23.10.2013
Traffic-monitoring systems (TMS) made by the VIDIT company and approved by PTB are used for video-based distance and speed measurements of vehicles. Within the scope of a joint cooperation with the approval holder and the DTC company, the measurement error of the traffic-monitoring system has, for the first time, been determined under the conditions of real driving manoeuvres.
The computer-aided traffic-monitoring system is based on the single-image analysis of a digitalized video recording of a traffic situation. The distance measurement is based on acquiring the positions of the two vehicles recorded in the video image by manual focusing with a measuring line. With the aid of marks placed on the road (so-called "ground control points" and "monitoring points") and by means of an individual file for each measurement point, it is possible to use a mathematical procedure to determine, from the positions of the measurement line, the distance between the two vehicles. The time information required to measure the vehicles' speed is determined automatically by the traffic-monitoring system with the aid of the two traffic situations at a fixed video image rate (40 ms).
The tests for the approval of speed-monitoring devices are carried out by means of comparison measurements in real traffic situations, using high-precision PTB reference facilities. In the case of the traffic-monitoring system, in contrast, the trueness of the distance measurement was verified with the aid of fixed reference points – a procedure which experts will easily understand. To this day, there had, however, been a lack of a universal demonstrative testing method which could be easily understood by the broad public. For this reason, the measurement error has now been analyzed in detail also under the conditions of real driving manoeuvres carried out by two test vehicles with deceleration and acceleration phases. Each of the test vehicles used was equipped with high-quality DGPS inertial systems manufactured by Oxford Technical Solutions Ltd. (see Fig. 1) which allowed the value of the distance between the two vehicles, which was used as a reference, to be determined with a measurement uncertainty of 3 cm. In addition, the position of the rear vehicle was determined as a function of the front traffic-monitoring system's ground control point with a relative uncertainty of 2 cm. This later allowed the reference distances (10 ms measuring rate) to be temporally synchronized with the distance data determined by the traffic-monitoring system.
Figure 1: Determination of the reference distance between the two test vehicles used (relative position) and of the distance to the front traffic-monitoring system's ground control point (absolute position).
A comparison with the reference curve plotted in dark blue in Figure 2 shows that the vehicle distances determined using the traffic-monitoring system are, in the whole measuring range (vehicle position up to 200 m away from the front ground control point), in favour of the monitored vehicle. Since the traffic-monitoring system systematically enlarges the measuring lines focused on manually by one pixel, the deviations for the positions of vehicles located further away are, as expected, larger than in the close-up range. Figure 3 shows the enlargement of a detail of the close-up range framed in red in Figure 2. Also in this case, the distances measured by the traffic-monitoring system are always and without exception in favour of the monitored vehicle. The light blue curve shows the physical distance value which results from the distances of the bumpers of the vehicles involved, i.e. taking the vehicles' overhang into account.
Figure 2: Comparison of the distances between two vehicles as measured by the traffic-monitoring system with the reference data provided by the DGPS inertial systems used (representation of the full measuring range of the traffic-monitoring system).
Figure 3: Comparison of the distances between two vehicles as measured by the traffic-monitoring system with the reference data provided by the DGPS inertial systems used (representation in the close-up range of the traffic-monitoring system).
All in all, the comparison measurements, which were carried out with great metrological effort, clearly confirmed that the distance values provided by the traffic-monitoring system always turned out to be generously in favour of the monitored vehicle, even taking further influence factors (e.g. statistic variations in multiple analyses) into account.
Contact person:
Frank Märtens, Dept. 1.3, WG 1.31, E-Mail: frank.maertens@ptb.de