| Summary: | 博士 === 國立成功大學 === 交通管理學系 === 88 === System validation is the last key procedure of system development, also the required stage to ensure the reality of a simulation system. However, the validation result is often very hard to justify as well as considering both the statistical objectivity and the data of random nature. In this study, therefore, "the Validation Method with Human Behavior Model" is proposed to solve effectively the difficulty of traditional validation methods, which fail to consider the statistical and practical side simultaneously.
The feasibility of this validation method is verified step by step with the example of developing the fixed based driving simulator. First one is to validate the simplified driving simulation device by comparing to the field test results within the car following model-related literatures. By collecting the real world data and establishing the lane changing decision model, the fixed based driving simulator is then developed and validated.
In addition to propose a new system validation concept, there are still other issues being discussed in this study,
1. the usability of simplified driving simulation device
2. the method to collect data and to make a microscopic lane change experiment design effectively
3. the method to construct and test the fixed based driving simulator as well as to collect the simulation data
4. the construction method for the optimal simulation environment and the proper simulation procedure
5. the results of the Validation Method with Human Behavior Model
6. the application find outs of the fixed based driving simulator
The results of each item found above are very close to the field experimental data. It indicates that the Validation Method with Human Behavior Model is applicable to other simulation system. Further all products generated in this study can be applied to other studies as well.
1. the validation procedure design for simplified driving simulation device and fixed based driving simulator for car following behavior and lane changing behavior.
2. high accuracy of field test car trajectory database, the data collection method.
3. the fixed based driving simulator, including five frequent used car manipulated interfaces, throttle, brake, clutch, driving wheel and gear, the interface between driving simulator and the image generation computer, the interface between driving simulator and data monitoring and collecting computer, the virtual car progression method.
4. the 3D computer graphics object library with a real world picture enhanced for Tainan Chun Whua bridge, driving situation library for lane changing on the urban overpass, the design and implement for image geometric calibration procedure of simulation experiment
5. 8 kind of lane changing simulation scenarios
6. the car trajectory databases for 645 different driving simulations, including 8 different test drivers, 8 different lane changing scenarios and 4 simulation items, the 13 data analysis VBA macros.
7. the analysis methods for driving simulator adaptation, simulation content effeteness, lane chnaging trajectory, lane changing influence factors.
|
|---|
