中文摘要：

提出了一种采用容量控制技术减少训练数据的结构风险最小化空时均衡器（SRM-STE）。该方法在利用通信信号中学习信息的同时,通过容量控制手段补充滤波器自身的结构信息,并自适应调整容量控制作用,部分降低滤波器对训练数据的依赖性。通过仿真实验,并与递归最小二乘空时均衡器（RLS-STE）、正则化递归最小二乘空时均衡器（rRLS-STE）和典型支持向量机空时均衡器（SVM-STE）比较,得出结论：SRM-STE可以在更少的训练数据条件下保持良好的跟踪性能,且鲁棒性较好,在复杂环境下的无线移动通信中具有一定的应用前景。

英文摘要：

A capacity control technology is presented to reduce training data in channel equalization,and a new space-time equalizer based on structure risk minimization criterion（SRM-STE） is designed.It not only takes advantage of the learning information in communication signals,but also appropriately exploits the structure information of the filter itself through adaptive capacity control.So,it can work better with a lack of training data.Simulation shows that when compared with recursive least squares space-time equalizer（RLS-STE）,regularization RLS-STE（rRLS-STE） and typical support vector machine space-time equalizer（SVM-STE）,SRM-STE has better tracking performance with much less training data,and costs little additional computation complexity.It is applicable for wireless mobile communications.