GPS World, September 2010

GOVERNMENT Emergency Response ÅFIGURE 8 Absolute error for IMU only and collaborative navigation solutions of C1 GPS outage ÅFIGURE 9 Absolute errors in collaborative navigation solutions of C1 local sensor measurements The collaborative navigation solution is formed by integrating the inter nodal range measurements to other nodes in a decentralized Kalman filter which is referred to as loose coupling of inter nodal range measurements The test results of different scenarios are listed in TABLE 3 For cases labeled 30 sec GPS outage the GPS outage is assumed at all nodes that are equipped with GPS The results listed in Table 3 indicate a clear advantage of collaborative navigation for nodes with tactical and consumer grade IMUs particularly during GPS outages When GPS is available see for example node A1 the individual and collaborative solutions are of comparable accuracy Scenario Min Max Mean Std No GPS outage 001 086 025 017 Outage on A3 003 098 027 018 Outage on A2 and A3 002 172 053 038 ÅTABLE 4 Statistics of collaborative navigation solution for C1 131 600 second m The next experiment used tight coupling of inter nodal range measurements at each nodes EKF in order to calibrate observable IMU errors even during GPS outages In addition varying numbers of master nodes are considered in this example The tested data set was 600 seconds long with repeated simulated 60 second GPS gaps separated by 10 second periods of signal availability The inter nodal ranges were 20 meters TABLE 4 and FIGURE 8 summarize the navigation solution errors for collaborative solution of node C1 equipped with consumer grade IMU only supported by varying quality other nodes The error of the individual solution for this node in the end of the 600 second period reach nearly 250 kilometers 2D Even for the case with a single anchor node A1 the accuracy of the 2D solution is always better than 2 meters Another 900 second experimental data with repeated GPS 60 second gaps on B1 node was analyzed with inter nodal ranging up to 150 meters TABLE 5 summarize the results for C1 node Future Work Collaborative navigation in decentralized loose integration mode improves the accuracy of a user with consumer grade IMU from several hundreds of meters 2D to 16 m max for a 30 s GPS gap depending on the number of inter nodal ranges and availability of GPS on other nodes For a platform with GPS and consumer grade IMU node B1 the improvement is from a few tens of meters to below 10 m Better results were obtained when tight integration mode was applied that is inter nodal range measurements were included directly in each EKF that handles measurement data collected by each individual node architecture shown in Figure 4 For repeated 60 second GPS gaps separated by 10 second signal availability collaborative navigation maintains the accuracy at 1 2 meter level for entire 600 s tested for nodes C1 and B1 Scenario See Challenged Positions page 56 Horizontal m Height m Min Max Mean Std Min Max Mean Std No GPS outage 000 614 122 093 000 1825 314 367 Outage on A3 002 1179 150 131 000 1877 358 387 Outage on A2 and A3 001 1559 217 212 000 3525 454 505 Loss of A3 B1 ranging used 001 1145 178 167 000 3885 424 570 ÅTABLE 5 Statistics of collaborative navigation solution for C1 GPS World September 2010 www gpsworld com 42