Dynamic traffic flow models are essential for obtaining information about the time evolution of variables describing the traffic flow phenomena and have a critical role in the development and implementation of real-time applications such as Intelligent Transportation Systems. Macroscopic traffic flow models that treat the traffic as a continuum are preferred for such applications. But, existing macroscopic models characterize homogeneous traffic, and may not be directly applicable to capture the vehicle heterogeneity seen on Indian roads. To address this issue, a non-continuum macroscopic dynamic traffic flow model based on the lumped-parameter approach was developed in this study. The model was developed based on the conservation of vehicles equation and a dynamic speed equation, incorporating an empirically developed traffic stream model, which is an important contribution of this study. Using this model, an estimation scheme has been developed based on the Kalman filtering technique to estimate traffic states in real time. The proposed scheme was implemented and corroborated for the heterogeneous traffic conditions existing in India. The performance of this scheme has been evaluated and the results obtained have been found to be promising.