Localization Using Neural Networks in Wireless Sensor Networks

Abstract

Noisy distance measurements are a pervasive problem in localization in wireless sensor networks. Neural networks are not commonly used in localization, however, our experiments in this paper indicate neural networks are a viable option for solving localization problems. In this paper we qualitatively compare the performance of three different families of neural networks: Multi-Layer Perceptron (MLP), Radial Basis Function (RBF), and Recurrent Neural Networks (RNN). The performance of these networks will also be compared with two variants of the Kalman Filter which are traditionally used for localization. The resource requirements in term of computational and memory resources will also be compared. In this paper, we show that the RBF neural network has the best accuracy in localizing, however it also has the worst computational and memory resource requirements. The MLP neural network, on the other hand, has the best computational and memory resource requirements.

Localization Using Radial Basis Function Networks and Signal Strength Fingerprints in WLAN

“…Subsequently, a location estimate is obtained when a new fingerprint s is presented to the network inputs. Recently, RBF networks have been discussed for indoor localization in Wireless Sensor Networks (WSN) by utilizing distance measurements [9] or series of successive RSS fingerprints [10]. Positioning techniques based on ANNs have also been applied to areas where WLAN infrastructure is available.…”

Wi-Fi-Based Localization in Dynamic Indoor Environment Using a Dynamic Neural Network

“…The main challenge to the techniques based on location fingerprinting is that the received signal strength could be affected by diffraction, reflection, and scattering in the propagation indoor environments. There are at least five location fingerprinting-based positioning algorithms using pattern recognition technique so far: probabilistic methods [3], k-nearest-neighbor (kNN) [3], [4], neural networks [5]- [7], support vector machine (SVM) [8], and smallest M-vertex polygon (SMP) [9].…”

Wlan–Based Indoor Localization Using Neural Networks

“…A neural network is composed of multiple neurons arranged in a set of layers; each neuron in a given layer is connected to every neuron in the successive layer through adjustable weights that work together to approximate a function [20]. In its training/offline phase an ANN learns the non-linear function between observed RSS values (inputs) and the geographical locations (target outputs) of the measurement, by adjusting its weights to minimize the error between the ANNs training outputs and its target outputs [21]. Once the network has been trained, it can estimate the unknown location of user equipment (UE) in the online phase by using RSS measurements only.…”

“…In constructing the ANN architecture for indoor localization, we have adopted the approach of [2], [20], [21], [22] [23] to use a feed-forward multi-layer perceptron architecture because it presents a good tradeoff between accuracy and memory requirements; the latter is desirable in view of memory and computational constraints of many handheld smart devices. We consider a threelayered MLP architecture, ie , a single hidden layer in addition to the input and output layers.…”