TMA (Truck Mounted Attenuators) alert system-development and testing

Abstract

Truck Mounted Attenuators (TMAs) play a crucial role in safety of work zones as they decrease the impact of the crashes, reduce fatalities and injuries, and increase safety. However, there are almost no solid solutions to decrease the number of crashes with the TMA truck while maintaining the safety of the work zone workers. In this study, we aim to alarm the drivers following the TMA truck to avoid collisions and consequently, decrease the number and severity of the crashes. We used Unity 3D as the simulator to create scenarios with a smooth 45- degree and steeper 90-degree curves. Then we ran the simulator with different vehicles volumes and speed, including aggressive drivers in the simulator, and overall, creating different crash scenarios involving TMA trucks making some scenarios that some cars crash with the TMA truck which are rare in real life. Furthermore, computer vision has been used to define a safety zone on simulator videos to automate triggering the alarm when necessary to avoid crashes when vehicles cross the safety zone boundaries on the same lane as the TMA truck. After that, we used field videos from a TMA truck to evaluate our system. Results show that the proposed system achieved an average accuracy of 76.6 percent and 65 percent in simulator videos and TMA field video respectively. The only downside is having a fixed safety zone which causes problems when the geometry of the road changes or the TMA truck rotates to some degree and causes false alarms for the vehicles passing in the other lane. Overall, this system showed promising results and can be implemented in real-time for the TMAs to reduce collisions.