快速路入口匝道的的非参数自适应迭代学习控制

Non-parameter adaptive iterative learning control for the freeway traffic ramp metering

基于快速路交通系统重复性和周期性的特征,引入"拟伪偏导数"概念,给出了宏观交通流模型沿迭代轴的非参数动态线性化形式.进一步,提出了快速路入口匝道的非参数自适应迭代学习控制(NP-AILC)方案.该控制方法本质上是无模型的,并且学习增益可迭代调节.收敛性分析表明当系统初始状态随迭代次数随机变化时,该方法可实现几乎完全跟踪性能.仿真结果进一步验证了方法的有效性.

Based on the repeatability and periodicity of the freeway traffic system, a non-parameter dynamic linearization of the macroscopic traffic flow model is developed by introducing the concept of ＂Mimic Pseudo Partial Derivative＂ . And then, a new non-parameter adaptive iterative learning control （NP-AILC） is presented for the freeway traffic ramp metering. This control approach is model-free in nature, and its learning gain can be adjusted iteratively. Convergence analysis shows that this approach can achieve an almost perfect tracking performance when the initial states are randomly varying iteratively. Simulation results further illustrate the validity of the presented method.