Autonomous Driving of Personal Mobility by Imitation Learning from Small and Noisy Dataset

The concept of personal mobility is getting popular, and its autonomous driving specialized for individual drivers is expected for a new step. However, a large driving dataset is difficult to be collected from the individual driver of personal mobility. In addition, when the driver is not familiar with the operation of the personal mobility, the dataset will contain non-optimal data as noise. This study, therefore, focuses on an autonomous driving method for the personal mobility with such a small and noisy personal dataset. To exclude noise while maximizing the use of the normal data, we introduce a new loss function based on Tsallis statistics that weights gradients depending on the original loss function and allows us to exclude noise in the optimization phase. The experimental results showed that while the conventional autonomous driving failed to drive, the proposed method learned robustly against noise and successfully drove automatically.