An Optimization-Based Multi-Sensor Fusion Approach Towards Global Drift-Free Motion Estimation

Ke Wang,Chuan Cao,Sai Ma,Fan Ren

DOI: https://doi.org/10.1109/jsen.2021.3064446

IF: 4.3

2021-05-15

IEEE Sensors Journal

Abstract:Precise and drift-free motion estimation is an essential technology for autonomous driving. Single-sensor methods such as laser-based or vision-based have proven to be inadequate. To solve the problem, we proposed an optimization-based fusion approach that incorporates information from complementary sensors to achieve high accuracy and global drift-free. The core idea is to construct a globally unified pose graph through a dual-layer optimization strategy. The local estimation layer obtains the relative pose through LiDAR odometry and visual-inertial odometry. Subsequently, by introducing the absolute geographic position information of GPS, the accumulated drifts are corrected in the global optimization layer. The performance of our approach has been evaluated both in real-world environments and public datasets. The result demonstrates that our approach outperforms other state-of-the-art algorithms, with an average translation error of 0.8045% and an average rotation error of 0.0043deg/m.

engineering, electrical & electronic,instruments & instrumentation,physics, applied

What problem does this paper attempt to address?

The paper attempts to address the problem of achieving high-precision and drift-free motion estimation in autonomous driving. Single-sensor methods (such as those based on laser or vision) have been shown to be insufficient to meet this demand. Therefore, the authors propose an optimization-based multi-sensor fusion method that achieves high-precision and globally drift-free motion estimation by combining information from complementary sensors. Specifically, the method constructs a globally unified pose graph using a two-layer optimization strategy, where the local estimation layer obtains relative poses through LiDAR odometry and visual-inertial odometry, while the global optimization layer corrects accumulated drift by incorporating absolute geographic location information provided by GPS. Experimental results show that this method outperforms other state-of-the-art algorithms in both real-world environments and public datasets.