駕駛跟車心理物理因素聯合不確定之實驗與分析—量子視覺流觀點

Abstract

道路駕駛者經由眼睛接受外在環境光線刺激，做為同車道內判斷跟車距離的重要參考依據。但目前跟車理論多以前車與目標車之物理距離量做為重要依據，鮮少將視覺的感知納入跟車理論中。故本研究主要目的係由量子視覺流理論之觀點出發，除分析駕駛跟車心理物理因素聯合不確定，例如跟車行為關鍵影響之駕駛者相對速度感知不確定與反應時間不確定外，亦進行環境不確定因素，例如晴天及濃霧天候條件因素之模式構建。 由於真實的跟車現象較為複雜，本研究先行嘗試以刺激-反應理論為基礎，以量子光學流理論觀點提出了一種隨機且具有動態發展潛力之感知模型來研究晴天及濃霧天候條件下相對速度感知的不確定性與駕駛者反應時間不確定性之間的關係。具體來說，該模型推測駕駛者察覺與前車相對速度和反應時間是時變的、不確定的且具有權衡關係，類似於海森堡測不準原理的形式。 本研究先進行定性分析，並以模擬道路駕駛視覺環境的駕駛模擬器，辦理晴朗天候條件下兩階段實驗，分析道路駕駛者對前車相對速度感知標準差與反應時間標準差之關聯性。結果說明了在跟車行為中駕駛者感知的不確定下，存在反應時間的標準差與感知相對速度標準差之間的關係，即由實驗發現說明了當知覺相對速度的標準差相對增加時，同時伴隨著其反應時間的標準差相對減小；反之亦同。再將其模型擴充應用至濃霧之不同天候條件下跟車行為中駕駛者感知的不確定。 本研究成果，例如反應時間部分數據，已應用於事件路段中可供自動駕駛車輛與手動駕駛車輛混合行駛車道之車流模擬。另可有助於描述在駕駛者感知的不確定下的跟車行為現象，促進道路安全之改善，亦可提供當前道路運輸安全中有關駕駛之人因工程與行為之基礎應用，甚至新的車流理論。

Road drivers receive external light stimuli via eyes as an important judgment to spacing on the same lane. The majority of car-following theories put many emphases on physical spacing between the leading vehicle and the following vehicle. However, an important basis for visual perception rarely was concerned in the car-following theory. The purpose of the research is to analyze uncertainties in car-following behavior and environment, such as the driver's uncertainty of perception of perceived relative speed and uncertainty of reaction time that play a key role on the affect of the car-following behavior, and to construct a model considering those uncertainties from a viewpoint of quantum optical flow theory. Because the real traffic phenomena are complex, the study grounds on stimulate-response theory and proposes a quantum optical flow theory that presents a stochastic and potential dynamic model on an optical flow point of view. It is a stochastic and potential dynamic driver perception model to investigate the relationship between the uncertainty of perceived relative speed and that of reaction time during car following in clear and foggy weather conditions. Specifically, the proposed model hypothesizes that driver perceived speed and reaction times are time-varying and uncertain, and correlate in a trade-off relationship mimicking the form of Heisenberg Uncertainty Principle. This study conducts qualitative analysis followed by a two-stage experiment rooted in quantum optical flow theory using data collected from a driver simulator. Then this study tests the relevant between standard deviation of perceived relative speed and that of reaction time of following vehicle driver under a clear weather condition. The results illustrate that under the driver's perception of the uncertainty in the driver’s behavior in the car following there is a trade-off relationship between the standard deviations of perceived relative speed time and the standard deviations of reaction time. That is, if the standard deviation of the perceived relative speed increases, then the standard deviation of reaction time decreases. On the contrary condition, it is similar to the trade-off relationship. Then this model was expanded to analysis those driving uncertainties in a foggy weather condition. In the application, some of the reaction time data has been applied to an automatic driving vehicle following control logic in a mixed lane where automatic and manual driven vehicles mix near the event area and adjacent to automated highway system. This study could not only help to describe in the driver's perception of uncertainty for car phenomenon, but also contribute to the improvement of road safety. Another can also provide the enhancement of current road safety in the basic application of human factors relating to driving and behavior even as the new development of car following theory.