# 交通資料分析 Mid-term 0856040 王立洋 ## 1,2. ![](https://i.imgur.com/eoDSFN6.png) ## 3 為了確保量測出平均速度的準確性,追蹤車輛的準確性是需要的,必須要能確定通過的是一致的車輛。penetration rate be set to 2-4% ## 4 ![](https://i.imgur.com/wANRgpN.jpg) ## 5 ![](https://i.imgur.com/57MrzLD.png) ## 6 ### US8972171B1 / US9644981B1 - 根據大量路段的車流旅行時間數據,算出一個地點到達另一個地點需要的最小時間,以及預估最快可以到達的路徑。 #### US9644981B1 For each route segment of a particular route, the server may estimate a travel time. The travel time may be based on estimated vehicle volume data generated from information received from other vehicles. #### US8972171B1 For each route segment of a particular route, the server may estimate a travel time. The travel time may be based on estimated vehicle volume data generated from information received from **other users vehicle Based on the estimated travel time for each route segment of a particular route** ### US9719792B2 / US9709411B2 #### US9709411B2 **PRIORITY CLAIM The present application is a continuation of U.S. application Ser. No. 14/635,685 having a filing date of Mar. 2, 2015, which is a continuation of and claims priority to U.S. application Ser. No. 12/757,178 having a filing date of Apr. 9, 2010 and issued as U.S. Pat. No. 8,972,171. Applicants claim priority to and benefit of all such applications and incorporate all such applications herein by reference.** - removed typos and useless words ## 7 1. 統計每一路口每一時闁區間,探偵車C=通過的次數-停車的次數 2. C值是否呈周期性變化? 3. 若2為是,則將時闁序列C值依考慮的週期長度摺疊後加總,加總後C值有最大的變異量,為正確的週期長度 4. 根據探偵車停車位置與其在正確週期內的停車時間點的闗係圖,估計紅燈週期起始時間點。 5. 根據探偵車停車位置與其在正確週期內的通過路口時間點的關係圖,估計綠燈的週期起始時間點,紀錄該路口交通唬誌之週期與相位 ### problems? - 可能會有車輛誤停情形,造成估計的誤差 ## 8 ![](https://i.imgur.com/LvU9D9d.jpg) ## 9 - 從CMS中提取出 Datetime, HW, direction, Event(事故, 施工, 故障車, 散落物, 壅塞), km0, km1, lanes(內中外)(12一二), oth_lane(路肩, 分隔島, 邊坡), closed(占用,封閉), ramp(出入口匝道)等十個attributes,因為這些attributes的對每個路段的影響程度可能不同,在不確定影響程度的情況下,我們可以分開路段訓練,並在訓練時固定好這十種inputs,給予他們不同的權重,在訓練途中,配合ETC資料、相關的VD資料,訓練這些權重大小,權重大的就代表這個attribute對此路段影響程度較大。