一維測試平台上高強度聚焦超聲波探頭的自適應運動控制

Abstract

高強度聚焦超聲波（High-Intensity Focused Ultrasound，HIFU）是一種醫學技術與治療疾病的方法，相比於其他治療方法，高強度聚焦超聲波有著非侵入性、無輻射治療環境與受術者能快速恢復等優勢。其中，以持續呼吸法來實現高強度聚焦超聲波手術最引人入勝，這是因為以持續呼吸法完成的手術能使受術者享有能更低的治療風險與更舒適的治療過程，整體的手術複雜程度也將下降許多，然而，相比於呼吸中止法，採用持續呼吸法來完成的高強度聚焦超聲波手術可能無法提供受術者良好的治療精準度，而本論文就是在致力於解決此問題。希望能以自動控制的角度進行精確地控制、追蹤與預測目標組織在手術中因呼吸運動所產生的位移，來達到精確的定位，讓受術者不僅能享受持續呼吸法所帶來的各項優勢，又不會降低手術精準度、犧牲效率與安全性。 本論文將利用微控制器、馬達驅動器、馬達、編碼器與線性滑軌等硬體設備，實際建構出一維運動平台，來模擬採用持續呼吸法之高強度聚焦超聲波肝臟燒灼手術的真實手術過程，並將以旋轉基底的QR與lattice遞迴最小平方法建構而成的自適應控制器應用於追蹤因呼吸所引起的肝臟運動上，期待自適應控制器的快速收斂、數值強韌與預測波形等等能力，能大幅提升超音波探頭追蹤肝臟運動軌跡的成效，以達成上述所說，藉由自動控制的角度對超聲波探頭進行精確地控制與定位，讓以持續呼吸法實現的高強度聚焦超聲波手術之受術者在不犧牲手術精準度的情況下，能享有持續呼吸法所帶來的各種優勢與好處。 在論文的最後，我們將有自適應控制架構的系統與單純只有比例與積分控制器的基準系統之肝臟運動波形追蹤結果做比較，並說明本論文所採用的自適應運動控制架構確實能有效的追蹤因人體呼吸而產生的肝臟運動，大幅提升以持續呼吸法實現的自動化HIFU手術之精準與有效性，也會討論不同的超音波影像解析度與影像處理時間延遲對於系統所帶來的影響，讓讀者更加理解，這兩因素的好壞程度，將對於系統產生巨大的影響。

High-Intensity Focused Ultrasound (HIFU) is a medical technique and treatment method that offers several advantages over other treatment modalities, such as non-invasiveness, radiation-free therapeutic environment, and rapid patient recovery. Among all methods, the use of continuous respiration technique in HIFU surgery is particularly intriguing. That is because it provides patients with lower treatment risks, a more comfortable treatment experience, and significantly reducing the overall surgical complexity. However, compared to apnea, HIFU surgery conducted using continuous respiration technique may not provide optimal positioning of ultrasound probe for patients. The purpose of this thesis is to address this issue from the automated control perspective. The aim is to achieve precise positioning by accurately tracking and predicting the displacement of the target tissue caused by respiratory motion during surgery. This approach allows patients to enjoy the benefits of continuous respiration technique without compromising precision, efficiency, and safety. In this thesis, hardware devices such as microcontrollers, motor drivers, motors, encoders, and linear slides will be utilized to construct a one-dimensional motion platform. This platform will simulate the real surgical process of HIFU liver ablation using continuous respiration technique. An adaptive controller, developed using the rotation-based hybrid lattice-QR RLS algorithm, will be applied to track the liver motion induced by respiration. The rapid convergence, numerical robustness, and predictive waveform capabilities of the adaptive controller are expected to significantly improve the effectiveness of ultrasound probe tracking of liver motion trajectories and achieve the purpose that mention above. Lastly, in the conclusion, a comparison will be made between the tracking results of the adaptive control architecture system and the reference system with only proportional and integral controllers. It will be demonstrated that the adaptive motion control architecture adopted in this thesis effectively tracks the liver motion caused by human respiration, thereby greatly enhancing the precision and effectiveness of automated HIFU surgery using continuous respiration technique. The impact of different ultrasound image resolutions and image processing time delays on the system will also be discussed to provide readers with a better understanding of how these factors significantly influence the system.