====== Projet Groupe 13 ====== ==== MAIN INFORMATIONS about the Project ==== === MEMBERS === == Name of the members of the Group == * Justal Kévin * Isoard Jean-Christophe * Tissière Alexandre == Equipments == * Phone : HTC 8S (A620e), Windows Phone 8 (updated to Windows 8.1) * IMEI : 358721050411700 === PROJECT === == Title of the Project : == Shake and luminescence recognition == Content of the Project == K-min clusterization of data samples to recognize when similar recorded events occurs. === Developped SOFTWARE === == Screenshots and GUI description == The GUI of our application contains 2 buttons used to launch and stop the record of values for the accelerometer and the luminosity. There is also a button and a text area to show the current state of shaking and luminosity acknowledged from recorded values and k-means algorithm. {{:projets:plim:20152016:apppl.png?200|}} == SOFTWARE PACKAGES of the Project == You can download the Visual Studio Project: {{:projets:plim:20152016:phoneapp1.zip|}} == Required SOFTWARE Environment == * Computer : * Windows 10 with Virtualization * Visual Studio 2015 with Windows Phone Developer Packages * Device : * Windows Phone 8.1 == How To build solution from project sources == * Open project solution in Visual Studio 2015 * Build app (PhoneApp1) === HOW TO USE IT and RESULTS === == Tutorial == The application needs some samples to perform the recognition, so you need to record some samples by using the buttons associated : there is one button for each capture (one for accelerometer and one for light sensor), you press one of the button to launch the capture, you put the phone in different situations during the record (moving normally and shaking hard for the accelerometer; low, normal and high luminosity for the light sensor), then you press again the same button to stop the capture. Once the samples recorded, the application will be able to recognize different situation for accelerometer and luminosity : the third button will show a special text when you shake the phone (you press it to clear the text) and the text area will show the actual level of luminosity (low, normal or high). Take care of correctly recording samples because the recognition is based on the k-means clustering (2 clusters for accelerometer and 3 for lightsensor), so if this step is done wrongly, the recognition might don't work. You can always record new samples using the associated buttons, the previous sample corresponding are erased and only the last record is used for recognition.