During the last decade in North America, there has been a significant of number of collisions involving first responders (e.g., field police officers) attending roadway emergencies. Extensive research has investigated in-vehicle collision avoidance systems (e.g., automatic braking systems) designed to protect drivers/passengers and pedestrians in the case of an emergency, but few studies have investigated systems designed to detect potential threats, such as fast approaching vehicles, and warn first responders that they need to take pro-active evasive actions to avoid a collision. This study aims to develop a real-time threat detection and warning system using advanced Internet-of-Things (IoT) devices (e.g., wearable warning devices) coupled with a radar-based threat detection system to protect first responders working at an emergency. The proposed system has three essential stages: 1) detection and localization; 2) threat assessment; and 3) targeted warning. In the first stage, a radar system detects various parameters such as the speed of an approaching vehicle, the vehicle’s distance from the first responder, and the time-to-collision second-by-second. In the second stage, a threat assessment using a fuzzy inference system estimates a threat value second-by-second. In the third stage, a threshold for the threat value is determined in order to decide the appropriate threat level and appropriate type of warning necessary to enable first responders to take pro-active action. This study describes and discusses the entire process and presents outcomes based on simulated collision scenarios. The simulated collision scenarios are based on data from real collisions that involved first responders in the line of duty.
