An online study was conducted (which is discussed more in detail in the research papers about the project) to gain deeper understanding on the user experience, and usability potential of the system. How and what people would be likely to use body mnemonics for? How many, and in what order differen items would be organised in the body space?

When we collated the results of all the participants of the study together, we could see a kind of a topographical map of the associational richness of different parts of the body.

Stephen Hughes developed a very compact acccelerometer board connected directly to the serial port of a Jornada Pocket PC. With his drastically improved sampling rate, range and form factor the system became truly feasible in everyday ubiquitous computing.

Ian Oakley was responsible for managing and programming the gesture recognition code running on the pocketPC. The first approach was to use Ari Benbasat's work in Inertial sensing units for portable applications, but it was found that the characteristics of the gestures his system was developed for weren't the same as what Body Mnemonics requires.

The next stage was to develop a modular frame work to test a range of simple recognition algorithms that could be combined together.

We are currently working together with the University of Ireland in Maynooth through a project called "body space" to engage in more fundamental level algorithmic development that specifically suits our purposes.

To get initial understanding of the potential of using acceleration sensing as the approach to detect the motions of the device, a "coffee cup" prototype was developed. It simply contains 3 single axis accelerometers, a switch and a stamp chip that transmits the values to the computer. With few simple approximations of the 2nd integral, we quickly found that direct position tracking would not be feasible without additional 3 gyroscopes to measure the rotational accelerations too. The drif was simply too much. Even with the affordance of the coffee cup, which is to be held upright, the slight deviations from the upright orientation shot the 3D detection off the wall.

The initial work was done at the Royal College of Art, where the concept was initially developed. Further development in the psychological and technological feasibility has been conducted at the Media Lab Europe.