基于树木切割机器人铅酸电池SOC特征值估计

Estimation of SOC Eigenvalues of Lead-Acid Batteries for Cutting Robot

为优化用于修剪靠近输电线路树木枝杈的树木切割机器人的特性,对机器人的铅酸电池展开研究。通过建模铅酸电池,选用电池荷电状态(State of Charge,SOC)作为模型的输出,完成缺失特征数值SOC的计算;根据电压、电流、电池容量以及温度建立神经网络模型,并通过数据拟合得到电池容量与电池剩余电量SOC的关系曲线,构建电池容量与剩余电量关系的模型。通过神经网络拟合验证算法的置信度,并使用卡尔曼滤波验证估计值与测试数据的关系。

In order to optimize the characteristics of a tree-cutting robot used to prune branches of trees close to transmission lines,the article investigates the robot’s lead-acid battery.By modeling the lead-acid battery,the article selects the State of Charge(SOC)as the output of the model and completes the calculation of the missing characteristic value SOC;establishes a neural network model based on the voltage,current,battery capacity,and temperature,and obtains the relationship curve between the battery capacity and the SOC of the battery’s residual power through data fitting,and constructs the relationship between the battery capacity and the remaining power of the battery model.The confidence of the algorithm is verified by neural network fitting,and the relationship between the estimated value and the test data is verified using Kalman filtering.