Previous works

There were three main problems in HoloLens interaction: gaze, pointing gesture, and voice input. We were to come up with a finger augmented device that will integrate the solutions for each problem - specifically, we wanted to create a ring with a trackpoint.

Specified goal

• Within finger augmentation, we discovered that most of the current input devices rely on movement only, such as the joystick or the mouse. On the other side, we found out there are some devices that are based on pressure, which are less common and therefore have more value in research than movement-based ones. We also saw the advantage of the pressure-based device as having no push-backs when using, and the potential to be utilized in many different ways.
• Moreover, the previous research question had some issues regarding the similarity. we wanted to compare the head gaze and our device and see how they would each work in conducting the original tasks in HoloLens, but we realized that our targets of comparison were too different from each other, which means there would be more independent variables to consider. Therefore, instead of simply creating a new device, we decided to develop and compare two finger augmented input devices, the pressure-based one and the movement-based one. In particular, we will compare the isometric trackpoint to the isotonic trackball using movement and selection tasks.

Prototyping

• We bought a thinkPad keyboard from Lenovo and and successfully disassembled the trackpoint from the keyboard. However, the problem was that the trackpoint cannot select but can only move the cursor. Therefore, we will implement a pressure sensor on top to recognize ‘click’s.

• Then, we connected the trackpoint to Arduino DUE using microUSB input. On the Arduino, a Bluetooth module was connected as well since HoloLens only recognizes external devices through Bluetooth signals.

• The trackball, on the other hand, will be connected to Arduino UNO because it originally is a part for Arduinos. And this will need a Bluetooth module for the same reason as trackpoint.

• In addition, about the ring shape - Considering the comfort, our ring prototypes will be made through 3d Prining using some elastic filaments such as NinjaFlex or SemiFlex.

Evaluation

• At first, we thought of using Fitts’ Law: representative model of evaluating selection, but we actually cannot use it because the trackpoint itself doesn’t move. Even if we cannot apply the law to our experiment, however, we will just use the paradigm or the factors of experimental design such as time and distance between each target.

• We also referred to a paper on target selection for augmented reality. The researchers used a point target selection task to measure its performance, and There were two types of targets: those near and inside the device Field of View, and far and outside the device Field of View. For each type, a total of 8 targets exist and they are positioned in 45 degrees apart from each other.

• Our tasks will resemble those of the paper. There are two types of tasks. In task A, the targets will be inside the field of view and for task b, the targets will be outside the field of view. We will measure factors such as the time consumed to select each objects, the error rate, the workload, social acceptance and user’s perceived difficulties. On the other hand, the control variables should remain the same – this will include the size of target objects, the distance between two targets, and the selecting order will be the same within each task.

To-do list for next week

• Finish building Unity app for experiment on HoloLens
• Connect hardware(physical prototype) to HoloLens
• Conduct experiments(evaluation)