针对交通流模型的强非线性、不确定性等特点,提出了基于近似动态规划的交通流模型参数辨识算法.该算法具有自学习和自适应的特性,不依赖于被控对象的解析模型.严格的理论推导证明了这种参数辨识方案的收敛性,仿真结果验证了所提出算法的有效性. In view of the strong nonlinearity and uncertainty of traffic flow model, a parameter identification algorithm of traffic flow model based on approximate dynamic programming is proposed, which has the characteristics of self-learning and self-adaptation and does not depend on the analytic model of controlled object. Strict theoretical derivation proves the convergence of this parameter identification scheme, and the simulation results verify the effectiveness of the proposed algorithm.