E-Mail Article     Request for image use     print page   

Speed Validation Project

Diamond Consulting Services Ltd (DCS) has developed the Idris patented vehicle detection system which, over the last ten years, has become the market leader for vehicle classification in axle-based tolling applications. More than 1500 lanes of Idris have been deployed across North America and Europe.

The core Idris product is a vehicle detection system based entirely on inductive loop sensors. As such, it is immune to weather variations and has a low maintenance requirement.

Speed and traffic light cameras have been rapidly deployed in many countries; however, next generation systems are being developed to incorporate 'speed on green light' and 'variable speed limit according to vehicle type'. Idris is ideally suited to combine with cameras, since it detects not only vehicle type but also vehicle speed, with enforceable accuracy and the ability to trigger a camera.

DCS has partnered with manufacturers of 'speed' and 'red light' camera systems, and has undergone independent verification of its speed detection performance to support qualification of the resulting products. Helping technology partners to progress into new areas with the security of knowing the technology incorporated in their systems meets specification and accuracy requirements is paramount to the success of DCS.

Results for Idris Speed Validation
All data accrued during tests were recorded on site by a TUV witness. These provide a physical and non-amendable record of all information. This data is stored in archive at TUV Product Service premises.

All tests on the Idris Speed Enforcement System indicated the system measurement accuracies were better than the specified Idris speed accuracy statement of:

+/- 2kph for speeds below 100kph (62mph).
and +/-2% for speeds over 100kph (62mph).

In addition the measurement accuracies were well within the tolerance requirements of the HOSDB Publication No 15/05, listed as:

Positive error not greater than 2mph or 3%, (over 66mph)
(Positive error not greater than 1.2kph or 3%, (over 41kph))

Negative error no greater than 5mph or 10%, (above 66mph)
(Negative error no greater than 3.1kph or 10%, (above 41kph))

Objectives
Track-based testing was undertaken to validate the speed accuracy and performance of the Idris Speed Enforcement system. The tests were carried out on two different roadside equipment platforms, from Nortech International and Advanced Traffic Systems (detailed later), each running the Idris Speed Detection algorithms.

The key initial requirements for validation were to:

• Equip a test track with in-ground inductive loops, suitable equipment housing for Idris hardware and an independent speed reference measurement system;
• Appoint an organisation with relevant experience and reputation to devise and oversee the testing and to produce the report, results and certification;
• Employ test vehicles and professional drivers capable of driving to the required standard and speeds.

Testing Parameters
The validation testing had to take into account:

• The current quoted Idris speed accuracy of +/- 2kph (1.2mph)for speeds below 100kph (66mph)and +/-2% for speeds over 100kph (66mph);
• Testing accuracy up to the maximum possible speed on the test track for cars, ideally over 200kph (125mph);
• The inclusion of an independent and sufficiently accurate reference speed measurement system to provide the 'truth' data;
• A requirement to test to a recogised international standard.

Validation Setup & Requirements
Loops: Loops were required to cover two lanes to allow testing of vehicles straddling two lanes. DCS used the expertise of Crown Cutting Services for the installation with an Idris Engineer to oversee the installation and set up. The loops had to satisfy the product and system specification for Idris.

Roadside equipment: Roadside equipment to analyse the loop data was run from batteries and was provided by the manufacturers and installed by DCS.

Independent speed assessment reference: An independent speed reference device was used and locked to an accurate clock to allow results from both reference and Idris to be aligned for comparison and validation. The products used to perform the independent assessment were supplied by Racelogic Ltd.

Test Specification: Testing was based on the UK's Home Office approval specification of HOSDB Publication No 15/05.

Test Track Speeds: Issues relating to the higher speed requirements had to be addressed and the appropriate actions put into place. This included vehicle type, driver qualifications and insurance. Bruntingthorpe Proving Grounds were ed to host the validation testing.  The company was able to provide a suitable test track, vehicles, driver and insurance.

Time Synchronisation: Accurate time stamping of Idris and the reference systems were a major necessity.

Validation Company Qualifications: The validation company used had to provide evidence of its standing in the industry as an independent verification and certification centre. TUV Product Service Ltd met the certification criteria and was chosen to witness, validate and produce the final report for the Idris Speed enforcement system.  TUV has overseen the validation of other Speed Measurement systems to the Home Office standard.

Independent Speed Reference Equipment for Validating Idris:
VBox Mini & VBox II Units supplied by Racelogic
The purpose of the VBox units was to provide the 'ground truth' speed information to which the Idris speed output was compared.
The units are mounted on the test vehicle and use the signals provided by GPS satellites to monitor its position, much like a normal Sat-nav unit, but to a much higher level of precision. Knowing the vehicle's position and its change over time allows the units to calculate speed (change of position over time), acceleration (change of speed over time) and from them, many other application-specific values such as lateral movement, for instance.

The units log all this information to a file on an SD memory card 10 times per second. Post-processing software (an application called VBoxTools) allows the data to be displayed in both graphical and text formats.

Setup
The units are installed in the vehicle and are usually powered by its cigarette-lighter adapter, although batteries can be used for vehicles such as motorcycles. A GPS antenna is magnetically attached to the vehicle's roof and this becomes the reference position on the vehicle (this means the  box can be moved around within the vehicle without affecting results).

Spatial Trigger
The VBoxes can  use a light beam accessory to provide a trigger signal when the vehicle crossed the Idris loops. This consists of a light emitter and sensor which are mounted externally on the vehicle and a reflector which is placed at the desired position by the roadside. As the vehicle goes past the reflector, the sensor picks up the reflected light and generates a trigger to the VBox. This can then be used as a reference to identify which speed sample provided by the VBox is the one on or closest to the loops.

Start, finish and splits
The VBoxTools application allows the user to define trigger positions which provide much the same functionality as the light barrier (one could think of them as 'virtual' light barriers). These coordinates can be saved in a file and used as trigger points in the post-processing of the data. This function was also used during our validation testing.

The following layout was defined for the trigger signals with respect to the two Idris loop sites (2x2m loops and 1x2m loops):

• the light barrier was placed exactly half way between the two sites;
• a Start Point was defined 20.13m before the light barrier;
• Split 1 was defined as close as possible to half way between the two pairs of 2x2m loops, 5.08m before the light barrier, thus providing the speed over the first loop site;
• Split 2 was defined as close as possible to half way between the two pairs of 1x2m loops, 3.44m after the light barrier, thus providing the speed over the second loop site;
• a Finish Point was defined 16.9m after the light barrier:
• the Start and Finish Points were used to separate the vehicle passes into 'runs'.

Lights, camera, action!
The speed tests consisted of 10 runs at each speed and were defined as 32kph (20mph) and every 16kph (10mph)  up to 225kph (140mph).

The VBox Mini had an advantage over the VBoxII for the driver in that it could be fixed to the screen as close to the driver's line of vision as convenient. This allowed the driver to monitor his speed and position the vehicle correctly at the same time.

VBoxTools has an online mode in which signals from the VBox are displayed in real-time. This was very useful for calibrating the light barrier, which consisted of matching the height and orientation of the emitter/sensor and the reflector so that a trigger would be generated for both test lanes.

It should be noted that it is not imperative for the driver to be at the correct or constant speed for the runs: we were interested in the (possible) difference between the Idris output and the VBox output. The speeds chosen were merely to group the runs.

On most runs the vehicle's cruise control was used to obtain a constant and repeatable speed. The exceptions were the 32kph (20 mph) runs, where cruise controls don't activate, and the runs at and above 190kph (120mph) because the Ferrari used didn't have cruise control!
The VBox Mini was used by the driver to set his speed correctly and the VBox II data sets were used for the results.

Validation Testing July 2009

1 Objectives

This part of the document describes the equipment setup for the Idris Speed Enforcement System validation process carried out at the Bruntingthorpe Proving Ground, Leicestershire, UK on 14th and 15th  July 2009.  A section of the 2 mile straight runway was used for the test track.

2 Equipment Setup

2.1 Idris Equipment
The tests were carried out on two different hardware platforms, both running the Idris speed detection algorithms:

Figures 1 and 2 show the equipment.

Figure 1: Nortech equipment



Figure 2: ATS equipment

3 Test Equipment

In accordance with the VBox manufacturers requirements the VBox II and VBox Mini were fitted into the test vehicles. The accuracy quoted by the manufacturers for the VBoxII is + or – 0.1km/h, while the Vbox Mini is + or – 0.2km/h. Both are an order better than the Idris specification.

The speed output from the Idris platforms was compared to measurements collected in the test vehicles by the two Racelogic GPS units:

• VBox II SX10Hz GPS datalogger equipped with a Light Barrier sensor, part number RLVBACS063, to provide the position reference.
• VBM01 (VBox Mini) 10Hz GPS datalogger.
  
  Testing was performed in accordance with the requirements of the HOSDB Speedmeter Handbook (Fourth Edition) 15/05, section 7.6 (speed accuracy) and 8.2 (speed measurement).
  
  For each test run the test vehicle made an attempt to maintain a constant speed throughout the length of the enforcement loop baseline.
  
  It is accepted that a vehicle traveling at constant speed through the entry/exit datum points is not necessarily representative of the manner in which actual drivers on the road would approach a speed enforcement system.

4 Loop Layout

Two sets of loop sites were installed, one with 2m loops and one with 1m loops (measured in the direction of travel). Both sets were installed in two lanes. The widths of the loops were slightly less than 2m and were chosen to suit the test conditions. The “as-cut” dimensions of the loops embedded in the test track are given in Figure 4.

Instead of defining two separate sites with two lanes each, it was deemed simpler from the data presentation standpoint to configure the first site (loops 1 to 4) as lanes 1 and 2 and the second site (loops 5 to 8) as lanes 3 and 4.

Each vehicle pass yielded two speed measurements, in either lanes 1 and 3 or 2 and 4, depending on the physical lane that was used. This configuration was used during testing of the Nortech system on  14th July 2009.  Most of the tests were done in the physical lane 2, as shown in Figure 3 below.

A simpler approach was taken on the second day for the ATS system, which is shown in Figure 5: only loops 3, 4, 7 and 8 were used and the lanes were defined as lane 3 and lane 4.

Figure 3: Main physical test lane

5  Position References

Two techniques were used to correlate the VBox data to the Idris measurements:

• The light barrier provided a VBox II trigger every time the vehicle passed the reflector at the side of the road.  This reflector was placed at the mid point between the two loop sets.
• The VBoxTools software allows for the definition of split points which can be considered a “virtual” equivalent of the light barrier triggers, i.e. when the vehicle crossed the split points an event was generated.
  
  Two split points were defined, one at each loop site centre. This enabled reports to be generated which provided the speed of the vehicle at both the split points and at the light barrier points, so the vehicle speed is measured as accurately as possible with respect to the Idris loops.
  
  It was observed that the light barrier was actually not as good a reference as the definitive split points. It proved to be unreliable in heavy rain. For this reason, test results for some vehicle runs had three test results (split 1, light barrier and split 2) and some only had two (split 1 and split 2).
  
  The splits, however, once defined and saved in a file, were available for the VBox II data analysis.  As the measurements were taken in the centre of the loop sites, they were more accurate than the light barrier triggered measurements which were taken between the two sites.

Figure 4: Loop site dimensions and lane definitions used for the Nortech system on 14-07-09 (all dimensions in cm) 

Figure 5: Loop site dimensions and lane definitions used for the ATS system on 15-07-09 (all dimensions in cm)

For further information, please click here

---

Bookmark and share