Bimanual Computer Input and Forearm Support Implemented and Evaluated in an Integrated System

Dan Odell, 2004 Ph.D. Thesis

Advisor: Paul Wright

This work addresses issues in providing more efficient, intuitive, and comfortable input to computer users. These are important problems because the costs associated with poor human-computer interfaces can be very high - ranging from steep learning curves to wasted time and, in the worst case, to disabling injury. An integrated approach to these problems is taken by providing both forearm support and bimanual (or, twohanded) input. Since there are both strong physical and cognitive aspects to this approach, work in this area requires development of both new hardware - to enable new forms of input, and new software - to use these new forms of input in a meaningful way.

On the hardware side, the benefits of forearm support are first explored through an experimental study testing a new forearm support device. This study verifies the benefits of forearm support for a variety of tasks both for reducing upper extremity muscle load and for improving subjective responses. Building on these findings, a new computer input device, called the Command Chair, is presented to provide more ergonomic computer input, including forearm support, while also providing bimanual input.

Following the approach of integration, the Command Chair approach integrates an office chair, keyboard input, bimanual pointer input, and forearm support into a single computer input system. This allows the workstation designer to consider the input system as a whole, rather than as a collection of individual devices. In turn, this allows for better control of body activities and postures, with the goal of providing user comfort over a full workday. This system uses the location of the keyboard to position the mouse, removing the need to switch between devices.

Workstation throughput tests developed for this work are presented in conjunction with conventional pointing tests to quantify the performance of the Command Chair relative to other workstations. Measures from these tests demonstrate that the Command Chair improves user wrist posture and reduces subjective wrist fatigue relative to a traditional input station, but also provides slower input speed - partially attributable to higher system inertia and poor bearing performance.

A new technique to evaluate the tradeoff between improved comfort and reduced efficiency in computer workstations is developed to interpret these results. This technique works by normalizing different metrics of workstation performance to a monetary cost basis. Considering the workstation operating costs as a function of time indicates that the Command Chair is less costly to operate than a traditional computer input station when used for more than thirty four days. This result verifies that the Command Chair provides improved overall workstation performance relative to a traditional computer workstation.

On the software side, a new bimanual software technique is developed to interpret the bimanual input that the Command Chair provides and apply it in meaningful and useful ways. This technique provides a new means of bimanual command selection, and is called "Bimanual Marking Menus." To verify the anticipated performance benefits of this technique, a test environment is used to compare its performance with five other techniques: static toolbars, hotkeys, grouped hotkeys, marking menus, and toolglasses.  This experiment builds on previous work by setting the comparison in a commonly encountered task, shape drawing. In this context, grouped hotkeys and bimanual marking menus are found to be the fastest. Subjectively, the most preferred input method is the newly developed technique: Bimanual Marking Menus. Toolglass performance is unexpectedly slow. The results of this experiment verify the speed, usefulness, and potential of bimanual input, and provide a new and powerful technique for its implementation.

As a whole, this work provides evidence supporting the benefits of both forearm support and bimanual computer input. Additionally, this work demonstrates how bimanual computer input and forearm support can be leveraged for improved usability, efficiency, and comfort in human-computer interaction - the ultimate goal of this project.