Evaluate Good Bus Driving Behavior with LSTM

Drivers’ behaviors and their decision can affect the probability of the traffic accident, pollutant emissions and the energy efficiency level, good driving behavior can not only reduce fuel consumption, but also improves ride comfort and safety. In this paper, a new concept, evaluation zone, is defined to distinguish special driving areas which has much influence on energy consumption and ride comfort. Then, based on reducing fuel consumption and improving ride comfort, evaluation zone based driving behavior model is proposed to obtain good driving behavior dataset for the long short-term memory (LSTM) to apply the driving behavior evaluation and driving suggestion providing tasks. By using 687# bus line’s driving data of Chongqing City, China, test results demonstrate that the developed model performs well and the LSTM could provide reliable driving evaluations and suggestions for drivers.