基于PSO-LSSVM的甘蔗破头率预测

A Prediction Method of Sugarcane Brocken Roots Rate Based on PSO-LSSVM

甘蔗收割机收割后的甘蔗宿根破头率是评价甘蔗收割质量的重要指标,破头率过高会严重影响下一年甘蔗产量及甘蔗收割机的广泛推广与应用,但甘蔗破头率的采集方式复杂、费时费力,是研究降低甘蔗破头率的控制策略中的一项难题。为此,提出了一种基于PSO-LSSVM的甘蔗破头率预测方法,通过在田间采集甘蔗收割机刀盘与行走子系统的工作压力、速度等信号,以此为输入数据,以甘蔗破头率为输出,建立了基于PSO-LSSVM的甘蔗破头率预测模型。结果表明:基于PSO-LSSVM的甘蔗破头率预测平均绝对误差为0.145,均方误差为0.0214,与LSSVM预测模型相比分别降低了60.7%、80.9%,与真实破头率的拟合程度高。预测模型为探究降低甘蔗破头率的控制策略提供理论基础。

The rate of sugarcane roots breakage after harvesting by sugarcane harvester is an important index to evaluate the harvesting quality of sugarcane. The high breaking rate will seriously affect the yield of sugarcane and the widespread promotion on of sugarcane harvester. However,the collection method of sugarcane head breakage rate is complicated and time-consuming,which is a difficult problem in the research of control strategy to reduce sugarcane head breakage rate.To solve this problem,this paper proposes a prediction method of sugarcane roots breakage rate based on PSO-LSSVM.By collecting the working pressure and rotational speed signals of the cutter subsystem and the walking speed of sugarcane harvester in the field,and taking these signals as the input data and the sugarcane breaking roots rate as the output,a prediction model of sugarcane roots breaking rate based on PSO-LSSVM was established.