Tracking Through Walls

Abstract

Tracking Through Walls is a project based on signal and processing which aim to project the objects location in an environment through an android application. The purpose of this project is to bring about a cost and energy efficient alternative to the traditional approaches that are being used to attain security in different environments. The localization and detection of objects has been done through various algorithms, the main concept of all of them being the display of results, objects‟ location, using radio tomographic imaging. Experimental data is gathered by placing radio transmitting sensors to detect changes in the movement. For this project, a data set already obtained by several experiments was used. The algorithm developed involved forming clusters for each node placed in a setup and then comparing the obtained values with the threshold set based on results. Instead of producing a tomographic image a diagrammatic figure is being displayed on the user‟s mobile phone screen as the real-time project of where the object(s) could be. However, training the algorithm required a hit and trial method to obtain accurate results. The distortions within signals and multiple nodes falling into same cluster values resulted into a slower response time which was eliminated by making the designed algorithm more efficient through regular programming practices. Currently the scope of the project is limited to set data set without providing a flexibility to the user to test just any environment. However, the project can be carried out further by setting up radio transmitters and gathering up live data to train the currently devised algorithm. To further involve more technology, image processing filters can be applied to a captured image for displaying results on to the mobile phone screen.