Linli Xu , Lie Guo , Xu Wang , Pingshu Ge , Longxin Guan
{"title":"利用集成认知架构预测驾驶员在紧急情况下的态势感知和响应时间","authors":"Linli Xu , Lie Guo , Xu Wang , Pingshu Ge , Longxin Guan","doi":"10.1016/j.trf.2025.07.007","DOIUrl":null,"url":null,"abstract":"<div><div>Because drivers are allowed to perform non-driving related tasks (NDRTs) in Level 3 (L3) automated driving, it is necessary to investigate the effects of NDRTs on driver behavior to ensure a safe transition of control. This study aims to examine drivers’ situation awareness (SA) changes and response abilities under NDRTs with different resource demands. Significantly, this study presents a novel cognitive mechanism to refine the takeover cognitive model under different NDRTs. Besides, time stress was considered as a latent variable manifested by the change in drivers’ SA recovery and takeover strategies after the takeover request (TOR). Seven single-task models were built using the Queuing Network-Adaptive Control of Thought Rational (QN-ACTR) cognitive architecture. The study combined them into a more complex takeover model naturally by utilizing the multi-task scheduling mechanism of QN-ACTR to simulate multi-tasking performance. Simultaneously, 80 drivers’ visual behavior and response time were recorded. The experiment results showed that visual resource demands significantly impaired drivers’ SA recovery and prolonged takeover decision time after the TOR. Physical resource demands primarily increased drivers’ motor readiness time and takeover time. In contrast, engaging in NDRTs that require only cognitive resources enabled drivers to remain alert during automated driving and respond more quickly after the TOR. These findings suggest that NDRTs with a moderate level of cognitive resource demand may be more appropriate for L3 automated driving. Regarding model fitness, coefficients of determination (<em>R<sup>2</sup></em>) for visual behavior before and after the TOR were 0.96 and 0.94, respectively. <em>R<sup>2</sup></em> was 0.98 for the response time. The model could fit the human data well. Overall, the modeling method and production rules in this study are reasonable representations of drivers’ takeover strategies and skills. The model can be used to explain drivers’ cognitive mechanisms under different NDRTs.</div></div>","PeriodicalId":48355,"journal":{"name":"Transportation Research Part F-Traffic Psychology and Behaviour","volume":"114 ","pages":"Pages 873-887"},"PeriodicalIF":3.5000,"publicationDate":"2025-07-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Predicting Drivers’ situation awareness and response times in the emergency situation using an integrated cognitive architecture\",\"authors\":\"Linli Xu , Lie Guo , Xu Wang , Pingshu Ge , Longxin Guan\",\"doi\":\"10.1016/j.trf.2025.07.007\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><div>Because drivers are allowed to perform non-driving related tasks (NDRTs) in Level 3 (L3) automated driving, it is necessary to investigate the effects of NDRTs on driver behavior to ensure a safe transition of control. This study aims to examine drivers’ situation awareness (SA) changes and response abilities under NDRTs with different resource demands. Significantly, this study presents a novel cognitive mechanism to refine the takeover cognitive model under different NDRTs. Besides, time stress was considered as a latent variable manifested by the change in drivers’ SA recovery and takeover strategies after the takeover request (TOR). Seven single-task models were built using the Queuing Network-Adaptive Control of Thought Rational (QN-ACTR) cognitive architecture. The study combined them into a more complex takeover model naturally by utilizing the multi-task scheduling mechanism of QN-ACTR to simulate multi-tasking performance. Simultaneously, 80 drivers’ visual behavior and response time were recorded. The experiment results showed that visual resource demands significantly impaired drivers’ SA recovery and prolonged takeover decision time after the TOR. Physical resource demands primarily increased drivers’ motor readiness time and takeover time. In contrast, engaging in NDRTs that require only cognitive resources enabled drivers to remain alert during automated driving and respond more quickly after the TOR. These findings suggest that NDRTs with a moderate level of cognitive resource demand may be more appropriate for L3 automated driving. Regarding model fitness, coefficients of determination (<em>R<sup>2</sup></em>) for visual behavior before and after the TOR were 0.96 and 0.94, respectively. <em>R<sup>2</sup></em> was 0.98 for the response time. The model could fit the human data well. Overall, the modeling method and production rules in this study are reasonable representations of drivers’ takeover strategies and skills. The model can be used to explain drivers’ cognitive mechanisms under different NDRTs.</div></div>\",\"PeriodicalId\":48355,\"journal\":{\"name\":\"Transportation Research Part F-Traffic Psychology and Behaviour\",\"volume\":\"114 \",\"pages\":\"Pages 873-887\"},\"PeriodicalIF\":3.5000,\"publicationDate\":\"2025-07-09\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Transportation Research Part F-Traffic Psychology and Behaviour\",\"FirstCategoryId\":\"5\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S1369847825002451\",\"RegionNum\":2,\"RegionCategory\":\"工程技术\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"PSYCHOLOGY, APPLIED\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Transportation Research Part F-Traffic Psychology and Behaviour","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1369847825002451","RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"PSYCHOLOGY, APPLIED","Score":null,"Total":0}
Predicting Drivers’ situation awareness and response times in the emergency situation using an integrated cognitive architecture
Because drivers are allowed to perform non-driving related tasks (NDRTs) in Level 3 (L3) automated driving, it is necessary to investigate the effects of NDRTs on driver behavior to ensure a safe transition of control. This study aims to examine drivers’ situation awareness (SA) changes and response abilities under NDRTs with different resource demands. Significantly, this study presents a novel cognitive mechanism to refine the takeover cognitive model under different NDRTs. Besides, time stress was considered as a latent variable manifested by the change in drivers’ SA recovery and takeover strategies after the takeover request (TOR). Seven single-task models were built using the Queuing Network-Adaptive Control of Thought Rational (QN-ACTR) cognitive architecture. The study combined them into a more complex takeover model naturally by utilizing the multi-task scheduling mechanism of QN-ACTR to simulate multi-tasking performance. Simultaneously, 80 drivers’ visual behavior and response time were recorded. The experiment results showed that visual resource demands significantly impaired drivers’ SA recovery and prolonged takeover decision time after the TOR. Physical resource demands primarily increased drivers’ motor readiness time and takeover time. In contrast, engaging in NDRTs that require only cognitive resources enabled drivers to remain alert during automated driving and respond more quickly after the TOR. These findings suggest that NDRTs with a moderate level of cognitive resource demand may be more appropriate for L3 automated driving. Regarding model fitness, coefficients of determination (R2) for visual behavior before and after the TOR were 0.96 and 0.94, respectively. R2 was 0.98 for the response time. The model could fit the human data well. Overall, the modeling method and production rules in this study are reasonable representations of drivers’ takeover strategies and skills. The model can be used to explain drivers’ cognitive mechanisms under different NDRTs.
期刊介绍:
Transportation Research Part F: Traffic Psychology and Behaviour focuses on the behavioural and psychological aspects of traffic and transport. The aim of the journal is to enhance theory development, improve the quality of empirical studies and to stimulate the application of research findings in practice. TRF provides a focus and a means of communication for the considerable amount of research activities that are now being carried out in this field. The journal provides a forum for transportation researchers, psychologists, ergonomists, engineers and policy-makers with an interest in traffic and transport psychology.