研究内容 / Research:

　中央大学理工学部 電気電子情報通信工学科 空間知能化研究室(橋本秀紀研究室)にて, 安心/安全なロボットを実現する制約条件を考慮した制御系設計および空間知能化を中心に研究しています.

　I am conducting research on robust and intelligent constrained control systems to achieve safe and reliable flexible human-assist robotics and the development of intelligent spaces.

研究テーマ / Research theme:

モデルベース制御手法では，モデル化誤差やパラメータ誤差による制御性能の劣化が一般的な問題として挙げられます．この問題に対し，これらの誤差要因を外乱として扱い，その影響を推定・補償・除去することで，ロバスト性を向上させる研究を行っています． 特に，外乱オブザーバ（Disturbance Observer: DOB）およびモデル予測制御（Model Predictive Control: MPC）の2つの手法に焦点を当て、，モデルベース制御の効率的かつ統合的なロバスト設計手法の開発を進めています．

Model-based control methods commonly suffer from degraded control performance due to modeling errors and parameter uncertainties. To overcome this challenge, my research treats these sources of error as disturbances and seeks to enhance robustness by estimating, compensating for, and mitigating their effects. In particular, I focus on two key control techniques: Disturbance Observers (DOBs) and Model Predictive Control (MPC). The goal of this research is to establish an efficient and unified robust design framework for model-based control that guarantees both stability and performance.

[ref] Takashi Ohhira, Keinosuke Yokota, Shuichi Tatsumi, Toshiyuki Murakami, "A Robust Hybrid Position/Force Control Considering Motor Torque Saturation," IEEE Access, Vol. 9, pp. 34515-34528, doi: 10.1109/ACCESS.2021.3059889, 2021.

[ref] Takashi Ohhira, Akira Shimada, "Movement Control Based on Model Predictive Control with Disturbance Suppression using Kalman Filter including Disturbance Estimation," IEEJ Journal of Industry Applications, Vol.7, No.5, pp.387-395, 2018. 

外乱推定技術（Disturbance Estimation Techniques）においては，外乱推定速度とノイズ感度がトレードオフの関係にあることが知られています． 本研究では，このトレードオフを解消するために，カルマンフィルタ（Kalman Filter）をベースとし，適応ゲイン則（Adaptive Observer Gain Law）を導入した新しい外乱推定手法の開発を行っています．さらに，外乱と状態を同時に推定することで，高精度な状態推定を実現するロバスト状態推定手法（Robust State Observer）としての応用にも取り組んでいます．加えて，非線形カルマンフィルタ（Nonlinear Kalman Filter, e.g., EKF, UKF）を用いることで，非線形システムへの拡張的研究も進めています．

Disturbance estimation techniques, including Kalman filters, disturbance observers, model error compensators, and extended state observers, commonly involve a trade-off between disturbance estimation speed and noise sensitivity, which is primarily governed by the observer bandwidth. To overcome this limitation, we are developing a novel Kalman-filter-based disturbance estimation method that incorporates an adaptive observer gain law to enhance robustness against measurement noise. In addition, by jointly estimating the system state and disturbances, this approach can be applied as a robust state observer capable of achieving high-accuracy state estimation. We are also extending this framework to nonlinear systems by employing nonlinear Kalman filters, including the extended Kalman filter (EKF) and the unscented Kalman filter (UKF), thereby broadening its applicability.

[ref] Takashi Ohhira, Akira Shimada, Toshiyuki Murakami, "Variable Forgetting Factor-based Adaptive Kalman Filter with Disturbance Estimation Considering Observation Noise Reduction," IEEE Access, Vol. 9, pp. 100747 - 100756, doi: 10.1109/ACCESS.2021.3097342, 2021.

[ref] Takashi Ohhira, Akihiro Kawamura, Akira Shimada, Toshiyuki Murakami, "An underwater quadrotor control with wave-disturbance compensation by a UKF," The 21th World Congress of the International Federation of Automatic Control, IFAC-PapersOnLine, Vol. 53, No.2, pp.9017-9022, July 2020.

近年，在宅や日常生活の中での非侵襲・非接触型の生体信号計測が注目されています．本研究では，Wi-Fi通信信号やカメラ映像を活用し，心拍数・脈波・血圧などの生体信号を推定することで，人が日常的に健康状態をモニタリングできるシステムの構築を目指しています．特に，Wi-Fi電波のチャネル状態情報（CSI）を用いた心拍変動や体動の抽出，およびカメラ映像（RGB・近赤外）によるイメージフォトプレチスモグラフィ（iPPG）信号解析を行い，非接触かつ高精度な生体センシングの実現を追求しています．

In recent years, non-invasive and contactless measurement of biological signals in home and daily-life settings has attracted increasing attention. In this research, we aim to develop a system that enables people to monitor their health conditions in daily life by estimating biological signals such as heart rate, pulse waves, and blood pressure using Wi-Fi communication signals and camera images. In particular, we focus on extracting heart rate variability and body motion from Wi-Fi Channel State Information (CSI), as well as analyzing imaging photoplethysmography (iPPG) signals from RGB and near-infrared camera images. Our goal is to achieve highly accurate and contactless biological sensing for practical health monitoring applications.

[ref] Ryoto Fujita, Takashi Ohhira, Hideki Hashimoto, "Stable Electrocardiogram Measurement Using Capacitive-Coupled Electrodes," IEEE ICM2023, March 2023.

[ref] Kazuya Tsubota, Takashi Ohhira, Hideki Hashimoto, "Heart Rate Variability and Body-Movement Extractions Using Wi-Fi Channel State Information," Proceedings of 2023 IEEE/SICE International Symposium on System Integration (SII2023), ThP1M1.4, Janu. 2023.

[ref] Keisuke Terai, Takashi Ohhira, Hideki Hashimoto, "Near-Infrared iPPG Body Motion Compensation Using Head Euler Angles During Sleep," Proceedings of 2023 IEEE/SICE International Symposium on System Integration (SII2023), WeP2M2.4, Janu. 2023.

磁気式エンコーダは，光学式エンコーダに比べて角度検出精度の面で劣る一方，高い耐環境性・小型化・低コスト化を実現できるという利点があります．本研究では，この精度面での課題を克服するための角度算出アルゴリズムの高精度化に取り組んでいます．特に，磁力信号に特徴付けを行った磁気式アブソリュートエンコーダを対象として，磁石の偏心誤差や組立誤差によって生じる角度推定誤差をモデル化し，Levenberg–Marquardt法などの数値最適化手法を用いた高精度角度推定アルゴリズムの開発を行っています・これにより、磁気式エンコーダの高精度化と実用性の両立を目指しています．

Although magnetic encoders are generally inferior to optical encoders in terms of angle detection accuracy, they offer several advantages, including high environmental robustness, compact size, and low cost. To address this accuracy limitation, this research focuses on improving angle calculation algorithms for magnetic encoders. Specifically, we investigate magnetic absolute encoders based on characterized magnetic field signals, and model angle estimation errors arising from magnet eccentricity and assembly errors. Using numerical optimization methods, such as the Levenberg–Marquardt algorithm, we aim to develop a high-precision angle estimation algorithm that improves the accuracy of magnetic encoders while maintaining their practical advantages. 

[ref8] Akishi Takeyama, Shota Komatsuzaki, Takashi Ohhira, Hideki Hashimoto, "Levenberg-Marquardt method based Precise Angle Estimation for Eccentric Magnetic Absolute Encoders," IEEE International Conference on Mechatronics 2023, March 2023.

モータは，電流を適切に制御することで定格トルクを超える出力を瞬時的に発生させることが可能です．しかし，過大な電流によってコイル温度の上昇や熱損傷が発生するため，高トルク動作を安全に維持するには熱特性を考慮した制御設計が必要です．本研究では，設計する熱モデルに基づくモータ温度推定と熱制御を組み合わせることで，熱損傷を回避しながら瞬時的な高トルク化を実現する制御手法の開発を行っています．

Motors can instantaneously generate torque beyond their rated value through appropriate current control. However, excessive current can lead to an increase in coil temperature and potential thermal damage. Therefore, control design that takes thermal characteristics into account is essential for safely sustaining high-torque operation. In this research, we develop a control method that integrates motor temperature estimation based on a designed thermal model with thermal control, aiming to achieve instantaneous high-torque operation while preventing thermal damage.

[ref] Momodayu Hattori, Subaru Murakami, Takashi Ohhira, Hideki Hashimoto, "A Coil Temperature Estimation for Disk Rotor Type Brushless DC Motors," IEEE ICM2023, March 2023.

[ref] Subaru Murakami, Momodayu Hattori, Takashi Ohhira, Hideki Hashimoto, "High-Torque Electric Motors with Coil Cooling Via Thermal Model and Peltier Element," Proceedings of 2023 IEEE/SICE International Symposium on System Integration (SII2023), WeP2M1.6, Janu. 2023.