JSET: Journal of Science & Engineering ... - JSET UniKL MFIjset.mfi.edu.my/wp-content/uploads/2017/01/jset0302_01.pdf · Earlier in 1960s, the MH-1 a computer operated dexterous hand

JSET: Journal of Science & Engineering Technology Vol. 3 Issue 2 (December) 2016 pp. 1-5

1

The Experimental Study on Arduino's Application Towards The Development of Dexterous Manipulator

Muazzin Mupit1,a, Abdul Syahid M.Rasidi2,b, Aiman Nazmi Rosli3,c

1 Section of Technical Foundation Technology UNIKL-MICET Alor Gajah, Melaka 2 BET (Hons) in Industrial Automation and Robotics UniKL-MFI, Bangi Selangor 3 Section of Chemical Engineering Technology UniKL-MICET Alor Gajah, Melaka

a [email protected], b [email protected], c [email protected]

Abstract— Hand disablement due to congenital or accident at works happens frequently. Looking at this scenario, dexterous manipulator project prototyping has been developed to be a preliminary stage or platform to cater the needs especially medical aspect. This project is to design and development the dexterous manipulator also known as Robotics Hand that able to respond to real-time reflexes of a human hand as a master via controlling glove. The objectives of this project development are to optimize the usage of Arduino interfacing towards the dexterous manipulator. Secondly, to analyze the data of flex sensor and response time between Arduino and manipulator. The strategies on developing the prototype are based on prior identification, which consist kinematics of human hand mechanized equivalent. A dexterous manipulator comprises of an embedded glove (master), a processing unit as medium and mechanized hand (slave). It used servo motors to move the respective fingers according to the concurrent user hand position. The controlling glove containing a palm and five fingers. The processing unit conducted through a Mat Lab and GUINO as well as Arduino Mega 2560. This experimental prototype able to be a potential impact selectively on niche markets especially to those who had deformed hand and feet to act like normal people.

Keywords—Arduino Mega; Dexterous Manipulator and Robotics hand

I. INTRODUCTION Earlier in 1960s, the MH-1 a computer operated dexterous

hand at MIT developed by [1] Heinrich Ernst was the first iteration of a robotic hand. This hand was a gripper that used two fingers to pick and hold some blocks using electric motors as actuators and touch sensors for the object identification. Moving along through 1980s, numerous researches are done on grippers and effectively established by early 1990. Even though early grippers do not look like hand but its applications far exceeds expectations. The finger count for the grippers ranges from two to three fingers, which results in a low DOF (Degrees of Freedom) that does not make the robot hand able to act dexterously. Theoretically the least number of DOF to achieve dexterity in a robotic hand is nine (Salisbury and Mason, 1985) proven by development of Stanford Hand [2].

The development of [3] MIT/UTAH hand was the beginning of more complex robotic hand structures. It was the first robotic hand capable of dexterously manipulating objects. It had four fingers and over twenty-five DOF including the wrist joint. Much research work has been done in dexterous manipulation performed by robotic hands and highly maneuverable and anthropomorphic hands have been reported.

Fingers are commonly actuated using electric motors

because of the position accuracy, velocity control and better force exertion for proper grasping. This thoroughly improves the overall design as opposed to early dexterous hand.

II. LITERATURE REVIEW Types of End-Users

Technology evolution since the 70’s is rapidly changing the industry landscape by revolutionizing the usage of robotic hand as the main drive behind the manufacturing plant. There are various types of dexterous manipulator, these are either 1.commercial, 2.prosthetic or 3.research hand. Development of dexterous manipulator in the early 80s began with the “Soft Gripper” by Shigeo Hirose from Tokyo Institute Technology that began in the late 70’s with just 1 Degree Of Freedom (DOF) [4]. Then, Bekey and Tomovic developed one of the first prototypes hand that is the “Belgrade / USC hand” after the World War II, four degrees of freedom (one for each pair of fingers and the thumb, two). At the same time, more research and development carried out in this field to improvise prototypes and technologies. For example Stanford / JPL hand has 9 DOF. Features like four fingers also designed for handling, incorporating fingertip combined with the finger strain gauge sensors [5]. Furthermore, the “Utah / MIT hand” developed in the 80s equipped with 16 degrees of freedom with tendons count up to 32. Sensor used for position and tendon tension sensing using Hall Effect. The fingers able to withstand forces up to 3kg, on par with the human level [6] thanks to the complex tendon mounting scheme.

JSET: Journal of Science & Engineering Technology (Special Issue: August) No. 01 (2016) pp. 1-5

2

A. Commercial Hands Due to the market demand in industrial sector, the

research and development of commercial robotic hand hastened. Barrett Technology’s “Barret Hand”, used four motor, three fingers has dedicated motor while another, motorized palm spread. This breakthrough technology allows the fingers to fit to the geometry of the object. It’s also includes optical encoder for position detection. At the fingertip, forces around 1.5kg exerted and weight of the hand about 1.18 kg [7], the Barret Hand is available commercially around USD 30,000. Subsequently, Gifu Hand is developed by Kawasaki & Mouri, Gifu University that is sold for USD 50,000 by Dainichi Company. Weighted around 1.4kg and fingertip forces recorded at 0.3 kg. Gifu Hand has 16 DOF (Two coupled joints except the last inch) combined with the detection of the pressure, but not detecting the precise position. A disadvantage is that its size is greater than man sensor size and is not too sensitive. [8]. A different commercial hand is “DLR / HIT hand” in developed by Gerhard Hirzinger, The Schunk Company sold “DLR/HIT Hand” at approximately $ 60,000. This hand is greater than the human size, which is the ability to maintain up to 0.75kg force at the tip of fingers with Hall sensors and the weight of the hand about 2.2 kilograms. It has 13-controlled DOF (last two joints of each finger are linked). [9] B. Research Hands

Akio Namiki and Masatoshi Ishikawa, University of Tokyo produced U.Tokyo hand. This hand has 14 DOF research and sensor mounting joint strength. A special feature of this hand is the accuracy of the cycle time of 1 ms for a vision-based system as a whole [10] control. Then the SBC hand developed Kyu-Jin Cho and Harry Asada of MIT. Its weight only 0.8 kg and 16 degrees of freedom with 32 actuators controlled shape memory alloy. This hand segmented binary control to overcome the non-linearity of the actuator. It has a peak force unknown, but the power to weight ratio should be high [11]. C. Prosthetic Hands

Finally, ACT hand by Yoky Matsuoka from University of Washington has three fully actuated with the human musculoskeletal structure (redundant actuation) fingers. This share target is to study the human control of hand movements because the passive and active dynamic that consistent hand with the hand of man. [12]

III. METHODOLOGY

There are two task involved in this study; 1. Developing the system model and 2. Data collection and analysis.

A. System Model

Dexterous manipulator divided into two parts, which are mechanical development that applied Solid works and Proteus ISIS. Whereas software development involved Arduino IDE with GUINO Graphic User Interface which able to interact as a medium between input glove (master) to dexterous hand (slave) as shown in figure 1.

Figure 1: System layout

B. Hand Glove and Flex Sensors The approach applied to develop the data glove by using

the concept patent that invented by Young L.Harvill [14] shows in Figure 2. Where the flex sensor attached on top of the glove by scratched the fiber near the finger joint to make it locally sensitive to bending. Flex sensor is only sensitive to one side of bending position.

Figure 2: Flex sensor equipped glove by Young L.Harvill and

project developed data glove

C. Data Collection All data be presented either in Mat Lab m-files and

figures as shown in Figure 3, Graph, Tables, and GUI Screenshots or in another form accepted, along with its calculations pertaining to this matter such as transfer function, forward kinematics, control systems and numerous others.

Figure 3: A Mat Lab screenshot showing collected data

plotted on graphs


3

• In electronics circuit design, as shown in diagram Figure 4, it was developed on Proteus ISIS for debugging, testing and simulation purposes. Those components are integrated and connected according to the configuration as shown in Figure 5.

Figure 4: Circuit Diagram on Proteus ISIS

Figure 5: Overall Configuration on Fritzing

IV. EXPERIMENTAL RESULT

Overview

To analysis this project, several aspect need to take into account in term of method of acquiring, storing and processing the data meanwhile, the output is mainly focus in terms of degree of bending angle. This analysis has been done in every stage of process as shown in Figure 6. Where the input analysis has been done by acquiring the value of input voltage, which is proportional to the bending angle of the flexible sensor.

Input Analysis Analog-to-

Digital Converter

Mapping Analysis

Output Analysis

Hand Glove

with Flex Sensors

ADC with 10 bits words

Bit Mapping

Angle of

mechanical finger with servo

motor

Figure 6: Analysis Block Diagram

A. Input Analysis

It is done by acquiring the value of input voltage that has to be proportioned to the bending angle of the flexible sensor. According to datasheet, the bending resistance range is from 30kΩ up to 110kΩ. Theoretically, each bending angle can be assigned to its own resistance value by using this transfer function (1) and the data tabulated in Table 1 as well as the graph plotted in Figure 7.

(1)

Table 1: Bend degree with corresponding voltage output

Bend Degree (θ°)

Resistance R1 (k Ω)

Voltage Output (V out)

0 10 2.50 20 21 1.61 40 32 1.19 60 43 0.94 80 54 0.78

100 66 0.66 120 77 0.57 140 88 0.51 160 99 0.46 180 110 0.42

Figure 7: Output voltage versus degree

of bending angle

B. Analog-To-Digital Converter

On the Arduino board, the pins labelled ‘A’ (A0 to A5) to indicated it able to read analog signal. The way an ADC works is fairly complex and the most common techniques using


4

analog voltage to charge up an internal capacitor and measured according to time discharge across an internal resistance. Theoretically each analog voltage is assigned to its own ADC decimal value by the transfer function (2).

(2)

C. Mapping Analysis

It is done by acquiring the value of output degree that is proportional the value of ADC value in decimal as shown in Figure 8.

Figure 8: Bit Mapping Block Diagram

D. Outpun Analysis

The control signal for a servo is a stream of pulses. The exact duration of these pulses, in fractions of a second, is what determines the position of the servo. Each pulse is nominally from 1000 to 2000 microseconds (µs) in duration — one microsecond is one millionth of a second. The pulses repeat about 50 times each second as shown in Figure 9.

Figure 9: Pulse Width controls Degree Output

Each degree of angle output be assigned to its own ADC

decimal value by this transfer function (3) and plotted graph shown in Figure 10.

(3)

Figure 10: ADC versus Bending Angle

V. CONCLUSION The development of the Dexterous Manipulator, which

closely resembles human hand and gesture with real-time data response and reflexes, are accomplished. Meanwhile, real-time inputs of the master (Data Glove) are able to control the movement of the slave (mechanical hand) and able to analyze the movement and demonstrate. It is also able to communicate using the Arduino IDE and MatLab software platform.

In future this project is able to replace by nerve sensor

instead of using master (Data Glove) which attached or embedded into human muscle to replicate the actual human hand gesture.

REFERENCES

[1] Heinrich Arnold Ernst, MH-1 A Computer- Operated Mechanical Hand, D. Sc. Thesis, MIT.

[2] K. Salisbury, M. Mason, Robot Hands And The Mechanics Of Manipulation, (MIT Press: Cambridge, MA. 1985).

[3] S. Jacobsen, E. Iversen, D. Knutti, R. Johnson, K. Biggers, "Design of the Utah/M.I.T. dexterous Hand," 1986 IEEE International Conference on Robotics and Automation. Vol.3, Pages. 1520- 1532, Apr 1986.

[4] Hannes Fillipi (2007) “Wireless Teleoperation of Robotics Arms”,Lulea University of Technology: Master Thesis

[5] FYP Central Committee (2012), Final Year Project Handbook: Policy and Procedure, 2nd Edition, May 2012, Universiti Kuala Lumpur, Kuala Lumpur

[6] Vitor F. Pamplona, Leandro A. F. Fernandes, João Prauchner, Luciana P. Nedel and Manuel M. Oliveira. The Image-Based Data Glove. Proceedings of X Symposium on Virtual Reality (SVR'2008), João Pessoa, 2008. Anais do SVR 2008, Porto Alegre: SBC, 2008, (ISBN 857669174-4). pp. 204–211

[7] T. Sonoda, I. Godler: Multi-Fingered Robotic Hand Employing Strings Transmission Named "Twist Drive", IEEE/RSJ International Conference on Intelligent Robots and Systems IROS 2010, Taipei, pp.2753-2738, October 2010.

[8] Jie Liu, Yuru Zhang, Robotics and Biomimetics, 2007. ROBIO 2007. IEEE International Conference on Author(s): Page(s): 829 – 834, 15-18 Dec. 2007

[9] The Carnegie Mellon Graphics, 2010, Hands Overview Slideshow. Slide Available at: http://graphics.cs.cmu.edu/nsp/course/16-899/. (Accessed on 23 March 2013)

[10] Shadow Robot, (2010), Shadow Hands [ONLINE]. Available at:http://www.shadowrobot.com/products/dexterous-hand/ [Accessed 16 April 13].

[11] J. D. Macgyver. 2012. How to Build a Robotic Hand with Haptic Feedback. [ONLINE] Available at:


5

http://www.instructables.com/id/How-to-Build-a-Robotic-Hand-with-Haptic-Feedback/#intro. [Accessed 18 April 13].

[12] Connor Lamon, Tyler Bellenfrant, Belsin Barkhosir Sensor Glove Connor Lamon, Tyler Bellenfrant, Belsin Barkhosir (2010) Sensor Glove. [image online] Available at: http://www.ece.ucsb.edu/drupal/ieee/node/181 [Accessed: 27 April 2013].

[13] Young L.Harvill, Thomas G. Zimmerman, Jean-Jacques G. Grimaud (March 17, 1992) Motion sensor which produces an asymmetrical signal in response to symmetrical movement. US Patent No.5097252

[14] K. Elissa, “Title of paper if known,” unpublished. [15] R. Nicole, “Title of paper with only first word capitalized,” J. Name

Stand. Abbrev., in press. [16] Y. Yorozu, M. Hirano, K. Oka, and Y. Tagawa, “Electron spectroscopy

studies on magneto-optical media and plastic substrate interface,” IEEE Transl. J. Magn. Japan, vol. 2, pp. 740-741, August 1987 [Digests 9th Annual Conf. Magnetics Japan, p. 301, 1982].

[17] M. Young, The Technical Writer’s Handbook. Mill Valley, CA: University Science, 1989.

JSET: Journal of Science & Engineering ... - JSET UniKL MFIjset.mfi.edu.my/wp-content/uploads/2017/01/jset0302_01.pdf · Earlier in 1960s, the MH-1 a computer operated dexterous hand

Documents