Abstract: An industrial safety system implements predictive method of detecting possible hazardous interactions between humans and machines or other hazardous entities. The system receives sensor data indicative of the locations of people and hazardous entities within the plant environment. Sensor data corresponding to each detected entity is aggregated, ranked, and weighed based on determined reliabilities of the respective data sources. The resulting ranked and weighed location information is tracked and analyzed by the system to determine current and predicted trajectories of the detected objects, and the system analyzes these trajectories to determine a probability of a future hazardous interaction between a human and a hazardous entity. In response to determining that a likelihood of a future hazardous interaction between a person and a machine exists, the system delivers a control output that places the machine in a safe state or a notification output warning of the hazard.