Task Assignment in Camera Networks: A Reactive Approach for Manufacturing Environments

Camera networks have been increasingly adopted in manufacturing to enhance workplace safety and maintain production quality levels. Effective management of cameras is crucial for the network to achieve its designated monitoring goal and fulfil task-specific requirements. In this paper, we investigate the online multicamera task assignment problem and propose a reactive scheduling approach that is suitable for manufacturing environments. In our problem, tasks occur stochastically, based on detecting or receiving specific information of activities occurring in the workcell. Task-specific information such as location, priority, required resolution of detection, and number of cameras needed to monitor the task are revealed at its release time, while the task lifetime remains unknown until its actual completion time. A utility function is derived to quantify each camera's ability to monitor a task occurrence. Estimation is based on the expected task quality, measured in terms of the resolution of detection, and each camera's cost, measured in terms of its pose change and workload. We develop a framework for experimentally designing and testing camera networks in a visually and behaviorally realistic virtual simulation environment. In particular, we demonstrate our solution on a manufacturing workcell. Results show the efficiency of the presented framework, which can be applied for establishing practical camera networks.