基于FFRLS-AEKF的6轮足机器人电池SOC估计

针对6轮足机器人动力电池的荷电状态(state of charge, SOC)估计精度低、电池模型准确度不高等问题，提出一种基于带遗忘因子的递推最小二乘(recursive least squares with forgetting factor，FFRLS)与自适应扩展卡尔曼滤波(adaptive extended Kalman filtering，AEKF)相结合的估计算法。首先通过FFRLS算法辨识建立动力电池等效模型参数；然后利用AEKF对SOC在线估计，并为参数辨识提供准确的开路电压；最后以机器人锂电池包为对象，在动态应力测试工况(dynamic stress test , DST)下实验验证了该算法可以准确地估算动力电池SOC，SOC估计相对误差在2.5%以内。 

SOC Estimation of Six-Wheeled-Legged Robot Battery Based on FFRLS-AEKF

In view of the problems of the six wheeled-legged robot, such as low estimation accuracy of state of charge (SOC) and low accuracy of battery model, an estimation algorithm based on forgetting factor-based recursive least squares (FFRLS) and adaptive extended Kalman filter (AEKF) was proposed. Firstly, the parameters of the power battery equivalent model were identified based on FFRLS algorithm. Secondly, AEKF was used to estimate SOC online and provide accurate open circuit voltage for parameter identification. Finally, taking lithium battery pack of the robot as an example, a validating experiment was carried out under dynamic stress test (DST) conditions. The results show that the algorithm can accurately estimate the SOC of power battery, and the relative error of SOC estimation is less than 2.5%.