TABLE 2

Performance Comparison Between ZSM and Conventional SM Based on a 3D Map of San Francisco

AlgorithmCentroid error w.r.t true location (m) [Cross-street, Along-street]Bounds (m) [Cross-street, Along-street]Average computation load across 100 runs (s)
OfflineOnline
Proposed ZSM14 satellites[1.2, 8.1][5.8, 16.2]162.54.4
[4.6, 58.8][9.8, 24.5]
10 random satellites[0.9, 6.8][6.1, 16.4]162.54.2
[4.8, 53.8][9.7, 13.1]
8 random satellites[1.4, 13.1][18.3, 26.8]162.53.0
[3.5, 54.3][7.6, 24.8]
Satellites with elevation > 20°[1.7, 10.2][0.5, 3.3]162.52.4
[1.9, 7.0][8.7, 20.2]
[4.9, 48.5][16.8, 29.9]
Conventional SM (all available 14 satellites)3 m[0.0, 3.0], [0.0, 6.0], [0.0, 9.0][173.2, 214.5]91204.02.8
[0.0, 12.0], [3.0, 12.0], [0.0, 48.0]
[3.0, 48.0], [0.0, 51.0], [3.0, 51.0]
[0.0, 54.0], [3.0, 54.0], [3.0, 57.0]
5 m[2.0,10.0][173.1, 215.6]34389.91.2
[2.0, 50.0], [3.0, 50.0], [3.0, 55.0]
10 m[2.0, 10.0][192.1, 221.6]6874.10.3
[2.0, 50.0]
30 m[12.0, 60.0][224.5, 261.8]1433.40.1
  • Note: For proposed ZSM, we performed comparison analysis across different GPS satellite subsets, while for conventional SM we compared the performance for different grid resolutions. The 14 satellites can be seen in Figure 10(c)