基于PCA-GA-SVM的甘蔗收获机运行状态识别

由于甘蔗收获机在收获过程中智能化水平较低,依靠人工操作很容易对甘蔗收获机的运行状态产生误判,从而造成物流通道堵塞、能源浪费、收割效率低。针对这些问题,提出一种基于主成分分析(PCA)、遗传算法(GA)和支持向量机(SVM)状态识别模型。首先,通过实地采集甘蔗收获机刀盘轴、行走轴、切段轴和风机轴扭矩和行驶速度特征信息,然后通过PCA进行数据降维,最后利用GA优化参数C、γ,使用每个特性信息来训练SVM,对甘蔗收获机运行状态进行分类。结果表明:PCA-GA-SVM状态识别模型对甘蔗收获机运行状态的识别准确率为93.75%,建模时间为3.688 s,与SVM(81.25%,9.487 s)、PCA-SVM(87.5%,5.817 s)和GA-SVM(90%,8.969 s)进行对比,该模型具有最高准确识别率和最快建模速度,具有较大的应用价值。

Sugarcane harvester operation status recognition based on PCA-GA-SVM

Due to the low intelligence level of sugarcane harvester in the harvesting process, it is easy to misjudgment by manually recognizing its operation status, which leads to problems such as clogged logistics channels,energy waste and low harvesting efficiency. Aiming at these problems, a method of state recognition based on Principal Component Analysis(PCA), Genetic Algorithm(GA)and Support Vector Machine(SVM) was proposed. Firstly, the torque information of cutter shafts, walking shafts, segmentation shaft and fan shaft were extracted. Then, PCA was used to reduce the dimension of feature variable space. Finally, GA was used to optimize the parameters C and γ, and each feature was used to train the SVM to identify the working status of sugarcane harvester. Results showed that the accuracy of our proposed method was 93.75%, the time taken was 3.688 s. These results were compared with those of SVM(81.25%, 9.487 s), PCA-SVM(87.5%, 5.817 s), and GA-SVM(90%, 8.969 s). PCA-GA-SVM had the highest accuracy and the shortest time,suggesting the great application value of the approach.