A continuous-time Markov compartment model of freeway traffic flow is presented and tested using simulated and real data. By using the method of large population approximation, the underlying stochastic process is approximated by the sum of a nonlinear deterministic process and a linear, time-varying Gaussian stochastic process. With this approximation a Kalman filter that tracks the density of a freeway section, broken down by destination, was derived. The filter was then tested using simulated data and actual freeway data obtained from Interstate 35W.
|Original language||English (US)|
|Number of pages||6|
|Journal||Transportation Research Record|
|State||Published - Dec 1 1994|