Audible Alert and TMA Lighting

Truck Mounted Attenuators (TMAs) are designed to mitigate crash severity. Currently, TMA drivers rely on visual checks via driving mirrors to manually trigger warnings thus placing the duty on drivers. To address this limitation, the Automated TMA Warning System (AutoTMA) replaces or augments manual driver interventions with an AI-enabled, sensor-fused platform. By integrating high-definition cameras, LiDAR, and radar with GPU-accelerated multi-task learning, AutoTMA continuously detects and classifies oncoming vehicles, segments lane and drivable areas, and calculates dynamic distance thresholds—safe, warning, and danger—in real time. Validation of the AutoTMA included comprehensive trials within a Unity 3D simulation environment and test-track deployments on Missouri Department of Transportation (MoDOT) TMAs. Through iterative refinements, the system’s response latency has been reduced from three seconds to 0.25 seconds, substantially improving both visual and audible alert accuracy. AutoTMA’s modular architecture and robust sensor calibration mechanisms ensure rapid component replacement and resilience in variable operational conditions. Drawing on insights from prior research, including National Cooperative Highway Research Program (NCHRP) 05 24, the system optimizes lighting and audio cues while integrating adaptive safety zone parameters to overcome the limitations of fixed configurations. Preliminary findings confirm AutoTMA’s ability to detect imminent collisions and deliver timely, context-sensitive warnings—significantly enhancing driver awareness and reducing the probability of TMA-involved crashes. AutoTMA marks a transformative shift in work zone safety protocols, offering a viable pathway for nationwide adoption. Future work will focus on expanding sensor modalities, further refining AI models to boost accuracy, and broadening field trials across diverse environments. By bridging the gap between manual vigilance and automated safety, the AutoTMA system not only improves operational workflows but also holds the promise of shaping policy and accelerating the integration of proactive safety technologies in transportation.

VIEW FINAL REPORT

Report number: cmr 25-007
Published: June 2025

Project number: TR202309

Authors: Yaw Adu-Gyamfi, Carlos Sun, Mark Amo-Boateng, Gahan Gandi, and Neema Jakisa Owor 

Performing organization: University of Missouri-Columbia/Missouri Center for Transportation Innovation