Key points are not available for this paper at this time.
This paper proposes a road surface recognition system based on a “laser radar” (LIDER), which is used to detect a lane markings for application to an automatic platooning system for trucks. To ensure the safety of automatic driving, there is a need to recognize the road surface conditions (dry, wet, etc.). This system proposes an integrated system that is not only capable of recognizing lane markings but also monitors the road surface using a laser radar scanning system. Our road surface recognition method relies on the multiple reflection intensities of laser radar and a machine learning algorithm. By using multiple reflection intensities, the recognition rate is improved. Moreover, to improve the recognition rate, an additional feature variable, called the “roughness index,” is proposed. In this paper, the concept of a road surface recognition system is proposed and six road surface conditions (dry old asphalt, moist old asphalt, flooded old asphalt, dry new asphalt, moist new asphalt, and flooded new asphalt) are recognized by the proposed algorithm. The quality of the road surface recognition is examined through comparison with long-term measurement data. The proposed method exhibits a high level of road surface recognition performance.
AKI et al. (Mon,) studied this question.