Due to the emerging demand for Internet of Things (IoT) applications, indoor positioning has become an invaluable task. We propose NDR, a novel lightweight deep learning solution to the indoor positioning problem. NDR is based on Noise and Dimensionality Reduction of Channel State Information (CSI) of a Multiple-Input Multiple-Output (MIMO) antenna. Based on preliminary data analysis, the magnitude of the CSI is selected as the input feature for a Multilayer Perceptron (MLP) neural network. Polynomial regression is then applied to batches of data points to filter noise and reduce input dimensionality by a factor of 14. The MLP's hyperparameters are empirically tuned to achieve the highest accuracy. NDR is compared with a stateof-the-art method presented by the authors who designed the MIMO antenna used to generate the dataset. NDR yields a mean error 8 times less than that of its counterpart. We conclude that the arithmetic mean and standard deviation misrepresent the results since the errors follow a log-normal distribution. The mean of the log error distribution of our method translates to a mean error as low as 1.5 cm.

Indoor localization has attracted much attention due to its indispensable applications e.g. autonomous driving, Internet-Of-Things (IOT), and routing, etc. Received Signal Strength Indicator (RSSI) was used intensively to achieve localization. However, due to its temporal instability, the focus has shifted towards the use of Channel State Information (CSI) aka channel response. In this paper, we propose a deep learning solution for the indoor localization problem using the CSI of a 2 × 8 Multiple Input Multiple Output (MIMO) antenna. The variation of the magnitude component of the CSI is chosen as the input for a multilayer Perceptron (MLP) neural network. Data augmentation is used to improve the learning process. Finally, various MLP neural networks are constructed using different portions of the training set and different hyperparameters. An ensemble neural network technique is then used to process the predictions of the MLPs in order to enhance the position estimation. Our method is compared with two other deep learning solutions; one that uses Convolutional Neural Network (CNN), and the other uses MLP. The proposed method yields higher accuracy than its counterparts.

Indoor Localization has attracted interest in both academia and industry for its wide range of applications. In this paper, we propose an indoor localization solution based on Channel State Information (CSI). CSI is a fine-grain measure of the effect of the channel on the transmitted signal. It is computed for each subcarrier and each antenna in the Multiple-Input-Multiple-Output (MIMO) antenna case. It is also becoming a trend for indoor position fingerprinting. By using a K-nearest neighbor learning method a highly accurate indoor positioning is achieved. The input feature is the magnitude component of CSI which is preprocessed to reduce noise and allow for a quicker search. The euclidean distance between CSI is the criteria chosen for measuring the closeness between samples. The method is applied to a CSI dataset estimated at an 8 × 2 MIMO antenna that is published by the organizers of the Communication Theory Workshop Indoor Positioning Competition. The proposed method is compared with three other methods all based on deep learning approaches and tested with the same dataset. The Knearest neighbor method presented in this paper achieves a Mean Square Error (MSE) of 2.4 cm which outperforms its counterparts.

Localization is the process of determining the position of an entity in a given coordinate system. Due to its wide range of applications (e.g. autonomous driving, Internet-of-Things), it has gained much focus from the industry and academia. Channel State Information (CSI) has overtaken Received Signal Strength Indicator (RSSI) to achieve localization given its temporal stability and rich information. In this paper, we extend our previous work by combining classical and deep learning methods in an attempt to improve the localization accuracy using CSI. We then test the generalization aspect of both approaches in different environments by splitting the training and test sets such that their intersection is reduced when compared with uniform random splitting. The deep learning approach is a Multi Layer Perceptron Neural Network (MLP NN) and the classical machine learning method is based on K-nearest neighbors (KNN). The estimation results of both approaches outperform state-of-the-art performance on the same dataset. We illustrate that while the accuracy of both approaches deteriorates when tested for generalization, deep learning exhibits a higher potential to perform better beyond the training set. This conclusion supports recent state-of-the-art attempts to understand the behaviour of deep learning models.