기계 학습을 활용한 구동 토크 예측 기반 차량 속도 프로파일 최적화

Abstract

A number of studies have been proposed in order to obtain the optimal vehicle speed profile for a given route based on dynamic programming(DP). In general, solving optimization problems requires a vehicle dynamics model to accurately calculate energy consumption. However, this model cannot exactly reflect the real characteristics of various vehicles because of the nonlinearity of the rolling resistance, air resistance, and gradient resistance. Therefore, this study proposes vehicle speed optimization by using a machine learning network model that is trained from actual vehicle driving data. The performance of the proposed method is verified by simulation where the driving environment is duplicated corresponding to real driving conditions. The effectiveness of the proposed optimal speed profile is evaluated by comparing with conventional cruise control driving. As a result, driving with the optimal speed profile for a given route of 27.3 km significantly reduces battery energy consumption by 8.4 %.