Home /Research /Time Synchronization Accuracy Verification for Multi-Sensor System
PERCEPTION

Time Synchronization Accuracy Verification for Multi-Sensor System

Songyu Yuan, Yuehua Li, Yong Chen, Chenghao Wang

Year
2021
Citations
6

Abstract

High-precision time synchronization is essential for applications such as perception and localization in autonomous mobile robots. For multi-sensor systems, high-precision time synchronization can reduce the ghosting in perception, improve the association accuracy between prediction and measurement in localization, reduce the number of corner cases caused by data asynchronous in fusion and improve the robustness of autonomous intelligent system. However, the present researches are mainly focused on the design of time synchronization system or the calibration of time shift between different sensors. Few methods are focused on verifying the accuracy of time synchronization multi-sensor system. In this paper, a novel verification method is proposed to verify the accuracy of time synchronization in multi-sensor system. The design idea of the method is to convert and amplify the small time deviation of the synchronized multi-sensor system into other physical quantities, which can be easily measured for calculation. Two different time synchronized multi-sensor systems are evaluated by our methods in the experiment and the time synchronization accuracy levels of the two systems are verified, which also prove the feasibility of our methods. Furthermore, our verification methods are suitable for various multi-sensor application systems containing camera, LiDAR or IMU, including Visual-Inertial SLAM, LiDAR-Inertial SLAM and Multi-Mode mapping.

Keywords

Computer scienceSynchronization (alternating current)Robustness (evolution)Real-time computingAsynchronous communicationInertial measurement unitTime synchronizationSensor fusionSimultaneous localization and mappingRobot

Related papers

Browse all PERCEPTION papers