Guidelines For Input Device

Guidelines For Input Device 1

Running head: GUIDELINES FOR INPUT DEVICE

Guidelines for the Development of an Input Device for Aircraft Technicians

Daniel Zinzow Carnegie Mellon University May, 2006

Guidelines For Input Device 2

Introduction The students of Spring 2006 Rapid Prototyping for Computer Systems course worked for the ongoing problem faced by the aircraft technicians. Aircraft maintenance is complex and the technicians need to constantly refer to technical manuals and procedures while on the field. The Interactive Electronic Training Manuals (IETMs) allow the maintenance staff to follow the complex procedures while they are on the field. The project created a system that combines the training manual and the IETMs by using various devices such as mobile computing, wireless networking and different input. After much research by the team, a dial was found to be the best input device for the mobile computing platform. The purpose of this following paper is to research information about anthropometrics and work on ergonomics to develop guidelines of how a dial should be created for this platform. With this information a design will be given for the optimal input device, the current dial will be tested against this design, and the faults of the dial will be given. This information will give guidance to the clients of how well the current dial will work and also the schematics for an optimal dial. This paper will also use the anthropometrics to give guidance in developing a wearable system that will fit the maintenance workers and this information will also be used to measure the quality of the wearable system developed by the class.

Design of Controls

Guidelines For Input Device 3

Overview Poorly designed controls alone may lead to inefficiency and breakdown in the man-machine system (Fitts and Jones, 1947). To develop a well designed control the operator’s task and physical capabilities need to be analyzed so as to be able to determine the level of force, accuracy, manipulation and so forth that is required.

Types of Controls There are two types of controls: discrete and continuous. Discrete controls are those that make discrete alterations in the machine state. An example is a switch that turns ‘on’ or ‘off.’ Continuous controls are ones that are used for making continuous settings. An example is a radio volume control. The dial has certain links that cursor can only go to,. The dial is not like a volume knob that can gradually turn to change loudness and stop at any of an infinite number of intensities within its operating range. Therefore, the dial is a discrete control. A list of controls that belong to each of these types are listed in Table 1 below.

Table 1. Types of Controls and Their Functions (McCormick, 1976)

Guidelines For Input Device 4

Each type of control has its advantages and disadvantages as can be seen above and also in Table 2 below.

Table 2. Characteristics of Common Controls (Chapanis and Kinkade, 1972; Damon, Stoudt, and McFarland, 1963; Murrell, 1965)

Guidelines For Input Device 5

The usefulness of any control can be limited by such features as the ease to which it can be identified, its location and size, its relationship to the appropriate display, and the type of feedback which it gives to the operator. (McCormick, 1976)

The input device chosen for the AME system consisted of a knob (which is referred to as a dial in the Final Report), four push buttons, and a cursor joystick. Design guidelines for knobs and push buttons will be discussed. The joystick on this input device is of a custom design built just for the AME system. Suggestions can be made for how it should be constructed but there will not be as much research backing it up as with the knob and push buttons.

Knobs

Guidelines For Input Device 6 A knob is a cylindrically shaped control that is operated by gripping the circumference of the knob by the thumb and forefinger. The knob is moved by moving the thumb and forefinger in opposition. The diameter should not be too small so as to prevent it from being gripped and turned easily, yet it should not be too large so as to not take up a lot of space or to be difficult to be gripped. The dial of the More system is a knob, which is a continuous type of control. The dial is used as a scrolling mechanism, moving from keyword to keyword. The dial was constructed to be a ‘palm grip’ knob as seen in the above figure on the right

Push-buttons

Push buttons are small, single-action controls that operate only in one direction. For the input device, there are four of these push buttons. There is the pane button on the top of the device, which moves the cursor through the panes, there is the tab buttons on the left and right side, which move from tab to tab in the browser, and finally the joystick is also a button which searches the selected item on the screen when pressed.

Resistance

Guidelines For Input Device 7 Inbuilt resistance in controls is desirable since it allows the operator to make his settings with a certain level of precision and it also helps to guard against the accidental activation of the control. On the other hand, if there is too much resistance or resistance of the wrong type, performance may be reduced and the operator could experience fatigue. Table 3, below, gives some explanations of types of resistance and their associated characteristics, advantages, and disadvantages. Table 3. The Characteristics of Static and Coulomb, Elastic, Viscous and Inertial Control Resistances (Oborne, 1982.)

When deciding on what amount of resistance to use it is important to keep in mind the strength of the operator’s fingers and hands. The pane button will be used by the index finger, the right tab button by the middle or ring finger, the left tab button by the thumb, and the selection/joystick button by the index finger. Below are statistics on

Guidelines For Input Device 8 the strength of fingers. The lower number is the smallest amount of strength out of all users and the highest number is the highest amount of strength out of all users. Strength of Fingers Index: 10.50 lb. – 16 lb. Middle: 9.88 lb. – 18.48 lb. Ring: 7.02 lb. – 14.58 lb. Little: 4.63 lb. – 9.69 lb. Refer to Appendix A for a more detailed chart of Finger Strengths.

It is also important to note the strength of the hand since a the dial is a palm-grip knob and will be moved by the hand. The lower number is the smallest amount of strength out of all users and the highest number is the highest amount of strength out of all users. Hand Strength Right hand: 42 lb. – 164 lb. Left hand: 39 lb. – 148 lb. Refer to Appendix B for a more detailed chart of Hand Strength.

It is important to keep in mind the minimum amount of strength, since designing a control with a resistance to strong for the lower limit will prevent the control from being useable by possible operators. Whereas, on the other hand, designing too weak of a resistance could cause those with stronger strength to accidental activate a control. To help guide the choice in resistance recommendations are presented below.

Guidelines For Input Device 9 It is suggested by Morgan et al. (1963) that resistance for all hand controls except knobs should not be less than 2 to 5 lb (9 to 22 N), since below this level the pressure sensitivity of the hand is very poor. As seen in Table 4 below, the minimum resistance for a hand push-button is 10 oz. (2.8 N) and for Knobs it is 0-6 oz. (0-1.7 N).

Table 4. Minimum Resistances Required for Different Controls (Oborne, 1982)

“It is difficult to set any maximum figure since it depends on the type of operator, location of the controls, frequency of usage, duration, direction and amount of control movement required.” More detailed information for recommended resistance levels for knobs by Woodson (1981) is shown in Table 5.

Table 5. Maximum Torques That Can Be Applied to a Round Knob as a Function of Knob Diameter and Depth (Woodson, 1981)

Guidelines For Input Device 10

Recommendations for resistance levels for push buttons are shown on the bottom of Table 6. Table 6. Push-Button Characteristics (Department of Defense, 1974; Moore, 1975; Marrell, 1965)

Guidelines For Input Device 11

Coding Different Controls It is beneficial to code controls along some tactile dimension, so as to free the eyes and allow them to accept other visual incoming information. Vision is better at perceiving differences than touch and as a result Moore (1976) even recommends that tactile identification of controls should only be used as a final method of identifying a control, not a primary method. With the case of the wearable computer system and the IETM, the maintenance workers need to keep their eyes on the screen while manipulating

Guidelines For Input Device 12 controls, therefore they will only be able to distinguish the buttons on the input device by touch. Touch may not be the preferred method, but in this case it is the only method.

Size The size and dimensions of the control clearly need to be related to the anthropometric dimensions of the limbs to be used.

Picture 3. Hand Length (Damon, Stoudt, & McFarland, 1971)

The lower and upper limit of hand length is 6.2 in. -8.7 in. Refer to Appendix C for a detailed table on Hand Length

Picture 4. Hand Breadth (Damon, Stoudt, & McFarland, 1971)

Guidelines For Input Device 13

The lower and upper limit of the hand breadth is 3.2 in. - 4.7 in. Refer to Appendix D for a detailed table on Hand Breadth.

Table 7. Hand Thickness (Damon, Stoudt, & McFarland, 1971)

The lower and upper limit of the hand thickness .8 in. -1.4 in.

Guidelines For Input Device 14

Just like with resistance it is important to take into account that designing for one extreme can hinder the other. If the dial is designed for the lower end of the size dimensions, then those with larger hands will have difficulty holding the dial with their palm. On the other hand, if the dial is made larger, it will be difficult for those with smaller hands to both grab the dial and also to press the buttons on top of the device. To help guide in the design of control sizes recommendations follow. Table 8. shows that for a palm-grip knob the diameter should be 1.5 in. – 3.0 in. and the depth should be .6 in.

Table 8. Knob Design Recommendations (Chapanis and Kinkade, 1972; Kellermann, van Wely, and Wellmens, 1963)

Guidelines For Input Device 15

Table 9. recommends the diameter of a fingertip button be .38 in. - .75 in. and have a displacement of .12 in. - .25 in. For thumb activated buttons, the diameter should be at least .75 in. and the Displacement should be .12 – 1.50 in.

Table 9. Push-Button Characteristics (Department of Defense, 1974; Moore, 1975; Murrell, 1965).

Guidelines For Input Device 16

It should be noted that these dimensions will be altered if the operator is wearing gloves. A further explanation of the effects of gloves is described below.

Gloves Appendix E contains a table of clothing increments to add to body sizes based on type of clothing and type of job a person has. Column 5 of the table is for someone in the Army that wears gloves, which is the closest to a maintenance worker. The maintenance worker is probably not decked out with all the other clothing associated with Column 5, but this column is the only one that contains gloves as the wardrobe. The increments to be added are 1.60 inches to hand breadth and .30 inches to hand length.

Guidelines For Input Device 17 This changes the hand sizes to: Hand length: 6.5 in. – 9.0 in. Hand Breadth: 4.8 in. – 6.3 in. As a result, the recommendations for sizes should also be slightly added to. There is no specific number to increase the dimensions of knobs and controls by, figuring out how to take into account the gloves will require more than doing research but more of actual testing. By considering how gloves add size and other effects, some suggestions can be made. Gloves are the most likely type of clothing to interfere with efficient control action and therefore may increase the necessary dimensions of the control and also affect the operator’s ability to use the control adequately. Oborne (1982) explains the difficulty of gloves by stating that: “The normal sensation of ‘grip’ probably results from the pressure perceived when the flexed fingers around the gripped object press against each other. If the working glove happens to be too thick in these regions, high pressures can be generated between the fingers before the hand is firmly closed around the tool handle or equipment control, which may result in an insecure grasp. Furthermore a thick glove can also obstruct the fingers from wrapping around the handle sufficiently for a firm grip. On the other hand, if the operator is aware of these problems he may grip the control unnecessarily tightly and firmly so increasing fatigue in his finger and other muscles.”

Bradley (1969) concludes that the two most important gloves parameters is snugness of fit and resistance to slipping. Gloves are often worn for just protection and snugness and resistance to slipping may, therefore, be absent. In such cases, the size of

Guidelines For Input Device 18 the controls should be increased so as to allow adequate manipulation and the controls should also be textured to reduce the possibility of slipping. Gloves can also impede the perception of any coded texture differences on various controls, therefore any texture used for coding must be made even more explicit. Further discussion of texture is below.

Texture The quality of the control action depends greatly on the extent to which the operator’s limb is able to remain in contact with the control. On one hand, surfaces of hand-held controls should not be so smooth as to make it difficult to grip firmly. Along with smooth controls being difficult to grasp, if they are highly polished they may also cause glare, which could adversely affect the operator’s performance on a visual task. On the other hand, surfaces which are grasped should be free from any abrasive properties. There is a balance that must be made, leading to the question to what extent should the control be textured. The solution given by Oborne (1982) is “by using a non-reflective, rippled coating, but the ripples should not be raised to the degree that they cause painful pressure spots. The directions of the ripples need to be considered in so far as they should be so arranged as to be at right angles to the likely direction of force.”

Hand Choice Estimates for incidences of left-handedness range from 2 to 29 percent (Hardyck & Petrinovich, 1977) and for different types of actions a single individual may have a

Guidelines For Input Device 19 different hand preference. This could cause problems from the point of view of determining which hand should operate which control. When rotating a control clockwise, a right-handed person will need to supinate his wrist ( palm moved upwards – wrist twisted away from the body). An operator with a left-handed preference, will need to protinate his wrist (palm moved downwards – wrist twisted towards his body). Supination allows a great torque and range of movement than does protination (Damon, Stoudt, and McFarland, 1971). Controls are often designed for use by the right-handed operator and, as explained above, a left-handed operator would have more difficulty turning this type of control than a right-handed operator. This difficulty could lead to fatigue and quite possibly to accidents.

Wrist Rotation Many tools are used that require the wrist to be bent either downwards or upwards. As explained by Bridger (2003) “the effect of this action is to cause the tendons which connect the finger muscles to the forearm bones in the elbow region to bend and to become subject to mechanical stress”. Under continuous operation, this effect will lead to muscular fatigue and as a result a loss of efficiency. It is therefore recommended that hand tools should be designed to allow the device to be operated with the hand and forearm longitudinal axes aligned as close as possible. Along with muscle fatigue due to tendons having to be bent continuously, having the wrist out of neutral position will cause static load to be placed on the wrist muscles.

Guidelines For Input Device 20 This static load can then cause fatigue. This effect can be minimized by distributing the weight on the control so that the wrist does not need to work to maintain its natural position.

Optimal Solution By taking all of the anthropometrics and ergonomics recommendations, an optimal solution for an input device that has the same functionality is addressed.

Palm Knob Taking into account that gloves will be worn, the diameter should be 2 inches and depth of 1.2 inches. The resistance level should be between 1.7 N to 13.22 N. Not wanting to cause strain on the hand or to make it too difficult to turn, the best resistance would be slightly greater than 1.7 N. Due to gloves, it is necessary to have a rippled coating with the directions of the ripple be at right angles to the likely directions of source.

Push Buttons Taking into account that gloves will be worn, the diameter of finger activated push buttons should be 1 inch and have a displacement of .25 inches. For thumb activated push buttons, the diameter of the button should be 1 inch and have a displacement of 1.5 inches. Resistance levels for finger activated push buttons should be between 2.8 – 11 N. Not wanting to accidentally activate the push buttons and to make it noticeable when a push button is activated, the resistance level should be

Guidelines For Input Device 21 closer to the higher. Therefore the best resistance should be around 9 N. Resistance levels for the thumb activated push buttons should be between 2.8 – 22.7. For the same reason as just stated, the best resistance should be around 20 N.

Current Solution

Guidelines For Input Device 22

Palm Knob The depth is 1 inch, which is a good amount as it is close to the optimal size. The diameter is 2.5 at the top and 2.87 at the bottom, which might be rather large, considering that gloves will add size to the fingers. The optimal size is 2 inches. There is no texture on the dial. It is smooth and the dial moves from a small top to a large bottom, which makes grip even more difficult.

Push Buttons The smallest width of the pane button is .75 inches, which is slightly lower than the optimal 1 inch size. The width of the selection button that is inside the dial is 1.25

Guidelines For Input Device 23 inches, which is a large enough size. The side buttons have to been fully completed, so those sizes are currently unknown.

Resistance Resistance cannot be measured without sophisticated tools that due to time limits and lack of access to such an instrument, has not been measured.

Further Steps The numbers and information in this paper is just recommendations and design guidelines. To really know the necessary sizes for knobs and buttons and resistance levels, it is pertinent to test a prototype of a dial in the actual field of use. Since creating a prototype of the optimal dial is rather impossible due to time limit and lack of access to tools to could make such a dial, an optimal measurements can not be tested. What can be tested is the dial that was constructed by the class. As can be seen above, the dial is relatively close to the optimal dimensions, but resistance and texture of dial is very important to test since these areas are very dependent of type of glove worn and the capabilities of the users. An experiment was conducted to test the dial constructed by the class to test its usability. An explanation of this experiment and its results can be found in the paper “Usability of User Input Device for the More system” which is attached.

Guidelines For Input Device 24

References

Bradley, J.V. (1969). Glove Characteristics Influencing Control Manipulability. Human Factors, 11, 21-36. Bridger, R.S. (2003). Introduction to Ergonomics (2nd ed.). New York: Taylor & Francis. Chapanis, A. & Kinkade, R. (1972). Design of Controls. In Van Cott & Kinkade (1972). pp. 345-379. Damon, A., Stoudt, H.W., & McFarland, R.A. (1963). Design Recommendations for Hand and Foot Controls. In Morgan et al. (1963), pp, 262 – 275. Damon, A., Stoudt, H.W., & McFarland, R.A. (1971). The Human Body in Equipment Design. London: Oxford University Press. Department of Defense. (1974). Human Engineering Design Criteria for Military Systems, Equipment and Facilities. MIL-STD 1472B, May 15, 1970. Fitts, P.M. & Jones, R. E. (1947). Analysis of factors contributing to 460 ‘pilot error’ experiences in operating aircraft controls. Aeromedical Laboratory Report TSEAA-694-12. July, In W. Sinakokio (ed.) Selected papers in the Design and Use of Control Systems. 1961. New York: Dover. Hardyck, C. & Pertrinovich, L.F. (1977). Left-handedness. Psychological Bulletin, 84, 385-404 Kellermann, F.T., van Wely, P.A., & Willems, P.J. (1963). Vademecum – Ergonomics in Industry. Eindhoven, Netherlands: Phillips Technical Library. McCormick, E.J. (1976). Human Factors in Engineering and Design. New York: McGraw-Hill

Guidelines For Input Device 25 Moore, T.G. (1975) Industrial Pushbuttons. Applied Ergonomics, 6(1), 33-38. Moore, T.G. (1976). Controls and Tactile Displays. In K.F.Kraiss and J.Moraal (eds.) Introduction to Human Engineering. Kiln: TUV Rhineland. Morgan, C. T., Cook , J.S., Chapanis, A. & Lund, M. (1963). Human Engineering Guide to Equipment Design. New York: McGraw-Hill Murrell, K.F.H. (1965). Human Performance in Industry. New York: Reinhold Publishing Company. Oborne, David J. (1982). Ergonomics at Work. Chichester: John Wiley & Sons. The Human Factors Section. (1983). Ergonomic Design for People at Work. Eastman Kodak Company: Author. Woodson, W.E. (1981). Human Factors Design Handbook. New York: McGraw-Hill.

Guidelines For Input Device 26 APPENDIX A

Table A1. Finger Strength: Flexion of the Finger – Palm Joint (Barter, Fry, and Truett, 1956)

Guidelines For Input Device 27 APPENDIX B

Table B1. Hand Strength: Dynanometer Squeeze (Damon, Stoudt, & McFarland, 1971)

Guidelines For Input Device 28 APPENDIX C

Table C1. Hand Length (Inches) (Damon, Stoudt, & McFarland, 1971)

Guidelines For Input Device 29 APPENDIX D

Table D1. Hand Breadth at Thumb (Inches) (Damon, Stoudt, & McFarland, 1971)

Guidelines For Input Device 30 APPENDIX E

Table E1. Clothing Increments for Nude Body Measurement (Damon, Stoudt, & McFarland, 1971)