Traffic data: Processing data /evaluation of automatic classifiers

DTV-Verkehrsconsult offers you substantial and freely definable analyses of the data of automatic classifiers. Therefore, we have a large supply of technical equipment at our disposal (e.g. modems). In addition we run several software tools to convert the different data standards and to enable an individual and suitable evaluation in a format to suit our customers.

In lots of technical standards and guidelines in the fields of road construction and maintenance as well as the traffic management the Annual Average Daily Traffic (AADT) represents a key value. The exact knowledge about this parameter is required to correctly dimension a road. The traffic values influence the necessary construction works and hence the costs for building, operation and maintenance of the road.

To gather detailed information about traffic characteristics so called automatic classifiers are installed. In Germany these are often induction loops, which are mainly placed on the higher-level road network. These automatic classifiers are supposed to detect the traffic by operating round the clock over the whole year. Alternative methods of data collection to suit the needs of our customers are available.

The AADT value results from the sum of all vehicles of a year divided by the number of days per year. If, for whatever reason, the counting is not complete for the whole year, the AADT will need to be calculated. The reasons for downtimes can be multifaceted. They reach from traffic related reasons (blocking of lanes, deviation of traffic without passing the loop) up to technical reasons (power breakdown, technical error, destruction by accidents or road works). Despite a high reliability of induction loops, counting gaps occur from time to time during the annual measurement.

To be able to calculate the AADT, it is necessary to somehow fill the counting gaps. Therefore special estimation procedures have been developed by us during the past decades and improved until today. DTV-Verkehrsconsult is one of the few companies in Germany, which have specialist experience in this field of work and in the application of those estimation procedures.

Another special challenge lies in the possibility that   the detected values can be incorrect. That is why the estimation of counting gaps is the last step within the data processing. Validity checking of the existing data is undertaken as a preliminary step. Implausible data sets are usually deleted.

We provide transparent documentation during all process steps  by marking changes with a code letter, so it is always visible if a data set was modified or not.
Due to the multitude of reasons for untypical traffic data, a fully automated value substitution has not been established yet. It has been shown that experienced human editors can identify the accuracy of untypical traffic data much more successful. That is why in our company software supported semi-automatic processes are used. Next to our own product DAUZ, which is in use with some of our customers, we also use the program TCM (Traffic Count Management) of the PTV AG.

These two programs are fundamentally different in their handling philosophy. While DAUZ is based on a top-down approach, TCM uses a bottom-up method. That means, that with DAUZ the processing first starts with the entire vehicle volume and afterwards the distribution to smaller time intervals or car types is carried out. This leads to an estimation procedure for the counting gaps with a higher statistical accuracy. With the bottom-up method the plausibility check is carried out separately for the different vehicle types. Here there are comparatively smaller vehicle counts available to correct so changes can have a comparatively larger effect on the overall results.
Of course not only daily or hourly traffic volumes can be edited, but also vehicle distributions, lane occupancy as well as weekly and monthly values within the data validation and correction process.

We are your contact for all questions concerning the processing and evaluation of automatic classifiers.  Whether weekly or monthly, we can analyse your data according to the common standards or your individual requirements. At our customers option we also work on your own servers using your preferred software. We are flexible to ensure our customer´s needs are met.

Contact us! We are looking forward to develop solutions tailored to your needs.