Muscle Computer Interface

MUCI


A combined effort between Microsoft Research, University of Washington, and University of Toronto has made the possibility of interacting with computers with nothing but your muscles a reality. In 2008, the researchers unveiled their muscle computer interface, abbreviated MUCI.  The hardware component of MUCI consists of an armband that the user attaches to their forearm. The armband uses six electromyography sensors (EMG) and two ground electrodes arranged in a ring around a person's upper right forearm for sensing finger movement, and two sensors in the upper left forearm for recognizing hand squeezes.

Example of a gesture recognized by MUCI

MUCI allows users to interact with computers and other devices without requiring the use of their hands. Though there are alternative hands free interaction systems such as voice control and camera based systems, these are vulnerable to inaccuracy and have privacy issues.

There are existing products such as prosthetics that rely on detecting muscle activity, but MUCI is the first commercial application. Unlike the electrodes used with prosthetics, user's do not have to worry about placing MUCI's electrodes on an exact position on their arms. After slipping the armbad on, MUCI's software will undergo a set of calibration exercises to recognize the position of the electrodes and to understand the user's movements. The calibration exercises rely on machine learning algorithms that improve in accuracy with time. The algorithms use three main components of data from the electrodes: the magnitude of muscle activity, the rate of muscle activity, and the wave like patterns that occur across sensors. These three components provide sufficient data to discern the type of muscle movement that the user is exerting. Preliminary testing on 10 subjects revealed that after calibration, the system has accuracy rates of up to 95% in recognizing movement of all 10 fingers.

Potential Applications

There are a number of potential applications for this technology, including the following:

• Opening car trunk with groceries in hand - When holding grocery bags in both hands, it can be extremely difficult to access the car keys and open the trunk. MUCI can alleviate this problem by allowing the user to open their trunk by completing a simple gesture such as touching two fingers. 
• Controlling an MP3 player while jogging - It can be awkward and time consuming for a user to take an MP3 player out of their pocket and change the song, increase the volume, etc. These actions can often force the user to stop jogging and stand stationary, something that is undesired. MUCI can allow the user to easily control their MP3 player while remaining in motion. 
• Accepting/ending phone call when driving - Having to reach for a phone and fumble for the small accept button can be an and inconvenient and dangerous task while driving. MUCI can allow the user to accept and end their calls without lifting their hands off the steering wheel.
• Playing video games such as guitar hero - As demonstrated in the video below, MUCI also has entertainment applications. Users could use MUCI as a controller in games such as Guitar Area, where their actions with an imaginary air guitar would be interpreted by the system. 

It should be noted that muscle computer interfaces are still very much in the research phase. Researchers are testing how well they work in real world scenarios, such as when people walk and run while wearing it. Future plans include creating arm bands that are easier to wear and that can be camouflaged as jewelry or an article of clothing. 

Sources

http://www.newscientist.com/article/dn13770-hightech-armband-puts-your-fingers-in-control.html

http://www.technologyreview.com/computing/23813/page1/

http://www.popsci.com/technology/article/2009-10/muscle-based-interface-lets-you-literally-point-and-click-no-mouse-required