This release note covers version RMGPS_GOTOP_MTK_T20260511 of the VisionAI GPS positioning module for ADPlus2.0 and M1Nv2 devices, which improves positioning timeliness, track accuracy, and responsiveness to speed changes.
This version replaces RMGPS_GOTOP_MTK_T20250505. The ezInstall companion version is Veyes_1.6.3.
Version change information
| Item | New version | Previous version |
|---|---|---|
| Version name | RMGPS_GOTOP_MTK_T20260511 | RMGPS_GOTOP_MTK_T20250505 |
| Models supported | ADPlus2.0, M1Nv2 | ADPlus2.0, M1Nv2 |
| Positioning module version | RMGPS_GOTOP_MTK_T20260511 | RMGPS_GOTOP_MTK_T20250505 |
| ezInstall version | Veyes_1.6.3 | Not applicable |
What's improved in this version
This version includes GPS performance optimization and improved filtering.
GPS algorithms are enhanced in the following areas:
-
Positioning timeliness Time-To-First-Fix (TTFF) is faster for both cold and hot starts.
-
Position and track accuracy Dead-reckoning algorithms are enhanced to reduce positional errors in tunnels and drift distance and mileage errors. This also improves accuracy when GPS lock is lost and then regained.
-
Responsiveness to changes in speed Previous GPS filtering did not respond quickly enough to changes in vehicle speed. This update improves speed accuracy and harsh-driving alert accuracy, including the three-axis acceleration match rate and the alert event match rate.
Performance evaluation
Stage 1: Lab and road testing
The test covered roughly 200 km across urban and rural areas, including 14 static scenarios (parked for at least 10 minutes each) and 22 dynamic scenarios: driving between tall buildings, mountain passes, and through a tunnel.
The evaluation compared positioning timeliness (cold and hot start performance), track accuracy (dead-reckoning error in tunnels and between high-rise buildings), speed timeliness (speed accuracy and delta speed), and harsh-driving alert accuracy (three-axis acceleration performance) against a stand-alone reference GPS device.
Most results are judged relative to the previous version rather than against an absolute target. The fixed acceptance criteria are dead-reckoning error under 3%, mileage error under 1%, cold start of 90 seconds or less, hot start of 5 seconds or less, and speed accuracy within ±5 km/h at a 95% or higher match rate.
Result: Pass. The new version outperforms the previous version on tunnel dead reckoning (1.32% average error versus 1.55%), speed match rate, speed behavior entering and exiting tunnels, and signal suppression during garage loading and unloading. Harsh-driving alerts and acceleration match rates also meet the acceptance criteria.
Stage 2: Field trials
Field trials are ongoing with approximately 80 devices upgraded.
No adverse performance has been observed in field trials to date.