2023 年 2023 巻 p. 19-27
The authors have been developing an assistance system for mobility scooters. The users of mobility scooters are usually elderly people or visually impaired people. Thus, an additional warning system can assist them when driving. The goal of this warning assistance system is not to achieve autonomous driving, but to assist human-driven vehicles with sound or vibration. Realizing autonomous driving not only requires high cost but also has the problem of attribution of accident responsibility. A more feasible solution is to provide a warning system to assist driving. This paper describes the design of our system, where the perception of various dangerous information is performed via vision algorithms and visual sensors. Finally, we demonstrate some preliminary recognition results through practical experiments.