Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
126 tokens/sec
GPT-4o
47 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Robust Stereo Visual Inertial Odometry for Fast Autonomous Flight (1712.00036v3)

Published 30 Nov 2017 in cs.RO

Abstract: In recent years, vision-aided inertial odometry for state estimation has matured significantly. However, we still encounter challenges in terms of improving the computational efficiency and robustness of the underlying algorithms for applications in autonomous flight with micro aerial vehicles in which it is difficult to use high quality sensors and pow- erful processors because of constraints on size and weight. In this paper, we present a filter-based stereo visual inertial odometry that uses the Multi-State Constraint Kalman Filter (MSCKF) [1]. Previous work on stereo visual inertial odometry has resulted in solutions that are computationally expensive. We demonstrate that our Stereo Multi-State Constraint Kalman Filter (S-MSCKF) is comparable to state-of-art monocular solutions in terms of computational cost, while providing signifi- cantly greater robustness. We evaluate our S-MSCKF algorithm and compare it with state-of-art methods including OKVIS, ROVIO, and VINS-MONO on both the EuRoC dataset, and our own experimental datasets demonstrating fast autonomous flight with maximum speed of 17.5m/s in indoor and outdoor environments. Our implementation of the S-MSCKF is available at https://github.com/KumarRobotics/msckf_vio.

Citations (414)

Summary

  • The paper demonstrates that integrating stereo vision with the MSCKF framework yields robust and efficient VIO, matching the performance of monocular methods.
  • It rigorously evaluates S-MSCKF on EuRoC datasets, showing stable, precise performance during aggressive maneuvers and under variable lighting conditions with reduced CPU usage.
  • The open-source implementation and novel dataset support further research in autonomous MAV navigation and real-time SLAM, paving the way for enhanced flight autonomy.

Robust Stereo Visual Inertial Odometry for Fast Autonomous Flight

The paper "Robust Stereo Visual Inertial Odometry for Fast Autonomous Flight" by Ke Sun et al. presents a sophisticated approach to stereo visual inertial odometry (VIO) that aims to address computational efficiency and robustness in the field of micro aerial vehicles (MAVs). Utilizing the Multi-State Constraint Kalman Filter (MSCKF), the research provides a stereo VIO solution that stands on par with state-of-the-art monocular methods in terms of computational cost, yet it offers enhanced robustness.

Methodology

The proposed approach, Stereo Multi-State Constraint Kalman Filter (S-MSCKF), builds upon existing MSCKF algorithms by incorporating a stereo configuration. The stereo vision setup is favored over monocular solutions to improve robustness against diverse environments and dynamic motion. The authors contend that stereo vision, contrary to common belief, does not incur significantly higher computational costs. Their implementation is open-source and available on GitHub, which fosters transparency and reproducibility.

Experimental Evaluation

The S-MSCKF was rigorously tested against leading VIO methods—OKVIS, ROVIO, and VINS-MONO—on standard datasets such as EuRoC. The results underline S-MSCKF's capability to achieve comparable accuracy with reduced CPU usage, making it suitable for onboard systems with limited computational resources. Specifically, the S-MSCKF excelled in scenarios involving aggressive maneuvers and variable lighting conditions, maintaining effective performance without the need for GPU acceleration.

In fast flight scenarios, where MAVs reached speeds up to 17.5 m/s, the algorithm demonstrated remarkable stability and precision. The novel dataset, which is also made publicly available, offers a benchmark for evaluating VIO systems under high-speed conditions.

Implications and Future Work

This research contributes significantly to the domain of autonomous MAV navigation by providing a robust, efficient VIO method apt for real-time applications. The implications are profound for both practical deployment in search and rescue missions and further theoretical exploration in SLAM and state estimation.

Future extensions could focus on optimizing planning trajectories to mitigate the growth of uncertainty in VIO estimates. This would potentially amplify the operational range and effectiveness of MAVs in diverse mission profiles. The paper hints at the possibility of integrating intelligent trajectory planning to enhance localization accuracy, thus opening avenues for further investigation.

In summary, the S-MSCKF's ability to deliver robust, efficient odometry in demanding scenarios positions it as a valuable asset in the development of autonomous flight technologies, paving the way for enhanced autonomy and adaptability in MAVs.

Github Logo Streamline Icon: https://streamlinehq.com
Youtube Logo Streamline Icon: https://streamlinehq.com