Research on Collision Avoidance Warning of High-Speed Maintenance Vehicle Based on Adaptive Algorithm
Year: 2025
Type: conference-paper
Abstract: Since highway maintenance vehicles often remain stationary or travel at low speeds during operation, they face a high risk of being rear-ended. In the event of such a collision, severe consequences may occur. To protect the life and property safety of personnel in the vehicle, a rear-end collision warning system centered on a 77 GHz millimeter-wave radar sensor as the core perception unit is designed. This warning model detects vector information such as the distance and speed of the target vehicle through the radar installed at the rear of the maintenance vehicle. It then judges whether the target vehicle is in the same lane as the maintenance vehicle based on the relevant information to eliminate interference from vehicles in different lanes on the warning system. If the target vehicle is in the same lane, the system calculates the current safe distance and time-to-collision (TTC) using the dynamic safe distance calculation model for highway maintenance vehicles under different operating conditions. After outputting the results, it compares them with the actual distance and TTC, and issues an acoustic and optical alarm of the corresponding level. This timely reminds both the maintenance vehicle driver and the approaching vehicles from behind, allowing both drivers sufficient time to adjust their vehicle's driving status. This forms a complete three-in-one system integrating "hardware-software-warning" to prevent traffic accidents. Field road tests show that the designed adaptive same-lane judgment algorithm can filter out interference from vehicles in different lanes on the warning system, avoiding the "warning flooding" problem caused by false warnings triggered by vehicles in different lanes. This algorithm is applicable to same-lane judgment, provides a core guarantee for the accurate triggering of subsequent collision warning processes, and demonstrates referenceable accuracy.
Authors: Zongxue Yan, Zeyu Du, Han Li, Jiacan Lu, Yidou Xiao, Nan Gao
Institution: Chang'an University
Language: English
Subfield: Automotive Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7127565829