基于SVM数据滤波的非同轴两轮机器人转弯自动平衡控制方法

Automatic turning balance control method for non coaxial two wheeled robots based on SVM data filtering

摘要:

非同轴两轮机器人由于其轮子的轴线为不共线、不平行构造,使其在转弯时容易出现不稳定的情况。为了提高该机器人在复杂环境下的控制性能,研究基于SVM数据滤波的非同轴两轮机器人转弯自动平衡控制方法。切换机器人传感器节点状态,并定义传感器锚点,以实现对机器人转弯运动数据的转化与分解。选取SVM滤波参数,通过分析该条件下机器人运动行为奇偶模特性的方式,实施对传感器数据的耦合重构,完成对机器人传感器数据的SVM耦合滤波处理。分析两轮机器人足端步态规律,按照转弯足端步态的非同轴规划原则,解算转弯轨迹,并推导具体的控制函数表达式,完成传感器数据SVM耦合滤波下非同轴两轮机器人转弯自动平衡控制方法的设计。实验结果表明,基于上述方法可将俯仰角、横滚角、偏航角同时控制在合理的数值范围之内,有效保障了机器人转弯过程中的运动平衡性。

Abstract:

Non coaxial two wheeled robots are prone to instability during turns due to the non collinear and non parallel construction of their wheel axes. In order to improve the control performance of the robot in complex environments, a non coaxial two wheeled robot turning automatic balance control method based on SVM data filtering is studied. Switch the status of robot sensor nodes and define sensor anchors to achieve the transformation and decomposition of robot turning motion data. Select SVM filtering parameters, analyze the odd even model of robot motion behavior under this condition, implement coupling reconstruction of sensor data, and complete SVM coupling filtering processing of robot sensor data. Analyze the gait pattern of the two wheeled robot"s foot end, follow the non coaxial planning principle of the turning foot end gait, calculate the turning trajectory, and derive the specific control function expression. Complete the design of the non coaxial two wheeled robot"s turning automatic balance control method under sensor data SVM coupled filtering. The experimental results show that based on the above method, the pitch angle, roll angle, and yaw angle can be controlled within a reasonable numerical range, effectively ensuring the motion balance of the robot during turning.