Malinen Matti, Huttunen Tomi, Kaipio Jari P
Department of Applied Physics, University of Kuopio, PO Box 1627, 70211 Kuopio, Finland.
Phys Med Biol. 2003 Mar 21;48(6):745-62. doi: 10.1088/0031-9155/48/6/304.
In this paper, a model-based optimization method is derived to control the thermal dose in biological tissues for ultrasound surgery. The optimization method uses the bioheat equation as a system model and quadratic cost criteria for the desired thermal dose. Time-harmonic quasi-stationary ultrasound fields are used as the heat source. In this method the optimal phase and the amplitude trajectories are found directly by minimizing the associated cost function. The approach also allows for maximum input amplitude constraints. The method is based on the Hamiltonian form of the system and results in a large dimensional nonlinear optimization problem which is solved with a gradient-type iterative scheme. The performance of the optimization method is tested with 2D simulations and it is shown that the approach is able to yield a feasible nominal solution. This nominal evolution would then eventually be sought to be maintained with the help of a feedback controller during the actual sonication.
本文推导了一种基于模型的优化方法，用于控制超声手术中生物组织的热剂量。该优化方法将生物热方程用作系统模型，并采用二次成本准则来确定所需的热剂量。时谐准静态超声场用作热源。在此方法中，通过最小化相关成本函数直接找到最优相位和振幅轨迹。该方法还考虑了最大输入振幅约束。该方法基于系统的哈密顿形式，导致一个高维非线性优化问题，通过梯度型迭代方案求解。通过二维模拟测试了优化方法的性能，结果表明该方法能够产生可行的标称解。然后，在实际超声处理过程中，最终会借助反馈控制器寻求维持这种标称演变。
