Design and Research of Novel Industry Robot Wrist Force Multidimensional Sensor

Transcription

1 Sensors & Transducers 214 by ISA Publishing, S. L. Design and Research o Novel Industry Robot Wrist orce Multidimensional Sensor 1 Yuan Chuanlai, 2 Zhang Yueyi, 1 Xiao Shenping, 1 Kong Lingshuang 1 College o electrical and inormation engineering, Hunan university o Technology, Zhuzhou, 412, China 2 Changsha vocational & technical college, Changsha, 41111, China Tel.: Received: 2 January 214 /Accepted: 7 April 214 /Published: 3 April 214 Abstract: The main unction o the wrist orce sensor in robot remote operation is to realize the orce eedback, one as a necessary part o the contact orce o the process itsel; another is the contact orce as a robot gripper with respect to the environment the exact location o the inormation source, wrist orce sensor provides inormation can reduce the robot to peripheral equipment precision requirement. Remote sensing operation environment, the manipulator and the environment rom the direct eect o the inormation, the wrist orce sensor detection, conversion, transmission to the master, or the establishment o virtual environment and the main basis and real-time interactive environment. The sensor circuit system design, designed the bridge circuit is applicable, overall on the measurement system is designed, and or sensor data acquisition sotware programming theory. Copyright 214 ISA Publishing, S. L. Keywords: Wrist orce sensor, Virtual instrument, Robot, eedback. 1. Introduction At present, the development o robots has entered the era o intelligent robot. One o the most important eatures o intelligent robot is equipped with the surrounding environment inormation timely eedback sensor more complete to the robot control system [1], robot wrist orce sensor is one kind o the important robot sensors. With the wide application o industrial robots and robot sensors, will need a lot o, but the robot wrist orce sensor is the robot contact tasks is security, robot remote task security. Through contact orce sensor to detect the manipulator and the environment and eedback on the robot's hand, enable the operator to produce be personally on the scene o orce eect, so as to control the robot with the eeling [2]. Wrist orce sensor is one o the most important in intelligent robot, as a orce eedback control system in detecting element, can rapidly relect the change o orce and moment, should have good dynamic characteristics [3]. Dynamic characteristics o wrist orce sensor is greatly aects the overall perormance o the robot system, because i the dynamic perormance o the wrist orce sensor is poor, cannot be ast, accurately relect the variation o the measured orce and moment, so that the whole robot system can t work normally [4]. The structure determines the sensor perormance in a certain extent. The six dimension wrist orce sensor is mounted on the robot wrist (away rom the end eect recent), ull inormation can also provide three bit space, has the widespread application o zero orce on the robot teaching, contour tracking, the bullet burr and grinding, matched with a shat Article number P_196 81

2 hole, bimanual coordination and automatic assembly operations, is a robot orce control and orce/position the most important hybrid control o sensor [5]. In this paper, on the basis o existing research results at home and abroad on the basic principle o orce transormation, sensor, six dimensional sensing orce Jacobi matrix elements, structure design, structure optimization design o additional elastic, dynamic perormance analysis o elastic body, sensor system design o several key basic theory and its application are studied systematically. To orce conversion principle o the sensor is analyzed, providing theoretical basis or design and orce sensor. Analysis o additional elastic component o the strain ampliication principle, we derive the analytical relationship between structural parameters o strain ampliication and additional elastic component, inds the actors aecting the additional elastic component strain ampliication, provides the theoretical basis or structure design and optimization o the additional elastic component. The sensor circuit system design, designed the bridge circuit is applicable, overall on the measurement system is designed, and or sensor data acquisition sotware programming theory. structure is a variant o the traditional Stewart platorm, and the general Stewart platorm structure is dierent. 6 ball hinges on the platorm o the new structure o the centre is not distributed on the same circumerence, 6 ball joints under platorm o the center is not distributed in the same circle, the orce sensing element is composed o an upper platorm, the 6 elastic connecting rod, the lower platorm and 12 spherical hinges connected [9]. Its characteristics are: 6 rigid rods are divided into 3 groups, each group o two separately along three mutually perpendicular to the direction o arrangement, plane and each rigid rod axis perpendicular to each other, to realize six axis orce and torque decoupling rom the structure; each additional elastic component o relative two side are respectively pasted 2 strain gauge [1]. The ull bridge circuit, corresponding to 6 rigid rods consist o six strain bridge, when acting on a platorm o measured orce makes the 6 rigid rod micro strain, additional elastic component corresponding to the rigid rod on the producing strain was ampliied, thus can obtain six road detection signal measurement, so as to realize the six axis orce and moment. Its basic structure was shown in ig The Structure Design o a New Type o Wrist orce Sensor 2.1. inite Element Analysis o Elastic Wrist orce Sensor The elastic body structure o robot wrist orce sensor is generally more complicated, because o the inluence o the principle o design and the manuacture error and other actors, makes the coupling exists in each o the sensor output channel. Classical analytic method is diicult to accurately analyze and describe this kind o coupling relation, usually using the method o experiment or calibration [6-8]. The experimental equipment calibration results are oten not accurate enough. D A1-B1 A2 -B2 A6 - B6 A1- B1 A2 - B2 A6 - B6 = B1 A1 B2 A2 B6 A6 A1- B1 A2 - B2 A6 - B6 + d = G ( + d) (1) (2) 2.2. The Basic Theory o the Wrist orce Sensor This paper study on a new type o wrist orce sensor, which belongs to the parallel wrist orce sensor, is the six dimension wrist orce sensor, a novel parallel decoupling and isotropic robot, its x A 4 A 3 B 3 B 4 z B 5 B3 A6 2c B 2 B 1 B 6 y A 2 2a ig. 1. Schematic diagram o mechanism and structure parameters o the sensor The Deinition o Jacobi Matrix Condition Number I the connecting rod orce error d, then the moving platorm stress also has certain error d, the above ormula can be written as: Then: d = G d G d (3) = [ P ] [ P ] (4) 1 1 A 1 L 82

3 In the ormula, * said the norm o vector or matrix, have the ollowing relationship between relative deviation o and : d G G d 1 [ ] (5) The vector into the ormulation to obtain the Jacobi matrix o six dimension wrist orce sensor, the model o parallel 1 1 d e 2 P 1 1 e3 [ D] ( e) EA[ G]( e) EA M = = = d 1 e4 d d e 5 d d e6 1 d 2.4. Establishment and Simulation Analysis o Sensor Model in ADAMS e (6) Automatic Dynamic Analysis Mechanical Systems is currently the world's most widely used mechanical system dynamics simulation and the most authoritative analysis sotware. Using ADAMS sotware, the geometric model o mechanical system users can quickly, easily create ully parameterized, then the simulation analysis and add the corresponding conditions, can help improve the design o various mechanical systems, rom the simple to the connecting rod mechanism o vehicle, aircrat, satellites and even complex body. The application ield o the sotware rom the automobile manuacturing industry, engineering machinery, aerospace industry, shipbuilding industry, mechanical and electronic industry, national deense industry, general machinery to the man-machine engineering, biomechanics, medicine and engineering consulting and so on. In every ield, or a variety o products, the virtual prototype technology can save money, time or users and provides a satisactory design. 1,, ( T T V = C = CV = H H) H t hen (7) The orce vector and the reading component: D a a a a a a x D y a21 a22 a23 a24 a25 a 26 2 D z a31 a32 a33 a34 a35 a 36 3 = x a41 a42 a43 a44 a45 a46 4 a y 51 a52 a53 a54 a55 a 56 5 a z 61 a62 a63 a64 a65 a The Hardware System and Static Calibration o Wrist orce Sensor 3.1. The Elastic Body Structure o New Type Sensor Thereore, in accordance with the method described above theory, in ADAMS simulation in order to load and load data tables and output data table is as Table 1. (8) Table 1. unit load calibration table Compact structure characterized by symmetry and parallel mechanism is particularly suitable as a multi-dimensional orce sensor prototype. This paper irst presents a new innovative way parallel wrist sensor, six telescopic rod parallel mechanism to replace six one-dimensional rod, install additional elastic element on the link, as a sensor or measuring the orce wrist sensor and said given sensor coniguration parameters. Then start rom the structural parameters o the parallel orce sensor analyzes the basic principle o the sensor measuring the orce that orce transormation, derived its power transer matrix theory, and analyzes the principle o the sensor and decoupled Jacobian matrix analytic number o conditions expression, the next step or the structural design and optimization o sensor laid the oundation The Hardware System o the Wrist orce Sensor Wrist sensor hardware systems usually consist o two parts processing and signal sensing head. Sensing head portion made o an elastic body and a measuring bridge preamp circuit or generating a orce signal and a complete pre-ampliier; stage signal processing section includes ampliication, iltering, sample and hold signal, A/D conversion and or system control, computing and online communications computer systems. 83

4 or dierent applications, the system structure will be dierent. The topic o the wrist sensor system ramework was shown in ig. 2, which has the ollowing unctions and eatures: 1) Weak signal sensor output ampliication, iltering; 2) The brightest signal decoupling, get the actual orce signals; 3) Can be easily with all types o robots and host computer connection. ig. 2. The topic o the wrist sensor system ramework Static Calibration o Wrist orce Sensor Because o this mutual coupling relationship is complex, diicult to describe accurately rom a theoretical, typically using the second method, the experimental method or calibration, which will directly aect the measurement accuracy o the calibration accuracy o the sensor when using this method both to reduce the sensor manuacturing process requirements, but also to get a more accurate measurement results. bridge circuit. The test computer data acquisition system block diagram was shown in ig A( p ) y( k) = B( p ) u( k) (9) 1 p y k y k ( ) = ( 1) (1) Ap ( ) = a+ ap + ap + + ap (11) n 1 2 n B( p ) = b + b p + b p + + b p (12) n 1 1 n Quantitative relationship between process output readings corresponding with the role o looking sensor on the sensor coordinate origin between static orce vector is the static calibration. nb Y( z) = X( z) 1+ az j j= na i= j bz 4. Result and Simulation i i 4.1. The Design o Sensor Circuit System (13) In the strain gauge measuring techniques, the most widely used measurement circuit is a Wheatstone ig. 3. The test computer data acquisition system block diagram Design o Strain Distribution and Bridge Road In a one-dimensional orce measurement under load output voltage ater a gradual change value o each channel, the need or a linear it to do all the measurement data to understand the impact o each dimension orce/torque coupling between that. Because o the inter-dimensional coupled aterburner in a single direction, the other direction will produce a orce or torque eect. Single-Axis orce/torque role under the corresponding voltage output was shown in ig Signal Acquisition and Data Processing System Main unction is required calibration personnel data analysis process is called sampling, the sampling loop wrist orce sensor 6-channel signal, the sampling data to append a way to write a data ile. 84

5 4 2-2 Mx-V x direction moment(kg*mm) My-V 4 shown in ig. 6. Bridge is designed primarily to improve the measurement bridge voltage sensitivity, wide measuring range, a simple circuit structure, high precision, easy to implement temperature compensation y direction moment(kg*mm) Mz-V 1 75 x z direction moment(kg*mm) ig. 4. Single-Axis orce/torque role under the corresponding voltage output. The main data ile by channel assigned six data iles, access the entire time course o the output signal or each channel, and then call the data processing and perormance analysis program to open the data ile, the data is processed to obtain a variety o calibration results. Sensor static calibration virtual instrument sotware components were shown in ig The Experimental Device and Experimental Method Sensor circuit system design includes the ollowing aspects: the bridge design, design signal acquisition system, the data processing system and graphics display system design was ig. 5. Sensor static calibration virtual instrument sotware components. or the purpose, the data processing system mainly consider ampliying the output voltage signal, signal iltering, analog to digital conversion o the signal, the data acquisition board selection (including sampling rate and sampling precision) and so on. The data processing system and graphics design system includes a static calibration sotware design, dynamic calibration sotware design, data processing sotware and graphics sotware. Theoretical study o this chapter ocuses on the preparation o sotware measurement circuit design and measurement systems acquisition. y M z ig. 6. The data processing system and graphics display system design. 85

6 4.3. Data Acquisition and Processing Data collection using the high-speed data acquisition card PCI-182L, its main characteristics related to the A / D has been introduced. Dynamic calibration sotware interace was shown in ig. 7, the dynamic calibration experiments, the sample rate is set to 1 K, the sampling time o 3 seconds, collecting a total o 3, points, ater the end o the program automatically saves the results o the sampling data, call LabVIEW data preprocessing and response curves plotted. ig. 7. Dynamic calibration sotware interace. 5. Conclusions This paper presents a new robot wrist orce sensor. Using the inite element method structural static analysis and modal analysis, quantitative analysis o the inter-dimensional coupling, and make a reasonable evaluation o its structural design, while the natural requencies and mode shapes type; wrist sensor or static calibration, the calibration matrix has been established correspondence between the measuring orce signal and the output voltage signal; wrist sensor or dynamic calibration o the sensor to calculate the natural requency and damping ratio dynamic perormance indicators based on dynamic calibration data, using time-domain modeling method to establish a variety o dynamic mathematical model o the sensor, and its spectral estimation method carried out in the requency domain analysis; relatively low damping or sensors shortcomings dynamic compensator design, compensation good results. Acknowledgements The work was supported by unding or Hunan science and technology planning project (212GK387), and the national natural science oundation o china und ( ). Reerences [1]. Yuan Chang Su, Chih Chung Chang, Jia Lin Wang, Construction o an automated gas chromatography/mass spectrometry system or the analysis o ambient volatile organic compounds with on-line internal standard calibration, Journal o Chromatography A, 121, 2, 28, pp [2]. Zhen Li, Endalkachew Sahle-Demessie, Ashra Aly Hassan, George A. Sorial, Transport and deposition o CeO2 nanoparticles in water-saturated porous media, Water Research, 45, 15, 211, pp [3]. Nadya Goldovsky, Viktor Goldovsky, Correlational gas analyzer, Measurement, 33, 3, 23, pp [4]. Li Tang, Yaming Li, Kailai Xu, Xiandeng Hou, Yi Lv, Sensitive and selective acetone sensor based on its cataluminescence rom nano-la2o3 surace, Sensors and Actuators B: Chemical, 132, 1, 28, pp [5]. Xiao An Cao, Zhen Yu Zhang, Xin Rong Zhang, A novel gaseous acetaldehyde sensor utilizing cataluminescence on nanosized BaCO3, Sensors and Actuators B: Chemical, 99, 2, 24, pp [6]. Zhi Ming Rao, Lin Jie Liu, Jing Yi Xie, Yu Yun Zeng, Development o a benzene vapour sensor utilizing chemiluminescence on Y2O3, Luminescence, 23, 3, 28, pp [7]. Zhen Yu Zhang, Chao Zhang, Xin Rong Zhang, Development o a chemiluminescence ethanol sensor based on nanosized ZrO2, Analyst, 127, 3, 22, pp [8]. Kao Wen Zhou, Xian Lin Ji, Nan Zhang, Xin Rong Zhang, On-line monitoring o ormaldehyde in air by cataluminescence-based gas sensor, Sensors and Actuators B: Chemical, 119, 2, 26, pp [9]. Su Yuanchang, Chang Chichung, Wang Jialin, Construction o an automated gas chromatography/mass spectrometry system or the analysis o ambient volatile organic compounds with on-line internal standard calibration, Journal o Chromatography A, 121, 2, 28, pp [1]. Zhen Li, Endalkachew Sahle-Demessie, Ashra Aly Hassan, George A. Sorial, Transport and deposition o CeO2 nanoparticles in water-saturated porous media, Water Research, 45, 15, 211, pp Copyright, International requency Sensor Association (ISA) Publishing, S. L. All rights reserved. ( 86