VS-pCoq ZSS

VS-pCoq Detector Count Server 

Automated statistics and quality control

The Detector Count Server calculates automated statistics and transforms them into forms. It works at night, so that the forms are ready at the start of business. Besides statistical forms it can also create forms for traffic and control quality evaluations. The forms can be accessed by the file system or via a document management system.

Process Steps

  • Calculation of statistical interval data:
    • Detector count data forms the base.
    • Data can be weighted and corrected.
    • Quality evaluations are added
  • Plausibility checking of the statistical data by using rules and reference curves.
  • Replacing of implausible data. Therefore there is always data for all intervals.
  • Automatic calculation of daily, weekly, monthly and yearly statistics. A plausibility value is generated for all measured values.
  • Automatic form creation on the results. Excel templates are used as base, PDFs will be generated as a result.
  • Integrated form management: you will quickly find the desired evaluation for a specific date.
  • Archive for forms and raw data

Form creation

Basically any forms can be generated. As an example, the forms for traffic counters are listed:

  • Per day:
    • Diurnal variation of hourly values ​​and
    • Average hourly values
  • Per week:
    • Weekly variation of hourly values ​​and
    • Average daily traffic
  • Per month:
    • Monthly variation of daily traffic
  • Per year:
    • Annual variation of monthly average daily traffic


Characteristics

The following data sources can be used:

  • Raw data in accordance with OCIT-I PD or OCIT-C PD
  • Interval data, for example according to TLS
  • Other formats, such as the import format of the Swiss Federal Roads Office (FEDRO) or a VPX archive

In a first step, statistical values ​​for short intervals are calculated. Typically they are 3 minutes long.

Then statistical values ​​for long intervals are calculated. Typically they are 1 hour long. During calculation, the long interval values are checked for plausibility on the basis of the short intervals. Other interval lengths are possible. They can be stored in export formats as well. Such export files can be used by other applications.

There are plausibility checks, which are based directly on the detector values, ​​and plausibility checks, which are based on traffic counter values, mostly already aggregated values. For each hour a count value is produced:

  • When data is plausible, the produced value is a genuine value.
  • The data can be marked as "suspect" and replaced by adapted surrounding values.
  • In case of implausible data the produced data is taken from adapted reference values.

Thus for every hour a count value with confidence information is generated.

Plausibility

Zero-value plausibility check

This plausibility check whether zero values ​​(no traffic in short intervals) are allowed. The following limits can be specified per hour by day-dependent reference curves:

  • A maximum number of consecutive zero intervals,
  • An absolute number of zero intervals.

Maximum plausibility check

This check is done as well by reference curves:

  • Short-term: it is checked whether the short interval values ​​do not exceed the physical maximum value (e.g. 1,800 veh / h).
  • Long-term: it is checked whether the long interval values ​​do not exceed the traffic engineering maximum (e.g. 1,200 veh / h).

Percentile Belt plausibility check

Percentile values ​​are a dispersion measure of short interval values while calculating long interval values. The 90% percentile value indicates, for example, under which value 90% of the short interval values ​​can be found. It does not matter how large the values ​​of the remaining 10% are.

Every hourly count is compared with the distribution of percentile reference curve and assigned to a percentile class. If the assigned percentile class exceeds or falls below the assigned percentile class by a given percentage, the value classified as suspicious or even as implausible value.

The multirow concept

Multirows consist of individual detectors. But the detectors are not required to be physically located where the logical multirow is placed. This makes it possible to count traffic in both directions where one will count it and not where the detectors are placed:

Unirow A1 consists of a single detector, which is located in the egress.

Unirows A2, B1 and C1 all consist of two detectors, which are located in the accesses, here at the stop line.

Unirow B2 does not exist physically. The traffic counted by this unirow is composed of the right turners of C1 and the straight through portion of A2.

This straight portion of A2 is composed of

  • the counts of the left lane and
  • a daytime-dependent proportion of straight-through traffic counted on the right lane, to be determined in a reference curve.