Progress slot 1: IKVM and Weka

Recently we have been working on a number of things to prepare for the start of the project.

1. Enhancing the board data collection performance

Last time we developed a C# program to display the collected data. But soon we found the program reacting a bit too slow that sometimes it failed to handle the large amount of input data, resulting in the program not responding.

Thus, we amended the way the data displays to enhance the speed of the program. At first, we put each of the input data into a list and display them accordingly to the user. The weakness of this method is that the system fails to handle the tremendous amount of data. Considering that it is not necessary to study all the information, we changed the way the data is displayed. This time we choose to extract and display at regular intervals so that not all information is shown.

2. Testing of the sensing board with variations in parameters

We tested the sensing board with different board size and distance and the nature of materials nearby (conduction/insulator).



Materials tested: 1. large sensing board; 2. small sensing board; 3. aluminium foil; 4. lead foil.

 We found that there are no observable differences in data values for different board sizes. But with a larger sensing board it increases the effective area that can be sensed.

For nature of testing materials, we tried to use the above materials to serve as sensing boards. We used bared hands to approach and finally touch the materials. We found that conductors (the foils) led to variations in data values, while insulators (wooden ruler) did not make observable changes in values.



A small sensing board



A large sensing board



Lead foil as a sensing board

The data value started to vary as the hand was around 15cm to the board. As it continued to approach the board, value increased. Finally when the hand touched the board, there was a significant increase in value immediately.



We connected two sensors in coils. This allows two independent sensors to sense a same particular area.

3. IKVM + Weka

We wrote lines to call out Weka for use in our C# program. However, Weka is compatible only with Java, thus cannot directly communicate with C#. We thus use IKVM to allow use of Weka on C#.

Reference:

IKVM with WEKA tutorial
http://weka.wikispaces.com/IKVM+with+WEKA+tutorial

C# code to run weka classifier
http://stackoverflow.com/questions/9820348/c-sharp-code-to-run-weka-classifier