Detecting, tracking, and warning of traffic threats to police stopped along the roadside

Abstract

Despite years of research into improving the safety of police roadside stops, reckless drivers continue to injure or kill police personnel stopped on the roadside at an alarming rate. We have proposed to reduce this problem through a "divert and alert" approach, projecting lasers onto the road surface as virtual flares to divert incoming vehicles, and alerting officers of dangerous incoming vehicles early enough to take life-saving evasive action. This thesis describes the initial development of the Officer Alerting Mechanism (OAM), which uses cameras to detect and track incoming vehicles, and calculates their real-world positions and trajectories. It presents a procedure for calibrating the camera software system with the laser, as well as a system that allows an officer to draw an arbitrary laser pattern on the screen that is then projected onto the road. Trajectories violating the "no-go" zone of the projected laser pattern are detected and the officer is accordingly alerted of a potentially dangerous vehicle.