中文摘要：

针对一类高精度随动控制系统中，跟踪目标轨迹短时不连续、跟踪指令的给定周期过大，导致随动系统跟踪响应变慢、特性变差的问题，提出了一种跟踪轨迹预测算法。该方法首先对被跟踪轨迹进行简单判定与分类，然后针对不同类型的轨迹，选择合适阶次的拟合多项式，采用改进最小二乘算法，对目标的未来轨迹进行预测。针对不同类型的轨迹，通过仿真确定了其合适的拟合多项式阶次；针对复杂的轨迹，通过仿真验证了预测算法的可行性；采用混合式两相步进电机系统进行实机试验，分析比较了轨迹预测效果。实验结果表明，采用轨迹预测方法，可以减小系统最大跟踪误差与跟踪延时。

英文摘要：

For high precision servo control, discontinuities of target trajectory or/and long-periodic instructions lead to slow response and bad characteristics in servo system. A trajectory prediction method was proposed. Targets were classified according to its differentiability and fitting polynomials with proper orders for different kinds of trajectories were chosen, and prediction with improved least square method based on trajectory classification was made. Simulation and experiment with two-phase hybrid stepper motor have been actualized to analyze the prediction result and prove the feasibility of this method, which can reduce the maximum tracking error and tracking system latency.