CNN-LSTM-based modeling estimation method for pneumatic clamping force of robot

The invention discloses a CNN-LSTM-based robot pneumatic clamping force modeling estimation method, which comprises the following steps: reasonably selecting input parameters of a clamping force estimation model by analyzing an industrial robot pneumatic clamping control system so as to reflect complex hysteresis nonlinear characteristics of the system, introducing a CNN feature extraction layer, optimizing an LSTM network by using the CNN, and carrying out modeling estimation on the robot pneumatic clamping force based on the CNN-LSTM. The problem that a traditional LSTM model is difficult to effectively obtain data association features is solved; compared with traditional electric claw control, the pneumatic clamping force effective estimation method is free of sensors and low in cost. 本发明公开一种基于CNN-LSTM的机器人气动夹持力的建模估计方法，先通过对工业机器人气动夹持控制系统的分析合理选择夹持力估计模型的输入参数来体现系统的复杂迟滞非线性特性，再引入CNN特征提取层，利用CNN优化LSTM网络，解决传统LSTM模型难以有效获取数据关联特征的问题；与传统电动爪控制的相比较，该发明是一种无传感器、低成本的气动夹持力有效估计方法。