Traction Controller Design
1.建立一个Vehicle Model
“We chose a design methodology for the vehicle model that would provide adequate simulation accuracy without being too resource intensive to design.We realized that by starting with a simple point mass model and adding known efficiencies and losses we could come very close to what we observed on track.”
Our model consists of a vehicle body and a drivetrain system. The vehicle body represents the point mass and contains load transfer and aerodynamic calculations. The drivetrain includes our motor, gearing, differential, and tire model. Many of the components require only simple physics to initially get results. Once the basics work, efficiencies, losses, and more advanced modelling can be added as required to get the simulation to match testing data.
The most difficult part of making the vehicle model was integrating the tire model. We ended up using MFeval, which is a free library available on the MATLAB file exchange, to do this. MFeval contains the necessary equations to evaluate magic formula models from TIR files.
Finally, we needed to validate the model so that it could reliably be used to tune traction control. We used testing data from past years to determine if the acceleration, wheel speeds, and slip response from our model matched a real acceleration test. We fed the model the logged throttle input and analyzed how its output matched the logged vehicle response. After some tuning and adjustment of correction values, we had a model which approximated our car’s longitudinal behavior.
核心思想:由简到繁,先建立点质量模型。要根据实际数据确定车辆模型的速度,加速度和打滑率与真实情况相符合,反复对比数据确定纵向模型的校正系数以使之准确
2.建立 Traction Controller
主要由两部分组成:
组成模块 | 作用 |
---|---|
纵向牵引力控制模块 | 计算需求扭矩以及滑移率导致减少的扭矩值 |
车辆状态解析模块 | 计算轴速、滑移率 |
3.模块连接仿真
可以根据simlink驾驶员模型demo更改