探討執行雙重作業下駕駛分心/心智負荷及警示音回饋的影響

Abstract

駕駛者分心與不注意力是發生交通事故的重要因素，當駕駛者分心與不適當的注意力轉移，容易導致車禍發生。有鑑於此，本研究是在虛擬實境的駕駛環境中，設計兩項實驗，以探討駕駛者執行雙重作業之分心與工作負荷對反應時間的影響以及警示回饋對反應時間的影響。分別為，研究主題一為駕駛者執行雙重作業（開車作業、認知作業），在不同時距設計(stimulus onset asynchrony, SOA)下的分心與工作負荷，及開車與認知作業之間關係。研究主題二為探討警示音回饋的影響。 研究主題一要求受試者於不同時距下，執行雙重作業，探討駕駛者產生分心/工作負荷現象，此研究假設開車作業與認知作業，會受到不同時距類型的影響。本研究以開車作業之反應時間、回答數學計算之反應時間及答對率為測量分析變數。共計有25位本校學生參與實驗，分析實驗結果指出，(1) 執行雙重作業狀況下，開車作業績效，不受次要作業干擾影響，分心效應沒有產生；(2) 次要作業(認知作業)於雙重作業狀況下，心智工作負荷比較高。此外本研究也發現，受試者於0ms SOA條件下之駕駛策略，乃以主要戶業開車安全為優先，次要作業數學計算將造成心智負荷過重現象。 研究主題二嘗試以單調開車場景，在駕駛者疲勞、注意力不集中或瞌睡的狀況下，探討何種頻率及警告類型警示效果比較好。除此之外，本研究也透過問卷來完整表達從受試者主觀感受。在結果分析方面，本研究透過Two-way ANOVA分析，以警告聲及頻率類型為自變項，反應時間為依變項。共計有13位本校學生參與實驗，分析實驗結果指出，(1) 警告聲音具有警示效果；(2)連續型警示音及1,750 Hz與3,000 Hz頻率警示效果較好。此外從問卷結果也發現，超過92%的受試者同意脈衝警告聲音有提醒效果，而且，有92%的受試者同意警告聲音比較有效。 本研究也根據研究主題一與二之研究結果，考量駕駛者能力限制及開車環境，結合人因介面及腦機介面的設計元素，發展雙重作業及偵測分心警告預警系統，提供降低心智工作負荷之建議及避免因駕駛者分心、疲勞或不注意力不集中，所造成的意外傷害。

Driver distraction and inattention are important issues of driving safety. It easily leads to traffic accidents when drivers shift their attention away from the driving task. Based on this circumstance, we conducted two studies in a virtual reality driving environment to explore the effects on reaction time of driver distraction and workload, and auditory warnings. In first study, the aim was to discover how answering math equations affected driving workload and found out the relationship between responding to car deviation and answering math equations. The second study was aim to investigate what kind of warning sounds and frequency could effectively give a caution and prevent traffic accidents caused by inattentiveness during driving. In first study, we hypothesized that the stimulus onset asynchrony (SOA) affects dual-task (driving & cognitive task) performance. We measured and evaluated the dependent variables in two aspects, driving-related task and workload-related task. The response time (RT) for responding car deviation was a dependent variable of driving-related task. Other dependent variables of workload-related measures were RT of giving correct answers and the number of incorrect answers of arithmetic task. The total of 25 volunteer participants was in this study. This experiment demonstrated that the performance of driving task did not be affected by interruption of arithmetic task, the distraction effect did not occur. We also found that the mental workload for secondary task (cognitive task) was much higher than that for the primary task (driving task). Additionally, the study found that participants’ driving strategy in the dual-task conditions took the driving manipulation as a priority. The more attentional resources consumed by the primary task, the fewer there are for the secondary task. Therefore, the secondary task caused participants' cognitive load to be maximized. Reaction time on the secondary task is thus seen as an indicator of demanded attentional capacity amount of the primary task. In second study, we attempted to identify what kind of auditory warning signals and characteristics could maintain driving performance. The study was carried out under seven conditions: with no warnings, with continuous warning tones (500 Hz, 1,750 Hz, 3,000 Hz), and with tone burst (500 Hz, 1,750 Hz, 3,000 Hz). The conditions were examined in random order and we ran different experimental sessions for the different conditions. Besides, we designed a questionnaire to measure individual experiences subjectively. To assess influence of stimulus characteristics on improvement in driving performance, this study conducted a test based on two-way ANOVA (type of sounds × carrier frequency, CF). The total number of volunteers was 13 participated in this study. The experiment result indicated that the mean response time (RT) of the warning sessions was also significantly less than for the without-warning sessions. Comparing the effectiveness of varying carrier frequency on RT, continuous tone, two sound frequencies, 1,750 and 3,000 Hz, were more effective in accelerating driver’s reaction time than the frequency of 500 Hz. In addition, more than 92% of the participants agreed with the statement that tone burst type warning sound gave efficacious warning. Also 92% of al1 respondents agreed that the warning sound frequency of 1,750 Hz (69%) and 3,000 Hz (23%) had better effectiveness. Based on the result of these two studies, the cognitive tasks could not significantly affect the performance of any assessed driving tasks. And the spectral characters in terms of the carrier frequency (CF) significantly influenced the warning effectiveness. Hence, the mental workload, driver’s distraction, and traffic safety are intrinsically intertwined. To reduce injury caused by driver’s distraction, fatigue, inattention, and mental load, we need to develop a real-time detection warning system of dual-task and driver’s drowsiness. It would combine design elements of human interface technology and brain-computer interface, and take into consideration the driver and their capabilities in the driving environment.