Fuzzy expert rule-based airborne monitoring of ground vehicle behaviour

This paper proposes an airborne monitoring methodology of ground vehicle behaviour based on a fuzzy logic to identify suspicious or abnormal behaviour reducing the workload of human analysts. With the target information acquired by unmanned aerial vehicles, ground vehicle behaviour is firstly classified into representative driving modes and then a string pattern matching theory is applied to detect pre-defined suspicious behaviours. Furthermore, to systematically exploit all available information from a complex environment and confirm the characteristic of behaviour, a fuzzy rule-based decision making is developed considering spatiotemporal environment factors as well as behaviour itself. To verify the feasibility and benefits of the proposed approach, numerical simulations on moving ground vehicles are performed using both synthetic and realistic car trajectory data.