Driving-PASS: A Driving Performance Assessment System for Stroke Drivers Using Deep Features

Abstract

Despite any doubts about driving safety, many stroke drivers drive again due to the absence of valid screening tools. The on-road test is considered a formal assessment, but there are safety issues in testing directly on stroke patients who are not fully capable of driving. A driving simulator is a promising tool since it provides meaningful information for identifying hazards to driving safety across different driver populations and driving conditions. Using the advantages of a driving simulator, we propose a Driving Performance Assessment System for Stroke drivers (Driving-PASS). Driving-PASS is designed not only to pre-screen invalid stroke drivers before the on-road test but also to provide problematic driving items for the use in driving rehabilitation. To design assessment classifiers, i.e., the core engine of Driving-PASS, we collect driving data from a total of twenty-seven participants in thirteen driving scenarios. Thereafter, we get subjective assessment results from ten driving evaluators in eleven assessment items. By using driving data and subjective assessment results, we construct eleven assessment classifiers for ten driving ability items and one driving suitability item. We addressed the technical challenges such as handcrafted features and imbalanced dataset by a feature extraction method using pre-trained CNN models and a resampling method. Through comprehensive performance evaluation, we build eleven accurate assessment classifiers in Driving-PASS by carefully selecting deep features in each assessment item. We envision that Driving-PASS can be used as a pre-screening tool for evaluating stroke drivers and will ultimately improve road safety.

A systematic review of major evaluation metrics for simulator-based automatic assessment of driving after stroke

“… GDS-300® developed by Gridspace Post-stroke patients: 30 Age: 47 ± 12.96 R: 16/L: 14/B: 0 Hird et al [27] , [27] Examined differences in driving performance between IS and SAH post-stroke patients. STISIM Drive® Post-stroke patients: 30 Age: 59.2 ± 11.5 IS: 15 (R: 8/L: 6/B: 1) SAH: 15 Healthy controls: 20 Age: 62 ± 10.6 Jeon et al [15] Developed a driving performance assessment system for post-stroke drivers utilizing AutoML for pre-screening and providing comprehensive details for rehabilitation purposes. STISIM Drive® Post-stroke patients: 18 Healthy controls: 9 …”

“…SA has been adopted for various use cases for post-stroke patients, such as predicting actual driving abilities [13] , as part of driving rehabilitation programs [14] , and as a pre-screening tool [15] , due to its numerous benefits. SA has several advantages over OA, including more affordable to test takers [16] .…”

“…By simulating various driving scenarios, SA provides a comprehensive assessment of driving performance, as some of the scenarios can be designed similar to driving license tests. The assessment assesses various capabilities of test takers, some found in driving license tests, including visual attention [20] ; reaction time [13] , [21] ; and decision-making based on traffic rules (e.g., [13] , [15] , [21] ), and various driving scenarios such as urban driving, highway navigation, and rural cruising [6] . Moreover, SA generates performance metrics comparable to those of OA, including signal uses [22] , [23] , speed (e.g., [13] , [15] , [21] ), and distance between vehicles [15] , [22] , [23] .…”

“…The assessment assesses various capabilities of test takers, some found in driving license tests, including visual attention [20] ; reaction time [13] , [21] ; and decision-making based on traffic rules (e.g., [13] , [15] , [21] ), and various driving scenarios such as urban driving, highway navigation, and rural cruising [6] . Moreover, SA generates performance metrics comparable to those of OA, including signal uses [22] , [23] , speed (e.g., [13] , [15] , [21] ), and distance between vehicles [15] , [22] , [23] . This comprehensive assessment capability positions SA as a valuable tool for evaluating driving proficiency.…”

Driving performance grading and analytics: learning internal indicators and external factors from multi-source data

“…Streams of studies have been conducted within the context of sensor data-enabled driving behaviour evaluation (see e.g. Toledo, 2007; Zhang et al , 2010; Sun and Elefteriadou, 2014; Granruth, 2014; Kline, 2015; Higgs and Abbas, 2014; Wang et al , 2017; Martinez et al , 2017; Ghandriz et al , 2020; Ma et al , 2020; Jeon et al , 2021; Xing et al , 2021). For instance, Sun and Elefteriadou (2014) estimated the probability of changing lanes as a function of driver characteristics and scenarios by using driver behaviour-related data.…”