Premium
Study on the Model Structure of Human Machine Operation to Evaluate a Coincident Timing Skill
Author(s) -
HASHIMOTO KOHJIRO,
DOKI KAE,
DOKI SHINJI
Publication year - 2018
Publication title -
electronics and communications in japan
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.131
H-Index - 13
eISSN - 1942-9541
pISSN - 1942-9533
DOI - 10.1002/ecj.12049
Subject(s) - driving simulator , hidden markov model , computer science , population , license , markov chain , simulation , markov model , engineering , machine learning , artificial intelligence , demography , sociology , operating system
SUMMARY With an increase in the elderly population, traffic accidents caused by elderly drivers have become an important social issue in Japan. The primary factor of traffic accidents by elderly drivers is a reduction in their driving skill. Therefore, elderly drivers are required by law to conduct regularly a driving aptitude test in Japan. If the driving skill of the driver is inadequacy, he is subject to punishment, such as driver license revoked. However, there is not appropriate evaluation index of driving skill. We have proposed a modeling method of driving behavior to support one. Proposed driving behavior model is generated based on the actual driving data of driver, and the generated model has some kind of habit and pattern of the driver. Therefore, it is considered that it is possible to evaluate driving skill of a driver by analyzing the generated model. According this reason, we aim to propose a modeling method of driving behavior to evaluate driving skill. Here, a coincident timing skill of driver is focused as a driving skill. In the case of driving behavior, the execution timing of driving operations is decided based on the prediction of subsequent change of the outside environment information. The accuracy of this decided execution timing of operation depend on the above prediction is called as the coincident timing skill. In the proposed model, the causality between the situation around a person and operation of driver is modeled by Hidden Markov Model, and timing probability distribution is added to Hidden Markov Model in order to evaluate this coincident timing skill. The timing probability distribution expresses the execution timing of next operation of driver. Therefore, it is possible to detect the information of the coincident timing skill of driver from the timing probability distribution. In the experiment, the controller operation of a radio controlled vehicle is modeled by the proposed method, and the usefulness of the proposed model is examined through some experimental results.