Controller data driving model identification method based on event triggering

The invention discloses a controller data driving model identification method based on event triggering. The method mainly comprises the steps of data acquisition and arrangement, model architecture search, initial model construction, model evolution, model prediction, triggering event calculation and the like. The method combines the advantages of the batch processing least square method and the nonlinear forgetting factor recursion method, and has the advantages of high precision, good real-time performance and good convergence. 本发明公开了一种基于事件触发的控制器数据驱动模型辨识方法，该方法主要包括数据采集与整理、模型架构搜索、初始模型构建、模型进化、模型预测与触发事件计算等步骤。该方法结合了批处理最小二乘法和非线性遗忘因子递归法的优势，具有精度高、实时性好、收敛性好的优点。