基于动力学特性的多叶光栅时间最优轨迹规划  

Time-optimal Leaf Sequence Planning of Multi-leaf Collimator Based on the Servo Dynamic

作　　者：张翔[1] 叶佩青[1,2] 解传滨[3] 张辉[1,2] 

机构地区：[1]清华大学机械工程系,北京100084 [2]清华大学摩擦学国家重点实验室,北京100084 [3]中国人民解放军总医院,北京100039

出　　处：《机械工程学报》2017年第6期16-25,共10页Journal of Mechanical Engineering

基　　金：国家科技支撑计划(2015BAI03B00);清华大学摩擦学国家重点实验室重点(SKLT12A03);北京市科技计划(Z141100000514015)资助项目

摘　　要：旋转放疗设备中,多叶光栅是实现调强放疗的关键功能部件,在轨迹规划中合理利用其动态特性对提高放疗精度和效率具有重要意义。针对某旋转放疗设备的多叶光栅,建立其动力学模型并获取其伺服驱动系统进给性能,建立由系统输入和干扰引起的几何误差预测模型;针对调强放疗中滑移窗治疗模式的多叶光栅轨迹规划问题,以治疗时间最优为目标,将伺服进给性能和几何误差作为约束条件,提出一种基于传动系统动力学特性的时间最优轨迹规划算法;针对临床射野的剂量分布,对此算法进行数值仿真和试验验证。试验结果表明,该算法的几何误差平均值比恒加速度轨迹规划减少20.55%,治疗效率提高20.51%;几何误差平均值相对于最大进给性能轨迹规划减少33.45%,但治疗效率降低17.22%,由此可见,此算法能够在保证治疗精度的前提下有效提高治疗效率。Multi-leaf collimator, which plays critical role in radiotherapy, can be implemented to achieve intensity modulated radiotherapy. Therefore, more and more attention should be paid to investigating the dynamic performance of the multi-leaf collimator to improve the treatment efficacy and accuracy in leaf trajectory planning. The mathematical model of a multi-leaf collimator＇s servo drive system is established firstly, where the range of the servo feeding performance can be identified and the prediction model of geometric error can be modeled according to the individual axis＇s tracking error caused by the position command and external disturbance. Afterwards, in order to utilize the shortest time to complete the treatment plan, a time-optimal trajectory sequencing algorithm based on the dynamic model of the multi-leaf collimator is introduced under the constraints of servo system＇s capabilities and geometric error by the technique of sliding window in intensity modulated radiotherapy. Finally, numerical simulation and experiments are conducted to validate the effectiveness of the algorithm by using a clinical radiation filed. The experiment result shows that the geometry error can be reduced by about 20.55% and 33.45% while the treatment time can be improved 20.51% and -17.22% compared to the leaf trajectory planning method of constant acceleration and maximum servo feeding performance respectively, which improves treatment efficiency on the basis of specified treatment accuracy.

关 键 词：多叶光栅 几何误差 动态误差 伺服进给性能 动力学特性 

分 类 号：TP23[自动化与计算机技术—检测技术与自动化装置]