GPS based high accuracy speed meter for vehicles is a special type of GPS speed meter which uses Doppler Demodulation of GPS signals to calculate the speed of a moving target. It is increasingly used as reference equipment in the field of traffic speed measurement, but acknowledged standard calibration methods are still lacking. To solve this problem, this paper presents the set-ups of simulated calibration, field test signal replay calibration, and in-field test comparison with an optical sensor based non-contact speed meter. All the experiments were carried out on particular speed values in the range of (40-180) km/h with the same GPS speed meter. The speed measurement errors of simulated calibration fall in the range of ±0.1 km/h or ±0.1%, with uncertainties smaller than 0.02% (<i>k</i>=2). The errors of replay calibration fall in the range of ±0.1% with uncertainties smaller than 0.10% (<i>k</i>=2). The calibration results justify the effectiveness of the two methods. The relative deviations of the GPS speed meter from the optical sensor based noncontact speed meter fall in the range of ±0.3%, which validates the use of GPS speed meter as reference instruments. The results of this research can provide technical basis for the establishment of internationally standard calibration methods of GPS speed meters, and thus ensures the legal status of GPS speed meters as reference equipment in the field of traffic speed metrology.
A field experimental standard for traffic speed measurement was set up at No.G92 Expressway of China from Hangzhou to Shanghai for field tests of vehicle speed-measuring devices in actual traffic, and it met the requirements of the recommendation in OIML R 91 about the metrological field tests of pattern approval. This paper firstly introduces the speed measurement principle of the standard equipment and evaluates the uncertainty of speed measurement. Secondly, a field test based on this standard equipment is designed and performed on a radar speed-measuring device to evaluate its actual speed measurement performance in actual traffic. Finally, the uncertainty of measurement of field test error in actual traffic is evaluated.