基于模型选择和优化技术的自动驾驶制动模型辨识研究

本文结合城市轨道交通列车自动驾驶系统实测数据,对电动客车制动模型的类型和参数的辨识进行较为深入的研究。首先,在分析最大制动率条件下实测制动数据后,对列车制动模型类型进行研究。然后,应用非线性优化技术在不同模型种类、有无参数范围约束等情况下,对制动模型进行辨识,并根据模型选择技术获得最佳制动模型及其参数。最后,研究不同制动率下列车制动模型的一致性,并基于模型选择技术给出制动率和制动减速度之间的模型。

Study on ATO Braking Model Identification Based on Model Selection and Optimization Techniques

This paper studied system identification of the electric train braking model with the field data from the ATO(Automatic Train Operation) system in urban rail transit.Firstly,after analyzing the field data under the maximal braking current,some possible types of models are proposed.Then,nonlinear optimization techniques are used to identify the model parameters of different models with or without constrained conditions,and the best braking model and its parameters are obtained by the model selection technique and expert know-ledge.In the end,the consistence of train braking models under different braking rates are studied and the best model describing the relationship between the braking rate and deceleration is chosen by the model selection technique.