| Project/Area Number |
62460162
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
交通工学・国土計画
|
|---|
| Research Institution | The University of Tokyo |
|---|
Principal Investigator |
KOSHI Masaki Institute of Industrial Science, University of Tokyo, Professor, 生産技術研究所, 教授 (70013109)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
AKAHANE Hirokazu Department of Civil Engineering, Chiba Institute of Technology, Associate Profes, 土木工学科, 助教授 (60184090)
KUWAHARA Masao Institute of Industrial Science, University of Tokyo, Associate Professor, 生産技術研究所, 助教授 (50183322)
|
|---|
| Project Period (FY) |
1987 – 1989
|
| Project Status |
Completed (Fiscal Year 1989)
|
| Budget Amount *help |
¥7,700,000 (Direct Cost: ¥7,700,000)
Fiscal Year 1989: ¥2,100,000 (Direct Cost: ¥2,100,000)
Fiscal Year 1988: ¥1,300,000 (Direct Cost: ¥1,300,000)
Fiscal Year 1987: ¥4,300,000 (Direct Cost: ¥4,300,000)
|
|---|
| Keywords | Bottleneck / Traffic congestion / Car-following theory / Sag / Tunnel congestion / Kalman smoother / 追従モデル / カルマンスムーザ |
|---|
| Research Abstract |
1. The dathering cars which can obtain the data of real car-following behavior and the data was developed. Using these test-cars and analyzing system, we accumulated the data of real car-following behavior on expressways.
2. The S-V relationship in congested flow was found to be convex rather than concave, but it had been believed to be concave.
3. Regarding the comments of the speed-space relationship function as parameters of the mathematical car-following model, the new way of identification of all these parameters was proposed. Suppose the three state variables (acceleration, speed and spacing) of the following-car as representatives of the car-following behavior simulated by the model. Then a parameter optimization algorithm was developed by using Kalman filtering theory, this algorithm can minimize simultaneously the weighted RMSE(Roat Mean Square Error)-values of these three variables between simulated ones by the model and collected data. Using this algorithm, sets of parameters of the model for each congested flow and non-congested flow were identified. Then the comparative study of the three reaction time lags of the model was investigated.
4. From re-examination of the of the car-following behavior (perception, judgment and action), another car-following model based on Fuzzy Theory was developed, and the identification of the parameters was also carried out. Then this model was found to have better fit with collected data than the mathematical model 5. Suppose all drivers behave in same parameters of the model (mathematical one or Fuzzy one), the traffic simulation was carried out, and the ability of reproduction of the real traffic characteristics was invested.
Then by using mathematical model, the simulated traffic became unstable and rear-end collision was frequently occurred. But by using Fuzzy model, it became stable state, so it was showed that this Fuzzy model had some ability of reproduction real stable state traffic.
|
