Speed Reduction Measure Based on Nudging Using Real-Time Vehicle Trajectory Acquisition With Thermal Cameras

Accident risk and severity are closely associated with speeding. Reducing vehicles' speeds can contribute to road traffic safety. Conventional speed reduction measures, such as speed limit enforcement or speed indicator devices, only have a local effect, so they are not suitable for longer road stretches with varying speed limit, e.g. narrow curves in motorway exits. We present a novel speed reduction measure based on the concept of "nudging". The measure consists of LEDs embedded in the road marking that light up in front of vehicles which exceed a predefined speed threshold. The speeds and positions of the vehicles are gathered in real-time by thermal cameras and computer vision algorithms in order to target and follow only the speeding vehicles without distracting other drivers. The speed reduction measure was implemented and evaluated at a motorway exit with a narrow curve in Eindhoven, the Netherlands. The effectiveness was analyzed using linear regression models in order to take external influences on the speed (e.g. daylight, weather) into account. The mean speed was reduced by up to 2.7 km/h and the ratio of speeding vehicles by up to 40%. The most substantial speed reduction was observed on the straight part of the exit lane, which means that drivers decelerate earlier in front of the curve rather than in the curve, where longitudinal and lateral acceleration would overlap.