﻿ 仿尺蠖爬壁机器人自适应吸附及摇杆控制
«上一篇
 文章快速检索 高级检索

 智能系统学报  2018, Vol. 13 Issue (2): 208-213  DOI: 10.11992/tis.201611016 0

### 引用本文

GAO Guoqing, WANG Tao, WANG Binrui. Adaptive adsorption and joystick control of an inchworm wall-climbing robot[J]. CAAI Transactions on Intelligent Systems, 2018, 13(2): 208-213. DOI: 10.11992/tis.201611016.

### 文章历史

GAO Guoqing, WANG Tao, WANG Binrui
College of Mechanical and Electrical Engineering, China Jiliang University, Hangzhou 310018, China
Abstract: When a vacuum-adsorbed wall-climbing robot crawls, it is necessary for the acetabulum to closely attach onto a wall surface, and this is not easy to manipulate. To solve this problem, a space geometry model on the acetabulum foot and wall surface was developed, and a kinematics model was established on the basis of the D-H parameters. A rocker control method was designed for the wall-climbing robot with symmetric structure according to the inverse kinematics for speed. The rocker shaft was mapped as the joint velocity loop, and the adaptive adsorption action was designed according to the inverse kinematics for location. When the distance between the acetabulum foot and the wall surface was less than the set threshold, the adaptive adsorption action was triggered automatically. The robot control system and adaptive adsorption device were built, and a crawling experiment that was carried out on the horizontal wall surface revealed that the method could reduce the difficulty of manipulation.
Key words: inchworm wall-climbing robot    vacuum adsorption    rocker control    modeling

1 机器人机构

2 机器人建模 2.1 位置运动学建模

J2J3J4关节是爬行运动的关键，将爬壁机器人简化为一个平面三连杆机构。建立D-H坐标系，如图2所示。

 ${ T} = \left[ {\begin{array}{*{20}{c}} {{{{c}}_{{{123}}}}} & { - {{{s}}_{{{123}}}}} & 0 & {p{}_x} \\ {{{{s}}_{{{123}}}}} & {{{{c}}_{{{123}}}}} & 0 & {{p_y}} \\ 0 & 0 & 1 & 0 \\ 0 & 0 & 0 & 1 \end{array}} \right]$ (1)

 ${p_x} = {a_1}{{{c}}_1} + {a_2}{{{c}}_{12}} + {a_3}{{c}}{}_{123}$ (2)
 ${p_y} = {a_1}{{{s}}_1} + {a_2}{{{s}}_{12}} + {a_3}{{{s}}_{123}}$ (3)

 $\varphi = {\theta _1} + {\theta _2} + {\theta _3}$ (4)

 $\varphi = \arctan \frac{s_{123}}{c_{123}}$
 ${\theta _2} = \pm {{arctan}}\frac{{w_x^2 + w_y^2 - a_1^2 - a_2^2}}{{2{a_1}{a_2}\sqrt {1 - {{c}}_2^2} }}$ (5)
 ${\theta _1} = {{arctan}}\frac{{({a_1} + {a_2}{{{c}}_2}){w_y} - {a_2}{{{s}}_2}{w_x}}}{{({a_1} + {a_2}{{{c}}_2}){w_y} + {a_2}{{{s}}_2}{w_x}}}$ (6)
 ${\theta _3} = \varphi - {\theta _1} - {\theta _2}$ (7)

2.2 速度运动学建模

 $\left[ \begin{array}{l} {{d}}{p_x} \\ {{d}}{p_y} \\ {{d}}\varphi \\ \end{array} \right] = {{J}}\left[ \begin{array}{l} {{d}}{\theta _1} \\ {{d}}{\theta _2} \\ {{d}}{\theta _3} \\ \end{array} \right]$ (8)

 $\left[ \begin{array}{l} {\omega _1} \\ {\omega _2} \\ {\omega _3} \\ \end{array} \right] = {{{J}}^{ - 1}}\left[ \begin{array}{l} {\nu _{{x}}} \\ {\nu _{{y}}} \\ {\omega _\varphi } \\ \end{array} \right]$ (9)

2.3 吸盘足与壁面夹角分析

 $\begin{array}{c}{{\alpha }}:z = 0\\{{\beta }}:Mx + Ny + Pz + Q = 0\end{array}$

 ${{{n}}_{{\alpha }}} = {\left[ {\begin{array}{*{20}{c}} 0 & 0 & 1 \end{array}} \right]^{{T}}}, {{{n}}_{{\beta }}} = {\left[ {\begin{array}{*{20}{c}} M & N & P \end{array}} \right]^{{T}}}$

 $\cos \theta = \cos \left\langle {{{{n}}_{{\alpha }}},{{{n}}_{{\beta }}}} \right\rangle = \frac{{| {{{{n}}_{{\alpha }}} \cdot {{{n}}_{{\beta }}}} |}}{{| {{{{n}}_{{\alpha }}}} || {{{{n}}_{{\beta }}}} |}}$

 $\theta = \arccos \frac{P}{{\sqrt {{M^2} + {N^2} + {P^2}} }}$ (10)

$\theta$ 的范围为 $0 \sim \text{π} /2$ $\theta = 0$ 时，吸盘足和壁面平行，是可靠吸附的必要条件。

3 摇杆操作与自适应吸附设计 3.1 摇杆与机器人动作映射

1) 吸盘足的抬起和放下。如图4(a)，当S1为支撑足时，摆动足S2向上抬起，若此时S2为支撑足，如图4(b) ，可以看到摆动足S1是向下运动的。支撑足不同时，相同的动作，导致摆动足在Y1方向上的运动方向是相反的。为使摇杆轴和机器人运动之间的映射不因支撑足不同而改变，当S1为支撑足时， $y \to {v_y}$ ；当S2为支撑足时， $y \to - {v_y}$ 。其中， $y$ 表示摇杆 ${{y}}$ 轴的值， ${v_y}$ 表示机器人摆动足沿Y1方向运动的速度，箭头表示映射。

2) 机器人躯干的伸缩。如图5，当S1S2为支撑足时，躯干伸长代表前进，躯干缩短代表后退，摇杆与机器人动作映射为： $x \to {v_x}$ ，其中， $x$ 表示摇杆 ${{x}}$ 轴的值， ${v_x}$ 表示机器人摆动足沿X1方向运动的速度。

3) 机器人转向。如图6(a)，当S1为支撑足时，摇杆控制J1 $z \to {v_{{{J_1}}}}$ ；如图6(b)，当为S2支撑足时，摇杆控制J5 $z \to {v_{{{J_5}}}}$ 。其中， $z$ 表示摇杆 ${{z}}$ 轴的值， ${v_{{{J_1}}}}$ 表示机器人J1关节转速， ${v_{{{J_5}}}}$ 表示机器人J5关节转速。

3.2 自适应吸附设计

4 实验测试 4.1 控制系统

PC作为上位机负责机器人的运动规划、状态监控、人机交互和图像采集；关节模块中，每个电机上集成有相对式编码器，伺服控制器负责底层电流、速度及位置闭环运动控制；吸附模块由微控制器最小系统和外围电路组成，微控制器作为下位机，负责响应上位机命令。PC通过CAN和5个关节模块进行通信，通过RS485和2个吸盘足上的微控制器进行通信。摇杆和图像采集设备通过USB和PC连接。

4.2 吸附装置

4.3 爬行测试

1) 按下B1按键，S1释放；

2) 按下B2按键，S2释放；

3) 操作摇杆 ${{y}}$ 轴，抬起或放下摆动足，当摆动足至壁面距离小于自动吸附阈值时，摇杆轴失效，机器人自适应吸附至壁面；操作摇杆 ${{x}}$ 轴，机器躯体伸缩；操作摇杆 ${{z}}$ 轴，机器人转弯。