Blog

Test Report | SV100 INS Series Performance Field Test Under Diverse Environments


To verify the real-world performance of the SV100 INS Series receivers in various GNSS-challenged environments, SingularXYZ conducted a structured dynamic vehicle test through diverse urban and semi-enclosed scenarios. This test focused on evaluating positioning accuracy, attitude stability, and robustness during GNSS outages.

 

Test Purpose

This field test aimed to evaluate the performance of SV100 INS Series under typical operating conditions for autonomous driving, mapping vehicles, smart transportation, and similar applications. Key objectives include:

  • Verifying positioning accuracy in open and semi-obstructed urban areas

  • Evaluating heading stability and trajectory deviation in elevated and partially covered roads

  • Assessing continuity of navigation during complete GNSS signal loss (e.g. in tunnels and underground garages)

 

 Test Devices

Three units from the SV100 INS Series were tested simultaneously:

No. Device Positioning Solution  Output Frequency
1 SV100 INS-1 (Standard) RTK + Dead Reckoning 10 Hz
2 SV100 INS-2 (Backup Unit) RTK + Dead Reckoning 10 Hz
3 SV100 INS-WI (Enhanced with Wheel Speed Input) RTK + Dead Reckoning + Wheel Speed 10 Hz

 

Test Method & Route

1. Test Methodology

The SV100 INS devices were installed alongside a reference system on a test vehicle according to standard installation requirements. The total driving test lasted over 3 hours, with each typical scene tested for approximately 10 minutes. After data collection, the positioning performance of the SV100 INS was evaluated by comparing it to the post-processed results from the high-precision reference system.

2. Test Route and Scenarios

The driving test route covered a wide variety of representative GNSS environments in the city of Shanghai, including:

  • Open highway environments

  • Urban canyons and dense residential areas

  • Elevated roads and overpasses

  • Tunnels and underground parking garages (not shown in the route map)

Each of these environments was selected to challenge GNSS signal availability and assess the SV100 INS system's performance under typical real-world conditions.

 

Test Results (by Scene)

All positioning results are illustrated in the following sections with corresponding visual graphs.

Explanation of Used Key Accuracy Indicators

 

Metric Definition Interpretation Typical Use Case
RMS Root Mean Square of all position errors Average error magnitude Overall system stability evaluation
CEP68 Radius containing 68% of position points 68% of errors fall within this distance Normal condition accuracy estimate
CEP95 Radius containing 95% of position points 95% of errors fall within this distance Reliability assessment in challenging areas

 

① Open & Semi-Obstructed Environments: (Nanpu Bridge, Residential Area)

These areas offer stable GNSS coverage with partial building occlusion.

 

② Elevated Road Segments: (Outer Ring Expressway & Donglong Elevated Road)

  • Outer Ring (open highway with occasional overpass)

  • Donglong Elevated Road (with a continuous underpass

 

③ GNSS Outage Conditions: (Renmin Tunnel, Underground Garages)

These are complete GNSS-denied environments lasting over 60 seconds per segment.

Summary of Key Findings

  • Centimeter-level precision achieved in open or mildly obstructed urban areas

  • Centimeter to decimeter level accuracy maintained in under-viaduct or elevated scenarios

  • Around 1 m accuracy sustained during full GNSS signal loss of 60 seconds and 2m of 120 seconds

  • Wheel speed input significantly enhances drift suppression during long outages

 

 Conclusion

The SV100 INS Series demonstrates excellent positioning performance in real-world conditions, especially where GNSS signals are intermittently or completely unavailable. Its tightly coupled fusion algorithm, high-frequency output, and support for auxiliary sensors like wheel speed make it a robust and adaptable solution for autonomous systems, vehicle navigation, mobile mapping, and smart infrastructure applications.

 

Post a new comment