Accuracy and precision evaluation of two low-cost RTK global navigation satellite systems

Two low-cost Real-time Kinematic Global Navigation Satellite Systems (RTK GNSSs) being the Emlid “Reach RTK” and the NavSpark “NS-HP” were evaluated in terms of positioning accuracy and precision. Each GNSSs’ rover unit was mounted on a field robot that travelled by manual remote control along a pre-defined test track in six repeated trials. The precision of the two systems was evaluated through F-test statistics. The combined accuracy of the two GNSSs was determined by comparing the positioning data to a fixed known distance between the GNSS antennas on the robot (472 mm). In three out of six trials, both GNSSs remained in fixed solution status, and showed a Root Mean Square Error (RMSE) of less than 50 mm, which was within the expected range. In two other trials, one of the GNSSs started in float solution status, and subsequently transitioned to fixed solution status. In these trials, the RMSE was still well within one meter, which was expected in float solution status. In one trial, a false fixed position status was encountered, where the NavSpark GNSS falsely claimed it was in fixed solution status. This issue needs to be alleviated in the future by improvements in signal conditioning, noise, and software, and/or by sensor fusion. Although the Emlid GNSS had superior localization performance, as its percentage of data in fixed solution status was 94.0% compared to 71.5% for the NavSpark GNSS, both were deemed promising for use on experimental field robots.