Tea_k screenshots

by Kristof Van Laerhoven,
Starlab NV/SA

This version aims at combining the work of TecO (serial reading and cues extraction) and Starlab (adaptive context recognition).

version 1.1

This version includes:
* sensorvalues plot
* Self-Organizing Map plot
* microphone wave plot
* Markov adjacency plot
* compact overview plot
* user-definable context descriptions
* user-definable inputs

All calculations and plotting are done real-time, the input is read from either a microphone or the prototypical TEA device, which is usually attached to the serial port (or both). All plots are resizeable.

version 1.2

The only big change is to be found in the Markov Model implementation: the plot has been changed to a drawing of a graph, while internally the contexts are linked to the states in the Markov model in stead of the neurons of the Self-Organzing Map. This leads to a simpler, and easier-to-understand Markov model. Also new is the text-to-speech output: the (user-defined) context descriptions are spoken by the software (mainly for future diagnostics on wearable platforms and demos). Apart from the microphone, there is also the possibilty to view the real-time outputs of one, predifined sensor (in the picture, the light sensor shows a harmonic signal since it is near a TL-light).