Simulator-based Assessment and Rehabilitation of Driving Abilities for Stroke Driver

A driving simulator can provide a safe and economical evaluation environment for stroke drivers to assess their driving skills in challenging traffic situations. Two measurement methods, an objective measurement method and subjective measurement method, have been developed to analyze results of a driving simulator. However, the former has a difficulty that numerical error results are not intuitive to interpret, and the latter is highly subjective due to a human intervention. To address these shortcomings, we propose the Simulator-based Driving Assessment System (S-DAS). S-DAS provides objective and intuitive assessment results by automating the subjective measurement method. For the automation, we first define driving ability items by interpreting driving errors that are currently used in an on-road test. We then compute the overall score by checking the defined ability items in each driving scenario. To verify the defined ability items, we employ the Rasch model based analysis. According to the analysis, S-DAS showed 83.3% (15/18) item validity. In addition, we compare the proposed automated measurement with the objective measurement. The Pearson correlation coefficient is .73 and Intra class correlation is .837, which indicated a strong positive correlation and acceptable reproducibility. We expect that S-DAS will be beneficial for both an examiner and an examinee because it enables the examiner to redesign driving scenarios for a patient-specific driving rehabilitation program.

Picture2

Picture1

People:

  • Sanghoon Jeon (topjsh0331@dgist.ac.kr)
  • Joonwoo Son (json@dgist.ac.kr)
  • Myoungouk Park (violet1211@dgist.ac.kr)
  • Bawul Kim (kimx2277@dgist.ac.kr)
  • Haengju Lee (haengjulee@hanyang.ac.kr )
  • Sang Hyuk Son (son@dgist.ac.kr)

Reference:

Sanghoon Jeon, Joonwoo Son, Myoungouk Park, Bawul Kim, Haengju Lee, and Sang Hyuk Son, “A Simulator-based Driving Assessment System for Stroke Drivers”, submitted to IEEE Transactions on Human-Machine Systems, 2016.