The network state estimation module - the start point for a traffic prediction and guidance generation - is used to obtain the best knowledge on the current network state. The network state to be estimated includes traffic flows, densities, speeds, travel times, and incident occurrences along with their locations and severities. These state variables can be estimated based on data obtained from the surveillance system module. For example, based on the measurements from the sensors, statistical methods may be used in estimating the traffic variables of interest. Alternatively, traffic simulation models may be used instead of or as a complement to the statistical methods. The advantage of the simulation-based approach is that it can provide network-wide estimates of the travel times on various paths. It may also have less dependency on the accuracy and completeness of the surveillance data. This is an important consideration because not all links and intersections are equipped with sensors and some sensors may not be operational. The mesoscopic traffic simulation model presented in Chapter 5 can be used in the laboratory for this purpose.
In this research, the default network state estimation module is treated as a dummy element that provides ``perfect'' estimation. In other words, we assume that the true network state - the state observed in MITSIM - is directly available to TMS. An estimation error can be easily added to the true network state to test a control and route guidance system's sensitivity to the quality of the network state estimation.