Cursor Control Devices Research Articles

Performance, characteristics, and error rates of cursor control devices for aircraft cockpit interaction

This document is the Accepted Manuscript version of the following article: Peter R. Thomas, 'Performance, Characteristics, and Error Rates of Cursor Control Devices for Aircraft Cockpit Interaction', International Journal of Human-Computer Studies, Vol. 109: 41-53, available online 31 August 2017. Under embargo. Embargo end date: 31 August 2018. Published by Elsevier. © 2017 Elsevier Ltd. All rights reserved.

The Effect of Pressurized Space Gloves on Operability of Cursor Controls, Mobility, and Strength

Space Human Factors Engineering (SHFE) researchers at Johnson Space Center (JSC), in collaboration with the Extravehicular Activity (EVA) Suit Team, performed a glove box study investigating operability of cursor controls, mobility, and strength of the participants while wearing pressurized gloves. Performance time was collected under a range of glove pressures (0, 0.8, 4.3, 6.3, and 8.1 psid), as well as bare-handed. Controls tested included a Castle switch, Rocker switch, and Trackball. The study was undertaken to determine impacts of operating controls under higher than nominal (i.e., > 4.3 psid) suit pressures. Operability when using cursor control devices was tested with an interactive software task representative of the types of actions that will be required to interact with space vehicle displays (display navigation and selection of a target). Results indicate that cursor control devices can be operated at pressures up to 8.1 psid, albeit with some difficulty. With respect to mobility, increased pressure seemed to affect thumb mobility more than the fingers. As the number of participants was limited in this initial feasibility study, further studies should be performed with a larger number of participants to evaluate performance with different hand/glove sizes, as well as with alternative device designs that are more ergonomically flexible and forgiving of hand and finger dimension changes brought on by increases in pressure. Results from this study may have implications for other gloved task environments.

Cursor Control Device Test Battery

The test battery was developed to provide a standard procedure for cursor control device evaluation. The software was built in Visual Basic and consists of nine tasks and a main menu that integrates the set-up of the tasks. The tasks can be used individually, or in a series defined in the main menu. Task 1, the Unidirectional Pointing Task, tests the speed and accuracy of clicking on targets. Two rectangles with an adjustable width and adjustable center- to-center distance are presented. The task is to click back and forth between the two rectangles. Clicks outside of the rectangles are recorded as errors. Task 2, Multidirectional Pointing Task, measures speed and accuracy of clicking on targets approached from different angles. Twenty-five numbered squares of adjustable width are arranged around an adjustable diameter circle. The task is to point and click on the numbered squares (placed on opposite sides of the circle) in consecutive order. Clicks outside of the squares are recorded as errors. Task 3, Unidirectional (horizontal) Dragging Task, is similar to dragging a file into a folder on a computer desktop. Task 3 requires dragging a square of adjustable width from one rectangle and dropping it into another. The width of each rectangle is adjustable, as well as the distance between the two rectangles. Dropping the square outside of the rectangles is recorded as an error. Task 4, Unidirectional Path Following, is similar to Task 3. The task is to drag a square through a tunnel consisting of two lines. The size of the square and the width of the tunnel are adjustable. If the square touches any of the lines, it is counted as an error and the task is restarted. Task 5, Text Selection, involves clicking on a Start button, and then moving directly to the underlined portion of the displayed text and highlighting it. The pointing distance to the text is adjustable, as well as the to-be-selected font size and the underlined character length. If the selection does not include all of the underlined characters, or includes non-underlined characters, it is recorded as an error. Task 6, Multi-size and Multi-distance Pointing, presents the participant with 24 consecutively numbered buttons of different sizes (63 to 163 pixels), and at different distances (60 to 80 pixels) from the Start button. The task is to click on the Start button, and then move directly to, and click on, each numbered target button in consecutive order. Clicks outside of the target area are errors. Task 7, Standard Interface Elements Task, involves interacting with standard interface elements as instructed in written procedures, including: drop-down menus, sliders, text boxes, radio buttons, and check boxes. Task completion time is recorded. In Task 8, a circular track is presented with a disc in it at the top. Track width and disc size are adjustable. The task is to move the disc with circular motion within the path without touching the boundaries of the track. Time and errors are recorded. Task 9 is a discrete task that allows evaluation of discrete cursor control devices that tab from target to target, such as a castle switch. The task is to follow a predefined path and to click on the yellow targets along the path.

Evaluation of a two cursor control device for development of a powered laparoscopic surgical tool

The current study was performed to test two electronic cursor control devices (a Touchpad and a MicroJoystick) for use in an articulating powered laparoscopic tool. A simple target acquisition test was conducted to test how well the cursor control devices could be manipulated and how accurate they were (including endpoint and movement path). The study varied the width (0.27, 0.54 and 1.07 cm) of the target as well as the hand position used (thumb and index finger control). Additionally, each participant was able to choose their ideal operating cursor speed for each cursor control device. The MicroJoystick had a higher throughput and movement variability than the Touchpad. In all other categories tested, the cursor control devices did not differ significantly. The speed of the cursor control devices did not affect the performance of the devices; therefore, the ideal cursor speed could be chosen by the participant. Finally, the hand position did not affect the performance of the devices. This experiment found both hand positions, cursor control devices and all speeds could be used to effectively manipulate an articulating end-effector in a powered, cauterising, laparoscopic tool. This article addresses the advantages and limitations of two control mechanisms for laparoscopic tool controls considering both the subjective and objective data. The controls are tested in two hand positions to test how well the participants can manipulate the device while minimising perceived fatigue.

INTEGRATION OF EMG AND EGT MODALITIES FOR THE DEVELOPMENT OF AN ENHANCED CURSOR CONTROL SYSTEM

While Eye Gaze Tracking (EGT) systems have demonstrated their potential as computer cursor control devices, their application outside the laboratory environment has been less prominent than originally expected. This may be due to limitations in the stability of the cursor controlled by EGT devices and in the potential for unintended selections ("clicks"). To address these EGT limitations, we have developed a cursor control system that integrates electromyogram-based control with EGT-based control. This integrated system allows a user to perform small cursor displacements and selection operations ("left click") by detecting distinct changes in electromyogram (EMG) signals generated during specific facial movements. Specifically, the system monitors and processes, in real time, EMG signals from four electrodes attached to the head of the subject. We compared the performance of this EMG/EGT integrated system, with the performance of a stand-alone EGT system and a hand-held mouse, in point-and-click trials. Statistical analysis of the results indicates that the EMG/EGT system is slower than the EGT system and the mouse, but it also revealed that it is significantly less error-prone than the EGT-only system. The analysis also revealed that the performance of the EMG/EGT system does not seem to be well matched to Fitts' model.

Evaluation of four cursor control devices during a target acquisition task for laparoscopic tool control

Current laparoscopic surgery instruments create awkward postures which produce fatigue and pressure points in surgeons. In order to alleviate some of this discomfort a new laparoscopic tool had been developed with the inclusion of an articulating end-effector manipulated by a trackball. The current study was developed to access the performance of four input devices which could replace the manual trackball in a powered laparoscopic tool. A simple Fitts’ law task was conducted and the devices’ performance was evaluated with both subjective and objective measures. This article makes three main contributions to the scientific community. First, it provides a comparison of four control devices (TouchPad, Mouse Button Module, MiniJoystick Module and MicroJoystick) for use in a powered laparoscopic tool. Second, it provides an understanding of how the non-traditional measure of target re-entry can be utilized to compare control devices and how this relates to the more traditional measures of throughput and error rate. Finally, it contributes to the understanding of how a user's familiarity with a control device could affect the subjective and objective performance of the device. The main results indicate that the TouchPad and MicroJoystick are the best candidate-devices for use in a powered laparoscopic tool. The article also provides support for utilizing the new measure target re-entry when comparing control performance. Although studied in the application of laparoscopic surgery, the results can be generalized for the design of any hand-held device in which the speed and accuracy of the control device is critical.

Cursor Control Device Test Battery: Development and Application

A test battery was developed for cursor control device evaluation. Four tasks were taken from ISO 9241–9, and three from previous studies conducted at NASA. The tasks focused on basic movements such as pointing, clicking, and dragging. Four cursor control devices were evaluated with and without Extravehicular Activity (EVA) gloves to identify desirable cursor control device characteristics for NASA missions: a large trackball, a Hulapoint mouse, a small trackball, and an aircraft trackball. Conclusions include: 1) the test battery is an efficient tool for differentiating among input devices; 2) devices used with gloves have to be larger, and should allow good hand positioning to counteract the lack of tactile feedback; 3) none of the devices, as designed, were ideal for operation with EVA gloves.

Comparison of Four Cursor Control Devices for Development of a Powered Laparoscopic Surgery Tool – Users' Opinions

Laparoscopic surgery greatly benefits the patient: it reduces recovery time, has a lower cost than other procedures, and has a greater cosmetic benefit. With every benefit, there is usually a cost. Laparoscopic surgery comes at the cost of the surgeons. Surgeons must cope with disadvantages caused from unergonomic handles. The current tools create awkward postures, pressure points and nerve legions. The purpose of the current study was to determine the best input device for a hot, powered, laparoscopic surgery tool. The current study tested four different input devices on a previously designed ergonomic laparoscopic tool handle using a target acquisition task. The input devices tested were the: Mouse Button Module, TouchPad, MiniJoystick Module, and MicroJoystick. Users' provided feedback via a questionnaire. Results show that the TouchPad and MicroJoystick are the best input devices for a powered tool with the MicroJoystick leading the pack.

Forward/up directional incompatibilities during cursor placement within graphical user interfaces

Within graphical user interfaces, an indirect relationship between display and control may lead to directional incompatibilities when a forward mouse movement codes upward cursor motions. However, this should not occur for left/right movements or direct cursor controllers (e.g. touch sensitive screens). In a four-choice reaction time task, 12 participants performed movements from a central start location to a target situated at one of four cardinal points (top, bottom, left, right). A 2 × 2 × 2 design varied directness of controller (moving cursor on computer screen or pen on graphics tablet), compatibility of orientation of cursor controller with screen (horizontal or vertical) and axis of desired cursor motion (left/right or up/down). Incompatibility between orientation of controller and motion of cursor did not affect response latencies, possibly because both forward and upward movements are away from the midline and go up the visual field. However, directional incompatibilities between display and controller led to slower movement with prolonged accelerative phases. Indirect relationships between display and control led to less efficient movements with prolonged decelerative phases and a tendency to undershoot movements along the bottom/top axis. More direct cursor control devices, such as touch sensitive screens, should enhance the efficiency of aspects of cursor trajectories.

MOUSE AND DISPLAY TABLETS AS CURSOR CONTROL DEVICES

Given the proliferation of touch-sensitive screen technologies, the factors contributing to efficiency of cursor control device were examined to better inform choice and design of Graphical User Interfaces. Fitts' law can be used to describe the relative efficiency of these cursor control devices. Experiment One required 18 participants to move a cursor between targets depicted upon a WACOM PL400 display graphics tablet, over various amplitudes, using a computer mouse or stylus, examining the speed and accuracy of cursor placement. The touch sensitive screen allowed faster acquisition of targets than the mouse. Since the mouse has a higher control/display ratio (gain) than the touch sensitive screen, Experiment Two addressed manipulation of the gain settings on computer mouse performance. Low gain offered some accuracy benefits. Possible posture problems associated with the touch sensitive screen were explored in Experiment Three. Screen orientation had little effect upon cursor positioning performance, suggesting the benefits of the touch sensitive screen can be transferred to user-friendly postures.

Cursor Control Device Characteristics

An analysis of cursor positioning may provide guidelines for improvements in cursor control. Four experiments addressed efficiency of cursor control devices (Mouse, Digitising Pen, Accupoint, Trackball). Participants moved a cursor leftwards, upwards or rightwards, positioning it in large or small targets situated in near or far space on the computer screen. Cursor coordinates were sampled every 5 ms. The number of submovements and the proportion of time spent in deceleration were analysed. Participants could not plan movements controlled by an Accupoint. Cursor trajectories were more variable in near space for detachable manipulanda due to potential cartesian coordinate system incompatibilities

User verification through pointing characteristics: an exploration examination

Previous research has highlighted that all data security systems rest on the assumption of accurate user verification. Although much research has focused on speech and keystroke characteristics, no study has examined the potential of indirect cursor control devices, such as mouse or pen, as a means of personal user verification. This paper presents the result of an exploratory examination of the measurement of individual users' behaviour elicited while using two indirect input devices (puck and pen). Ten subjects undertook two series of 500 trials using the puck and pen in a counterbalanced design. Each trial involved guiding the cursor to one of ten possible targets, projected upon one of five angles of projection and one of two distances. Data recorded comprised a series of ( x, y) coordinates and associated time stamps of each component of the movement. Five parameters for each full movement were extracted-total time to reach the target, maximum velocity of the movement, maximum acceleration for the movement, time to maximum velocity of movement, time to maximum acceleration of the movement. Results of a series of Multivariate Analyses of Variance indicated that for both the puck and the pen data, there were statistically significant differences between subjects for each of the five parameters. This set of discriminating parameters was then embedded in a verification paradigm. Results indicated that the average error rates varied between 39% (S.D.=2%) and 14% (S.D.=2%) for the puck and 38% (S.D.=2%) and 12% (S.D.=2%) for the pen. Error rates were dependent upon the number of samples used when making the verification decision. This research suggests that pointing device measurements using the current parameters are not a viable method of user verification by themselves. However, with average error rates as low as 12%, it may be reasonable to integrate them into a multi-modal security system. Practical implications and future research directions are discussed.

A comparison of three computer cursor control devices: pen on horizontal tablet, mouse and keyboard

It might seem likely that an electronic pen manipulated on a desk top tablet would elicit better performance and greater user preference than other cursor control devices. This paper presents four experiments that examine this hypothesis by comparing speed, accuracy, and preference for a pen device with the two most widely used cursor control devices (mouse and keyboard) in a drawing (graphics) environment. In the context of these experiments pen use was never significantly faster than mouse, while both pen and mouse use were always significantly faster than the keyboard. In contrast, keyboard use was always more error free than the mouse, which in turn was always more error free than the pen. These error differences were significant in a majority of the cases. Subjects expressed preference for the mouse in general work (significant). For highly accurate work, subject preference was for the keyboard (significant).

The Effects of Physical Attributes of Computer Interface Design on Novice and Experienced Performance of Users with Physical Disabilities

It is well accepted that with even very simple tasks, a user's performance with a cursor control device improves substantially over some period of time before stabilizing. Although no systematic studies are available concerning how particular attributes of screen or device design affect the rate at which users learn to interact with a system, past studies with input devices have shown that the overall period of time required to learn to physically interact with a system is generally quite short. Hence the lack of attention paid to the “learning” phase with respect to physical interaction is probably justified. For users with mobility impairments, however, not only may the overall physical learning phase be significantly longer than for nondisabled users, but certain features of the interface design may require a longer learning period than others. Depending on how different “initial” performance is from “practiced” performance, systems meant for “walk up and use” or casual use may need to be designed differently to allow easy access for persons with mobility impairments. In addition, adaptive interfaces which change the physical design of features over time as a user becomes more proficient may facilitate access for individuals with impaired motor control. Twenty persons with impaired hand and arm function (as a result of spinal cord injury) performed a target acquisition task with five cursor control devices. The task required that the user select targets of different sizes and distances using both “point and click” and “drag” modes of interaction. Time and errors were recorded. The results indicate not only that some physical design attributes negatively effect performance, but that the magnitude of the effects differ for “initial” performance and “practiced” performance. In fact, in some cases attributes which had no effect once performance had asymptoted were shown to have a significant effect on novice performance. Also, some features required significantly longer periods of time for the users to become proficient at using than others. The implications for interface design are discussed.

Cursor Control Device Use by Persons with Physical Disabilities: Implications for Hardware and Software Design

Computer technology has the potential to offer individuals with physical limitations greater levels of independence and increased opportunities for meaningful employment, but this can only be realized when the individual can interact efficiently with the computer. Choosing a cursor control device is particularly important given the growing popularity of direct-manipulation style interfaces. Twenty persons with impaired hand and arm function (as a result of a spinal chord injury) and 10 nondisabled persons performed a target acquisition task with five cursor control devices: a mouse, trackball, cursor keys, joystick, and tablet. Even persons with profound impairment were able to compensate for their disability and operate each device by using minor device modifications and/or unique operating strategies. These modifications and compensation techniques are described. Regardless of the physical skill level of the user, the rank ordering of the five devices with respect to target acquisition time was the same. The mouse, trackball, and tablet provided better performance than the keys, which provided better performance than the joystick. Dragging was particularly problematic for persons with motor control limitations, as was acquiring small targets. The implications of the results for hardware and software design are discussed.

The evaluation of cursor control devices for space station freedom

Human-Computer Interface researchers at the Johnson Space Center are faced with the challenge of designing a cursor control device for use in zero gravity so that crewmembers can take full advantage of direct manipulation interfaces. Over the past year, the Crew Interface Analysis Section (CIAS) and the Astronaut Space Station Support Office at the Johnson Space Center in Houston have been working together to generate a research-based recommendation for the design of a microgravity cursor control device for Space Station Freedom.

A Comparison of Cursor Control Devices on a Graphics Editing Task

Six cursor control devices (absolute touchpad, mouse, trackball, relative touchpad, displacement joystick, and force joystick) were compared on seven graphics editing tasks. Analysis of subjects' performance data showed better task completion times (TCT) for the trackball and mouse than the remaining four devices. Preference rankings by subjects reflected the performance results.

Comparison of Six Cursor Control Devices Based on Fitts' Law Models

Six cursor control devices were compared on a target acquisition task which required subjects to move a cursor into square targets of varying sizes and at various screen distances. The target acquisition performance data were fitted to movement time models proposed by Fitts, Jagacinski, and Kvalseth. Regression analysis results indicated good predictions of target acquisition performance for the six cursor devices. The best fit was obtained with the trackball across the three models.