GPS World, March 2018

SIMULATION one of the buildings that was simulated as a simple concrete box was more complex in the real environment Therefore we applied some modifications to scene as in FIGURE 9 After those changes a general improvement in the results was visible but most importantly the missing satellites could finally be tracked by the receiver FIGURE 10 SAN JOSE DYNAMIC TEST RESULTS Similar results were obtained with the dynamic test in San Jose FIGURE 11 shows the results obtained for satellites PRN12 and PRN24 The walking trajectory included two points where the antenna was stopped because of a traffic light Those points correspond to a relatively flat C N 0 that can be clearly seen in the field test and simulation data for both PRNs When instead the antenna was moving a higher variation in the C N 0 is noticeable in both simulation and field test FIGURE 12a illustrates the positioning error obtained from simulated red and field test blue The first part of the simulation produced an error smaller than the one obtained from field data However from the time 1948 a good agreement can be seen The satellite availability is also shown in FIGURE 12b This last result was obtained with the improved model described in Figure 9 CONCLUSIONS AND FUTURE WORK A new real time system for multipath simulation is designed to generate realistic multipath that depends on time position and type of urban environment The 3D scene is used to calculate the multipath reflection and diffraction caused by the buildings and objects around the antenna Some first results demonstrated that realistic multipath can be generated by simulating reflections and diffractions even with a simple 3D model However the inclusion of finer details in the model can improve the simulation and make it even closer to reality As always simulation interest is a tradeoff between reliability in all 34 GPS WORLD WWW GPSWORLD COM MARCH 2018 conditions and efforts to adapt that is to specify a generic and simple model The added value of our model consists in its simplicity and its good compliance with field data Ray tracing techniques coupled with geometrical optics and uniform theory of diffraction are efficient and simple methods to simulate the propagation of GNSS signals in complex urban environments Their efficacy is demonstrated by a good agreement between simulation and field measurements Some discrepancies still exist and are due to the limitations of such a model The accuracy of the model is never perfect and as ray tracing is a deterministic method the returned results strongly depend on the FIGURE 10 Satellite availability for field data blue and simulation after scene improvement FIGURE 11 Carrier to noise ratio top pseudorange residual middle and doppler residual bottom for PRN 12 left column and PRN 24 right column FIGURE 12 a Positioning error for field test blue and simulation red b satellite availability for field data blue and simulation red after scene improvement