Experimental Evaluation of a Head-On Collision Warning System Fusing Machine Learning and Decentralized Radio Sensing

This paper presents the idea of an automatic head-on-collision warning system based on a decentralized radio sensing (RS) approach. In this framework, a vehicle in receiving mode employs a continuous waveform (CW) transmitted by a second vehicle as a probe signal to detect oncoming vehicles and warn the driver of a potential head-on collision. Such a CW can easily be incorporated as a pilot signal within the data frame of current multicarrier vehicular communication systems. Detection of oncoming vehicles is performed by a machine learning (ML) module that analyzes the features of the Doppler signature imprinted on the CW probe signal by a rapidly approaching vehicle. This decentralized CW RS approach was assessed experimentally using data collected by a series of field trials conducted in three different two-lane vehicular scenarios: a high-speed highway, a rural road, and an urban road. Detection performance was evaluated for three different ML algorithms: a support vector machine radial basis function kernel, K -nearest neighbors, and Boosted Trees. The obtained results demonstrate the feasibility of the envisioned head-on-collision warning system based on the fusion of ML and decentralized CW RS.