力場模式建構市區幹道混合車流機車微觀模擬

Abstract

台灣的市區道路中，汽機車混合的車流形式在台灣隨處可見，而不同車種互動的關係極為複雜，時常造成交通壅塞，並且日益增加的機車數量已經在台灣造成了嚴重的交通問題，在過去十年，台灣引進了許多國外的微觀車流軟體，用以評估交通改善方案，但國外軟體面對混合車流的情況仍難以有效呈現混合車流間的特性，為解決此問題，必須建構一套具有混合車流特性的微觀車流軟體，因此必須先針對車流狀況，建構其微觀車流模式。 在混合車流之微觀車流模式中，機車的行為最為複雜，不僅需考慮縱向影響因素，由於機車的機動性能高，其側向影響因素也須納入考慮，且機車在路段行駛時並不如汽車般須限制在車道線中行駛，其側向偏移動作與汽車大相逕庭，故無法以傳統微觀車流區分跟車以及變換車道行為表達，因此為了描述此複雜的行為，本研究以無車道力場模式描述機車與周遭車輛間的互動，將車輛看作原子，以原子間作用力表達機車與其他車輛間的交互關係，並考慮不同車種間影響的差異，建構機車與其他車輛間相互作用行為。 由模式驗證中看出本研究模式能夠有效預測機車的行徑，並表現出機車受到其他車輛影響後的加減速及閃避動作，從微觀驗證中可看出，本模式對於行進方向的預測十分精準，並可在模式中重現機車的錯車以及鑽行行為，巨觀驗證的角度來看，模式中車輛間的交互作用符合混合車流的特色，能夠滿足現況之驗證指標，故可說明本模式對現況車流的解釋能力。 以此模式的解釋能力，建構屬於台灣混合車流下的微觀模擬軟體，使之更加符合真實環境，用以提供交通工程改善案評估工具，以解決台灣目前的交通狀況，使未來車流更加順暢，為本研究之期許。

In Taiwan, it is very common to see motorcycles riding with cars on the same roads everywhere, which kind of traffic flow is called “mixed traffic flow”. Under this kind of traffic flow, the interaction between different types of vehicles is very complicated, which often causes traffic congestion. In addition, the increasing quantity of motorcycle has already caused severe traffic problems in Taiwan. Over the past decade, we had been introduced many foreign microscopic traffic simulation software to assess transportation improvement programs; however, these foreign simulation software faced the same problem on efficiently presenting the characteristics of mixed traffic flow, which commonly happened in Taiwan. In order to solve this problem, it is a must to construct the software, which sets with mixed traffic microscopic traffic characteristics. Constructing the microscopic traffic simulation models based on this condition is the important part before starting to design the software. The behaviors of motorcycle forwarding are the most complicated in the microscopic traffic simulation models of the mixed traffic flow. Not only the longitudinal effects are needed to be considered, but also the lateral effects should be taken into consideration for the high mobility of motorcycles. Also, when a motorcycle rides on the road, it has no lane line restriction like a car that has to drive in the lanes. Since the lateral movements between motorcycles and vehicles are very different, the behavior of motorcycles cannot just be separated into the two: car-following and lane-changing behavior In order to describe those complex behaviors, this study took vehicles as atoms with different sizes, described the interactions between motorcycles and other vehicles under the model of non-lane force field. Also, we consider the impact of different types of vehicles and construct the behaviors of motorcycles as interacting between other vehicles. From the model validation in this study, it can be confirmed that this model can predict the behavior of motorcycles, and show the movements such as swerving and deceleration when motorcycles influenced by other vehicles. From the microscopic validation, this can be known that the model can precisely predict the direction of travel, and is able to rebuild the behaviors of motorcycles such as ‘crossing’ and ‘drilling’. From the view of macroscopic validation, the interaction mode between vehicles accords to the moving characteristics of motorcycles in the mixed traffic flow. Knowing that all validation items can pass the statistical test, this model has a high level of ability of explanation. By using the ability of explanation of this model to construct micro-simulation software for the mixed traffic flow in Taiwan, we can make it meet the real environment. Also, this can provide traffic engineering for improving the current traffic situation in Taiwan, and make the in Taiwan traffic run much smoothly.