In work that was undertaken for Mouchel Essex on behalf of Essex County Council, Minnerva exploited the capabilities of OmniTRANS to develop trip matrices for Chelmsford.

Roadside interview (RSI) data was available for some of the area, but coverage was partial in nature and included multiple observations. Using the OmniTRANS matrix estimation facilities a procedure was established that processed the RSI data to provide a set of observed data for the study area. This accounted for sampling levels, unobserved trip directions, and multiple observations. The processing encompassed the ‘MATVAL’ methodology that merges multiple observations according to the level of sampling (calculated as an Index of Dispersion)

Gaps in the data were filled using a mixture of Gravity type modelling for local trips and information from a regional study (LOIS) matrix for external trips. Trip end information for the local area was calculated from the TRICS database and from car park surveys. The matrices were developed not only by time period and vehicle type, as required for the Paramics traffic simulation modelling being undertaken in Chelmsford, but also by trip purpose. This was a result of also using the matrix estimation procedures of OmniTRANS that allow information on trips to be by vehicle type and trip purpose. For example, the traffic count data used in the estimation was variously supplied in terms of total trips (from automatic traffic counters), by vehicle type (from manual classified counts), and by type and purpose (in the case of RSI sites)

The image on the right, which shows car trips in the form of bandwidths colour-coded by purpose, is therefore significant in supporting not only traffic management modelling but also policy analysis – education trips being of particular interest in this case, as well as multi-modal modelling, for the matrices can also be provided in terms of person trips as well as vehicle trips

The project also involved the use of GIS to enhance the precision of the geocoded origin and destination data in the RSI surveys and subsequent coding to traffic zones.

(Contact: Martin Bach)