Recent years have seen growing concern with bicyclists in terms of protection for vulnerable road users (VRU) [1]. One instance of this is the establishing of bicyclists to the Euro NCAP protocols for Automatic Emergency Braking (AEB) assessment.
In a real-world environment, however, even if AEB operates, there is still a possibility that the vehicle will not slow down sufficiently because of the timing of that operation, the vehicle environment, or the road surface environment. In this case, impact between the vehicle and the bicyclist may occur.
Consequently, there remains a necessity for technology to detect impact with bicyclists.
For this paper, the investigation was carried out with a focus on the detection of bicyclist impact in two modes. They are run-out mode and rear-end impact mode, in which detection is assumed to present a challenge because the bicyclist and the vehicle bumper do not come into contact.
As a method for detecting impact with a bicyclist in the above modes, a detection system was devised that integrates a conventional pedestrian detection system, deformation sensors mounted around the bumper surface periphery, and ADAS information. This will be referred to as the integrated impact detection system.
Bumper surface deformation sensors are used to detect minute deformations of the bumper surface caused by a bicycle. ADAS information is used to control the threshold for impact judgment. The conventional pedestrian detection system is used mainly to detect the impact of a bicyclist near the central portion of the vehicle.
We conducted CBU tests and simulations on the integrated impact detection system we invented this time and confirmed and investigated the bicyclist impact detection performance.
As a result, it was found that minute inputs to the bumper surface caused by a bicycle can be effectively detected by the deformation sensors. The deformation sensors alone are sensitive to inputs to the bumper surface. However, the reliability of impact detection can be maintained by using ADAS information to control the deformation sensor threshold when operation of the vulnerable road user protection device is required. Incorporating the conventional pedestrian detection system into this system also maintains reliability.
It was determined from these results that the combination of the conventional pedestrian detection system, deformation sensors, and ADAS information in an integrated impact detection system presented possibilities for detection of impact with bicyclists in run-out and rear-end impact modes.