A Method for Making a Fair Evaluation of Driving Styles in Different Scenarios With Recommendations for Their Improvement

Abstract

Evaluating driving from the point of view of efficiency is a complex task. The environment is continuously changing and there is no direct relationship between all the factors that affect fuel consumption. In this paper, we propose a method to evaluate driving that can fit any scenario. It is based on Data Envelopment Analysis (DEA). This technique allows us to measure the efficiency of the drivers and to ascertain which behaviors they should improve. There are many solutions to evaluate driving in the literature. However, these solutions have difficulties adapting to changes in the driving environment and the result of the evaluation could be inaccurate. A driver could be classified as less efficient when in fact the driving could be affected by factors beyond his or her control. In addition, they do not provide the sources of inefficiency. With our method, we can analyze and quantify the cause of inefficiency. This proposal has been used to evaluate drivers of a bus fleet in Spain. The results show that drivers who were rated as efficient achieve lower fuel consumption. They can save up to 9.21% in comparison with the least efficient drivers. It is also observed that anticipation is a skill that saves a large amount of fuel and that it is an aspect that most bus drivers have a lot of room for improvement.