探討模擬駕駛平台上事件相關之腦波動態變化

Abstract

許多駕駛人曾表示在長時間、單調的駕駛環境下有打瞌睡的情況，開車打瞌睡更在每年都造成車禍及傷亡。駕駛者如果不對於路上的突發狀況做出反應的話，交通事故的發生往往在幾秒鐘，甚至不到一秒內發生。本論文的目的在研究駕駛員從清醒至打瞌睡，對於發生車輛偏移事件之前、中及後的腦波動態變化過程，以及動態刺激對於駕駛者的腦波的影響。實驗上以虛擬實境的方式模擬夜間高速公路且單調之駕駛環境，受測者須將車輛維持在指定車道中，並在車輛發生隨機向左或向右偏移時轉動方向盤將車輛拉回指定車道。實驗分成兩部份，第一部分在靜態的模擬平台下實驗(平台不動)，第二部分延續第一部分的設定，但使用動態平台，車輛會隨著方向盤的轉動及偏移做出對應的翻轉或傾斜，以提供駕駛者適當的動態刺激。在腦波訊號處理上使用獨立成分分析(independent component analysis)及時頻分析(time-frequency analysis)。實驗結果可觀察到駕駛者的腦波從清醒至打瞌睡在偏移事件發生前在頂葉及枕葉的alpha波有強直(tonic)方向的明顯變化；在平台動與不動的實驗上，轉動方向盤開始之前到結束之後在頂葉、枕葉及體感覺運動區(somatomotor)的alpha及beta波上皆有發現到能量的抑制(suppression)，或者稱事件相關非同步(event-related desynchronization)；在停止轉動方向盤後，在頂葉、枕葉及體感覺運動區的alpha及beta波上出現明顯能量的回彈(rebound)，或者稱事件相關同步(event-related synchronization)；而在平台動的情況下，事件相關非同步及事件相關同步的能量變化都比平台不動還要大。我們希望經由本論文對於開車之事件相關腦波動態變化上的基礎研究，來提供發展安全駕駛偵測系統上必要且足夠的資訊。

Many drivers have reported that they became drowsy after long hours of driving in a monotonous environment. Driving while drowsy has resulted in crashes and injuries every year. Accidents could occur in a matter of seconds or less if the driver does not promptly respond to sudden events on the road. The first aim of this thesis is to investigate the electroencephalogram (EEG) activities before, during and after lane departure events from alertness to drowsiness during continuous driving. The second aim of this thesis is to study the influence of kinesthetic stimuli on EEG patterns. The event-related lane departure paradigm was implemented in a virtual reality based driving simulator on a motion platform. Subjects were required to compensate for the randomly generated lane departures in order to maintain the vehicle in the cruising lane. Each subject participated in two experimental conditions. In the motionless condition, the motion platform was inactive. In the motion condition, the motion platform simulated kinesthetic stimuli during the lane departure event. EEG data were analyzed with independent component analysis (ICA) and time-frequency analysis. The results show tonic increase in alpha-band power in occipital and parietal clusters before deviation onset as reaction time increased. Between response onset and offset, suppression or event-related desynchronization (ERD) in alpha and beta band power occurred in the somatomotor, parietal, and occipital regions in both motion and motionless conditions. Following the response offset, significant rebound or event-related synchronization (ERS) in alpha and beta band occurred in the parietal, occipital, and somatomotor clusters. The strength of ERD and ERS was stronger in the motion condition than that in the motionless condition. These results may provide necessary foundations for the development of a driving assistance system in the real world.