Automatic classification of motorcycle motion sensing data

Motorcycles are cost-efficient transportation but regarded as unsafe. To build safer motorcycle, the vehicle motion of motorcycles should be clarified. The motion depends on that of a rider much more than that of the body of a motorcycle, and thus motorcycle companies do not have enough motion data. Our research group has researched a sensing network with motorcycles and a vehicle-motion corpus of motorcycle. In this paper, we propose a method to attach an appropriate label to sensed data from the sensing network.