To address the difficulty of achieving continuous high-precision positioning for in-vehicle Internet of Things systems in urban scenarios and indoor parking lots where global navigation statellite system（GNSS）signals are limited

a multi-sensor fusion positioning solution based on low-cost commercial sensors is proposed.This approach fuses measurement data from on-board motion sensors with 3D point clouds obtained by LiDAR scans and registers them to a high-precision 3D digital map

achieving precise motion-compensated georeferencing and efficient map matching.Real-world road test results demonstrate that the system maintains decimeter-level accuracy in complex environments

with root mean square values of average horizontal and vertical positioning errors of 16.3cm and 14.3cm

respectively.The error is less than 50cm for 99.8% of the time

and less than 30cm for 95.1% of the time.It is confirmed that this system is a practical and reliable solution

which can provide continuous high-precision positioning capabilities for autonomous vehicles and has the application potential to support the collaborative construction and updating of 3D urban maps.