What Can Data-driven Calibration Do for 6DOF Inertial Odometry?

• Item type: SIG Technical Reports(1)
• 公開日: 2022-08-29
• タイトル: What Can Data-driven Calibration Do for 6DOF Inertial Odometry?
• タイトル / 言語: en
• タイトル: What Can Data-driven Calibration Do for 6DOF Inertial Odometry?
• 言語: eng
• キーワード / 主題Scheme: Other
• キーワード / 主題: 行動認識:IMU
• 資源タイプ / 資源タイプ識別子: http://purl.org/coar/resource_type/c_18gh
• 資源タイプ: technical report
• 著者所属: Nara Institute of Science and Technology
• 著者所属: Nara Institute of Science and Technology
• 著者所属: Nara Institute of Science and Technology
• 著者所属: Nara Institute of Science and Technology
• 著者所属: Nara Institute of Science and Technology
• 著者所属(英): en
• 著者所属(英): Nara Institute of Science and Technology
• 著者所属(英): en
• 著者所属(英): Nara Institute of Science and Technology
• 著者所属(英): en
• 著者所属(英): Nara Institute of Science and Technology
• 著者所属(英): en
• 著者所属(英): Nara Institute of Science and Technology
• 著者所属(英): en
• 著者所属(英): Nara Institute of Science and Technology
• 著者名: Huakun, Liu; Monica, Perusquía-Hernández; Naoya, Isoyama; Hideaki, Uchiyama; Kiyoshi, Kiyokawa
• 著者名(英): Huakun, Liu; Monica, Perusquía-Hernández; Naoya, Isoyama; Hideaki, Uchiyama; Kiyoshi, Kiyokawa
• 論文抄録 / 内容記述タイプ: Other
• 論文抄録 / 内容記述: For low-cost IMU, the uncalibrated bias and noise will quickly propagate 6DOF odometry errors over time. This paper proposes a data-driven accelerometer calibration method based on a dilated convolution network. Then, with a state-of-the-art gyroscope calibration method, we comprehensively analyze the impact of data-driven calibration on 6DOF inertial odometry. The experimental results show that our data-driven accelerometer calibration can reduce the bias by a factor of 5 to 10 and decreases the noise by a factor of 2 to 5. Through our exhaustive evaluations and analysis of data-driven calibration methods, the primary finding is that the data-driven calibration methods can slow down the error growth rate by 40-200 times. However, the effect of accelerometer calibration is only noticeable after calibrating the gyroscope. This fact would be experimental support for the design of future data-driven 6DOF inertial odometry.
• 論文抄録(英) / 内容記述タイプ: Other
• 論文抄録(英) / 内容記述: For low-cost IMU, the uncalibrated bias and noise will quickly propagate 6DOF odometry errors over time. This paper proposes a data-driven accelerometer calibration method based on a dilated convolution network. Then, with a state-of-the-art gyroscope calibration method, we comprehensively analyze the impact of data-driven calibration on 6DOF inertial odometry. The experimental results show that our data-driven accelerometer calibration can reduce the bias by a factor of 5 to 10 and decreases the noise by a factor of 2 to 5. Through our exhaustive evaluations and analysis of data-driven calibration methods, the primary finding is that the data-driven calibration methods can slow down the error growth rate by 40-200 times. However, the effect of accelerometer calibration is only noticeable after calibrating the gyroscope. This fact would be experimental support for the design of future data-driven 6DOF inertial odometry.
• 書誌レコードID / 収録物識別子タイプ: NCID
• 書誌レコードID / 収録物識別子: AA12628327
• 書誌情報: 研究報告コンシューマ・デバイス＆システム（CDS）; 巻 2022-CDS-35, 号 8, p. 1-10, 発行日 2022-08-29
• ISSN / 収録物識別子タイプ: ISSN
• ISSN / 収録物識別子: 2188-8604
• Notice: SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc.
• 出版者 / 言語: ja
• 出版者: 情報処理学会