective means of transport are used. The key issues of the autonomous transport system are then researched, and an improvement of the multiple object tracking algorithm is proposed. We demonstrate the performance of our proposed system with a designed workspace. Based on the demonstration and experiment, results show that the proposed concept and the tracking algorithm are appropriate and robust to be implemented in real-time applications.