Coco-LIC: Continuous-Time Tightly-Coupled LiDAR-Inertial-Camera Odometry using Non-Uniform B-spline (2309.09808v1)

Abstract: In this paper, we propose an efficient continuous-time LiDAR-Inertial-Camera Odometry, utilizing non-uniform B-splines to tightly couple measurements from the LiDAR, IMU, and camera. In contrast to uniform B-spline-based continuous-time methods, our non-uniform B-spline approach offers significant advantages in terms of achieving real-time efficiency and high accuracy. This is accomplished by dynamically and adaptively placing control points, taking into account the varying dynamics of the motion. To enable efficient fusion of heterogeneous LiDAR-Inertial-Camera data within a short sliding-window optimization, we assign depth to visual pixels using corresponding map points from a global LiDAR map, and formulate frame-to-map reprojection factors for the associated pixels in the current image frame. This way circumvents the necessity for depth optimization of visual pixels, which typically entails a lengthy sliding window with numerous control points for continuous-time trajectory estimation. We conduct dedicated experiments on real-world datasets to demonstrate the advantage and efficacy of adopting non-uniform continuous-time trajectory representation. Our LiDAR-Inertial-Camera odometry system is also extensively evaluated on both challenging scenarios with sensor degenerations and large-scale scenarios, and has shown comparable or higher accuracy than the state-of-the-art methods. The codebase of this paper will also be open-sourced at https://github.com/APRIL-ZJU/Coco-LIC.

• The paper’s main contribution is a non-uniform B-spline representation that dynamically adjusts control points to motion complexity for improved trajectory estimation.
• It leverages a frame-to-map reprojection factor to tightly fuse LiDAR, IMU, and camera data, enhancing both accuracy and computational efficiency.
• Extensive real-world experiments demonstrate the method’s robustness and superior performance compared to state-of-the-art approaches.

Overview of Coco-LIC: Continuous-Time Tightly-Coupled LiDAR-Inertial-Camera Odometry using Non-Uniform B-spline

The paper "Coco-LIC: Continuous-Time Tightly-Coupled LiDAR-Inertial-Camera Odometry using Non-Uniform B-spline" presents a novel approach to LiDAR-Inertial-Camera Odometry (LICO) that aims to improve real-time pose estimation accuracy and system efficiency. This is achieved through the innovative use of non-uniform B-spline-based continuous-time trajectory representations, which allow for the adaptive distribution of control points based on the motion dynamics. The significance of this approach lies in circumventing the inefficiencies of uniform control point distribution and improving sensor data fusion.

The core contribution of the paper is threefold:

1. Non-Uniform Continuous-Time Representation: The authors employ non-uniform B-splines to parameterize the continuous-time trajectory, dynamically adjusting control points according to motion complexity. This contrasts with traditional approaches utilizing uniform B-splines, which may face difficulties with real-time performance due to under- or over-parameterization.
2. Efficient Sensor Fusion: The proposed system tightly couples LiDAR, IMU, and visual data without interpolation by formulating a frame-to-map reprojection factor. By leveraging map points from a global LiDAR map, the system avoids expensive depth optimization of visual pixels, thereby enhancing both accuracy and computational efficiency.
3. Comprehensive Evaluation: Extensive real-world experiments demonstrate the system’s heightened accuracy, robustness in face of sensor degeneration, and efficiency in both challenging and large-scale environments. Quantitative results show improvements over state-of-the-art methods in terms of accuracy and runtime performance.

Implications and Future Developments

The system's efficient and adaptable design suggests significant implications for robotics and autonomous systems, particularly in environments where varied and dynamic movements are prevalent. The Coco-LIC system proves beneficial in mobile robotic platforms with heterogeneous sensor configurations, potentially enhancing navigational precision and robustness against environmental challenges.

The theoretical underpinning of non-uniform B-splines offers a potent methodological advancement, enabling systems to more accurately model complex trajectories without excessive computational overhead. This work indicates a promising direction wherein dynamic adaptability is crucial to advancing continuous-time SLAM systems, setting the groundwork for future research involving real-time multi-sensor fusion in robotics.

Potential areas for future research could involve the exploration of more sophisticated control point placements, integration with advanced mapping techniques, and extending applicability to sensors like event cameras. Moreover, further investigations may examine scalability aspects and performance trade-offs when integrating additional sensor modalities into the continuous-time framework.

In conclusion, this work contributes significantly to the field of robotics and sensor fusion by addressing key challenges associated with real-time trajectory estimation and heterogeneous sensor data processing. The insights gained from this research offer promising avenues for enhancing the accuracy and efficiency of future autonomous systems.