Performance Analysis of Machine Learning Techniques in Device Free Localization in Indoor Environment

The potential benefits of device free localization technique has accelerated the research in target detection domain in indoor environment. Received signal strength based target localization technique is commonly used as it is easily equipped in commercially available modules. Machine learning models that understand the variation in received signal strength due to the presence of target and assessing the position of the target is more reliable in a noisy and changing environment compared to conventional models. In this work, two popularly used machine learning models, deep neural network (DNN) based deep learning model and support vector machine models are used to assess the location of the target in an indoor. The performance analysis of the models is assessed with and without measurement error variance condition in four different node configuration setup using mean localization error parameter. The performance of the deep learning model is found to be better in higher node configuration scenario.