透過物聯網通訊改進停車位的佔用率與停車用戶的滿意度

Abstract

隨著全世界的汽車銷售量逐年上升，駕駛對停車位的需求也急劇上升。此外，由於全世界的人口密度逐年上升，可以用來建設成停車位的土地愈來愈少。雖然隨著環保意識抬頭，公共運輸系統持續不停的完善，目標是成為經濟實惠且環保的交通替代方案，但是在許多情境下，民眾還是傾向自行開車前往目的地，因為行程較具彈性且耗時較短。因此，都市停車位的嚴重缺乏對於駕駛與停車場管理者而言，是一項難以解決的挑戰。 對駕駛而言，停車時希望能愈快找到停車位，如果可以，停車位離目的地愈近愈好。此外，當把車停在一個巨型停車場後，駕駛希望回來取車時，可以輕鬆快速的找到愛車。對停車場管理者而言，希望能知道與停車場管理相關的因素，並能制訂出一個可以滿足各類駕駛的停車需求，又能維持停車位高使用率的最佳停車場政策。本研究聚焦在如何達成管理者的目標。我們的研究方法是使用模擬系統，我們使用SimPy開發出一個停車模擬系統。SimPy是一個用Python開發出來的基於行程的離散事件模擬開放原始碼框架。 管理者可以使用我們的模擬系統來評估不同的停車場政策帶來的成效，從中找出最佳停車場政策。停車場政策是由管理者訂定的一系列規則，當駕駛想要使用該停車場，就要遵守所有規則。這些規則包含將停車位區分成不同類型，以及訂定不同的停車收費模式。管理者也可以使用我們的模擬系統來評估智慧停車系統對停車場管理帶來的影響。智慧停車系統不僅會追蹤空停車位的數量與位置，並且能透過物聯網通訊提供駕駛有用的停車場狀態資訊。 我們設計三組實驗來研究上述管理者面臨的挑戰。在實驗中，我們使用滿意度與佔用率來呈現實驗結果。模擬結果顯示，我們的系統可以檢測各式各樣的停車場政策，並從中找出在管理者給定的佔用率條件下，能得到最大滿意度的最佳停車場政策。此外，模擬結果顯示不論面臨何種停車場使用情境，智慧停車系統總是能改進滿意度。

The demand for parking space has increased dramatically due to the increase of vehicles in the world. At the same time, it is hard to create new parking spaces because of the continual growth of population density. Although public transit systems have advanced as an economical and ecofriendly alternative to cars, people prefer to drive the car to their destination because it takes less time and is more flexible for many trips. As a result, the significant shortage of parking spaces in cities creates a challenging problem for both drivers and parking space operators. For drivers, problems include searching for an available parking space quickly, locating the space that is the closest to the destination, and afterwards finding the car without pain in a big parking garage. For operators, problems include identifying all parameters of parking space management and developing the best management policy to satisfy needs of different customers while maintaining a high utilization rate. In this research, we focus on problems of operators. We develop a method of simulation to study these problems. We implement the parking simulation system by using SimPy, a process-based discrete-event simulation framework based on standard Python. The operator can use our method to evaluate different parking policies and find out the best policy. Parking policy is a set of rules established by operators. Drivers are required to obey these rules when they use the parking lot. Parking policy is created by defining different types of parking spaces and prices. The operator also can use our method to estimate the impact of Smart Parking System (SPS) on parking space management. An SPS not only tracks the number and locations of available parking spaces in an area but also utilizes Machine-to-Machine communications (M2M) to provide drivers useful information about the status of a parking area. We create three sets of experiments to study these problems and use the satisfaction rate and the occupancy rate to evaluate the performance. Based on the results obtained from our simulation, we are able to identify the best policy among all alternatives which achieves the highest satisfaction rate under the constraint of maintaining certain occupancy rate. Furthermore, our simulation shows that the SPS can improve the satisfaction rate regardless scenarios under different assumptions.