Combing multiple linear regression and manifold regularization for indoor positioning from unique radio signal

Traditional learning methods for indoor positioning are based on a multitude of wireless radio signals synchronously, while only one or two Access Points (APs) can be perpetually and steadily received by users in the real-world indoor environment. In this paper, we propose a novel indoor positioning method by two aspects. On the one hand, we establish multiple linear regression to estimate the Euclidean distance between reference AP and mobile terminals. On the other, we propose manifold regularization approach to predict the intersection angle drew from reference baseline. Experimental results show that our proposed method achieves an acceptable and effective room-level precision using unique radio signal in the deployed indoor test-bed.