Radar-based Rear Collision Prevention System for Snowplows

Snowplows play a vital role in maintaining road safety during winter, yet they frequently suffer collisions with other vehicles due to reduced visibility, drivers' inattention, and hazardous conditions during snowplowing. Existing collision avoidance systems primarily focus on forward object detection, whereas research on rear-approaching vehicles is limited. This paper presents a snowplow protection system that continuously tracks the trajectories of vehicles approaching from the rear using a low-cost radar and a nonlinear observer. Once a potential collision is detected, an audiovisual warning system is activated from the snowplow to alert the approaching driver. Unlike conventional vehicle tracking models that rely on global coordinate systems, this study introduces a relative motion model that accounts for the ego vehicle's rotational motion effects, improving tracking performance during turning maneuvers. Additionally, a nonlinear observer is implemented to enhance estimation accuracy across varying vehicle speeds using a single observer gain. The proposed system offers a cost-effective and practical solution for snowplow safety in harsh weather conditions, with its performance validated through both simulations of various scenarios and real-world vehicle experiments.