BumbleBeat is a culmination of a semester long project for my Intro to HCI class(Georgia Tech) to design a tool to assist the deaf and the hard-of hearing in learning dance.

Bumble Beat is a wearable wristband that provides vibro-tactile feedback to its users to help them identify beats in the song, in order to understand the rhythm of the song. The feedback the device proves is fairly simple and involves a low learning curve. The band’s form factor is aesthetic and does not stand out conspicuously.

I worked in a team of 3 with Fang(Baron) He, Regis Verdin. As Team Out of the Blox, we went through the User Centered Design Process and documented the process and results in 4 reports (P1 through P4) along the way.

Timeline: 4 months, August-December 2014.

Techniques/ Methods: Literature review, Brainstorming, Personas, Storyboarding, Poster Session, Rapid Prototyping: Arduino, Soldering, Illustrator, User Evaluation, Heuristic Evaluation

Understanding the Problem

In order to understand the problem, we conducted a literature review to see what had already been done in order to understand this problem. We also read some blogs to see how the deaf and hard-of-hearing community felt about dance and the methods they employed to carry out the dance.

We set out to  design a system to assist the deaf and the hard of hearing (HoH) in the activity of learning a dance routine. An apparent problem they face when it comes to dancing is that they are not (either entirely or partially) able to listen to the music or words of a song, which are important for dance. However, being able to listen is not a requirement of dance. This has been exemplified by many deaf dancers and dance groups in the past. Dance has so many movements and physical gestures, which deaf people often excel at because of their reliance on gestures to communicate. While cochlear implants and hearing aids do exist, these can be intrusive and prohibitively expensive. Moreover, these devices are not able to address all the causes of hearing impairments or reduced hearing, especially with the “profoundly deaf” population that we are developing our system for. With our system, we hope to encourage those of the deaf and HoH community that wish to dance, by assisting them with dance practice.

Design Criteria:

We identified key design implications that guided our design process. These implications are as follows:
Our designs should employ Sensory Substitution, which refers to translating the characteristics of one sensory modality into stimulus for another sensory modality. (Source-Wikipedia)
System should be light and portable if it is something that the user has to wear so as to not affect the mobility of the dancers. Especially, if they have poor balance: might fall, etc.
System should not involve bone-conduction, as this is not effective for sensorineural hearing loss.
System should be non-invasive in that it shouldn’t interfere with other senses that are used in dance. Moreover, it should be transparent to the user and act as a tool rather than to distract them from the task they are performing.
Given that the deaf and hard of hearing do not like themselves to be labeled as disabled the system should be non-”pathologizing” of deafness. Rather, it should be seen as an experience enhancing tool.
The System should perform in real-time as it is being used for live dance and the feedback should be such that the dancer finds it useful in dance. Delayed responses to stimulus would be negative in this context.
Given that there may be users with varying levels of language proficiency, all interfaces must display easy to understand language with redundancy in the form of graphical symbols.

Keeping this in mind, we conducted an “informed brainstorming” session. We created a Google Doc where we each listed our ideas to encourage “divergent thinking”. We then sat down together to “converge” on ideas and to elaborate them.

In the process of brainstorming, we considered various modes that could be used to convey information either about the music or the choreography. Some of these ideas are:
Different colors can be applied to describe the emotion that is conveyed by music
Series of graph to visualize beat, speed, pitch and loudness of music
Symbols can be used to convey dance instructions
Vibration to represent speed, beat, and emotion of music

Finally,we zeroed in on three ideas from the perspective of vision, tactile sense and a combination of the two.

Design alternatives:



Software Prototype:BumbleBeatSoftware

Bumblebeat software interface has, generally, been divided into three functional modules as displayed on the above image: 1. Toolbar  2. Music categories and devices Explorer 3. Files Viewer.

Band Prototype:BumbleBeat Band

We used a piece of cloth to sew our LilyPad Arduino onto we stitched this on top of an elastic wrist band.  Using conductive thread and wires we made hard-soft connections to connect the Arduino to a vibration motor and LED lights.

Details about our Prototyping process can be found here: BumbleBeat Prototype



We conducted an evaluation of both the band and the software prototypes.

For the software prototype we conducted a Heuristic Evaluation with the help of other students from the MS-HCI program acting as experts. We documented the errors as well as perceived severity and ease of fix rating.

Due to limited resources and time, we were unable to conduct a test of the band with our intended users. However, we did conduct a test with proxy users. The test involved a simple reaction time test to see if the band’s vibration was perceived timely and easily. We also administered a questionnaire to guage participant’s response to the band.

Survey questions are listed as following:

7 point likert scale questions:

  1. The product was enjoyable to use.
  2. I would want to wear this product while dancing.
  3. I would feel comfortable wearing this product in public
  4. The vibration was easy to feel.
  5. I would use this product often.
  6. The product is easy to use.
  7. The product was comfortable to wear.

The 7 point likert scale is mainly used for test subjects’ attitude toward a specific aspect when they use the wristband prototype. And we can specifically improve one part of our design based on the information we get from likert scale.

4 free-response questions:

  1. What did you like about the device?
  2. What did you dislike about the device?
  3. How did the band hinder your movement? (Answer only if you answered yes to the previous Question)
  4. Do you have any other comments or suggestions?

While none of us endorse using proxy users, we had to adopt this work-around.

More details about the evaluation can be found here: BumbleBeat Evaluation