基于深度图像的轮足机器人坐姿平衡控制仿真  

Simulation of Sitting Balance Control for Wheeled-legged Robot Based on Depth Images

作　　者：熊长智 周慧 贾杰[2] XIONG Chang-zhi;ZHOU Hui;JIA Jie(College of Artificial Intelligence,Jiangxi University of Applied Sciences,Jiangxi Nanchang 330000,China;College of Information,Nanchang Hangkong University,Jiangxi Nanchang 330000,China)

机构地区：[1]江西应用科技学院人工智能学院,江西南昌330000 [2]南昌航空大学信息工程学院,江西南昌330000

出　　处：《计算机仿真》2023年第10期435-439,共5页Computer Simulation

基　　金：国家自然基金项目(61263012);江西省教育厅科技项目(GJJ203014);江西省教育厅科技项目(GJJ203002)。

摘　　要：现有的轮足机器人坐姿控制方法无法根据采集图像判断环境障碍体,导致坐姿控制的响应速度不够快,且易受到周边环境干扰。现提出基于深度图像处理的平衡控制方法。分层处理深度图像,划分危险预警层、动态调整层及安全层,以此判断环境内障碍体。利用最大类方差法分离图像中背景与障碍物,完成轮足机器人坐姿动作初步避障。建立姿态平衡模型,分析机器人坐姿动作动力学,计算轮足的静摩擦系数,确定轮足平衡驱动因素,明确控制器平衡性指标,得到坐姿平衡最优参数。利用混合零密度优化及加权变量函数,完成补偿控制。经仿真对比证明,所提方法的平衡控制过程更稳定、控制器响应速度快,且有效提升了方法的抗干扰能力。The existing methods are unable to judge the environmental obstacles by the collected images,leading to slow response of posture control.Therefore,a balance control method based on depth image was proposed.Firstly,depth image was layered into early-warning layer,dynamic adjustment layer and safety layer.On this basis,we could determine the obstacles in environment.Then,we used maximization of interclass variance to separate the background and obstacles in images and thus to complete the preliminary obstacle avoidance in sitting position of the wheeledlegged robot.Secondly,we built a postural balance model to analyze the dynamics of the sitting posture of robot and calculate the static friction coefficient of wheel,thus determining the driving factors of balancing wheel and leg.After that,we defined balance indexes of controller and thus to obtain the optimal parameters of the sitting balance.Finally,we used hybrid zero-density optimization and weighted variable function to complete the compensation control.Simulation results prove that the proposed method is more stable during balance control,and the controller response speed is fast.The anti-interference ability is also improved.

关 键 词：深度图像 轮足机器人 平衡控制 动力学分析 平衡优化参数 静摩擦系数 

分 类 号：TH112[机械工程—机械设计及理论]