Motion Gestures are intentional, voluntary, 3D movements of body parts detected through body-worn inertial sensors
Comoge is a first of its kind rapid prototyping mobile tool that allows interaction designers to design, prototype and quickly test if the motion gesture they designed would work in a user setting. Think Sketch for motion gestures. It is an ecosystem that consists and Android app that is used in conjunction with any Bluetooth-enabled wearable devices to record gesture samples for prototyping and testing purposes.
First things first
Comoge was my undertaken by me for the required credits of a two-semester long Masters project at Georgia Tech. Comoge was awarded the best Masters project prize among 55 other projects that were presented with it.
Comoge was the product of previous research conducted by Dr Aman Parnami. His interest in contextual gesture design started with his curiosity to understand the failure in the widespread adoption of natural user interfaces (NUI's). To understand more about the subject he conducted interviews with several Interaction Designers, field observations into the gesture design process and also did an extensive literature review into the subject of motion gestures. His research revealed the following facts
- Embodied interactions through NUI's are easier to access and faster to perform
- NUI development is mostly in the hands of developers and tech-savvy individuals who lack the ability to make such interactions usable and a part of a cohesive ecosystem.
- Interaction designers who can make NUIs usable are unable to contribute due to their lack of technical aspects such as machine learning and hardware sensors and poses a high entry barrier
- The gesture design process if it exists, is very fragmented
- Furthermore, the gesture design process is done in laboratory settings which fail to take the end user contextual factors into account which leads to low end-use recognition rate for gestures.
Contextual factors that affect gesture design
Goals & Constraints
Based on the problem space that we were tackling we had the following goals to achieve
- The end product should be a rapid prototyping tool for motion gestures
- The end product should enable contextual sample capture - which meant that the product had to target the mobile and wearable ecosystem
- The end product should hide the technical complexity but still enable Interaction Designers to leverage it for prototyping motion gestures
The simplified gesture design process
Prototyping and Exploration
Evaluating the gestures
That's an overview of Comoge and the gesture prototyping process with Comoge. To learn more about the evaluations of v1 and v2 designs and the changes in design between v1 and v2 read on!
Details about the process
Evaluating v1 designs
As mentioned in the Project overview before, I started the project by tweaking some of the designs provided by Apurva Gupta as a part of her Master's project. I then set out to build a working application which we could use to answer the following questions
Does Comoge support rapid, iterative, in-situ prototyping of motion gestures by Interaction Designers?
Does Comoge promote reflection on contextual factors?
Is Comoge usable?
To answer these questions we conducted a usability study with the working app. I helped Aman Parnami design the study and it was as follows
Results - the good aspects of v1 designs
Comoge supports Rapid Prototyping
- All participants created a total of 1456 samples
- Participants recorded an average of 16.32 samples per gesture
- On an average, participants spent around 5.44 mins per gesture
Comoge supports context reflection
- Participants changed their gesture design based on context constraints after exploring
- Participants reflected and commented on social acceptability of gestures, vehicular motion and device form
Comoge provides a design process
- Participants felt that the tool provides much-needed clarity in terms of providing a streamlined process they can follow
- Aspiring and novice Interaction designers especially liked that the tool scaffolded a design process for gestures
Results - the bad aspects of v1 design
Information Architecture & Feedback need work
- Tab layout with floating action buttons was confusing
- Testing was perceived to be a minor action
- Participants were unclear on when they had enough samples to start testing
Information models of Projects and gestures not enough
- This was especially important for professional Interaction designers. They asked questions like - "What if I want to test 2+ versions of gesture for a task?"
- A middleman between gestures and projects would enable more robust testing
The interface lacked a consistent visual design
- Participants felt that the app lacked polish and didn't look professional
- Different icon styles and inconsistent typography with muted colors affected the aesthetic appeal of the app
Redesign - fixing the concerns from v1
Evaluating v2 designs
To evaluate the v2 designs I created an InVision prototype using the mockups I had created. The purpose of this evaluation was to confirm that the new designs address the concerns from the previous design without compromising on the good aspects of the previous design. To evaluate this design I designed a user interface evaluation study and a comparative study with the following protocol
The results from both the User Interface evaluation and the comparative study we can conclude that the v2 designs improve upon the v1 designs while at the same time preserving the rapid prototyping, in-situ core-functionality of Comoge. Furthermore, it not only scaffolds a design process for designers it adds extra flexibility and shortcuts for professionals to use as they see fit.