International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 10 Issue: 08 | Aug 2023
p-ISSN: 2395-0072
www.irjet.net
LIGHTWEIGHT INDOOR POSITIONING IN A REAL ENVIRONMENT BASED ON WIFI FINGERPRINTING TECHNIQUE AND M-WKNN ALGORITHM Jean Ndoumbe1*, Soubiel-Noel Nkomo Biloo1 and Agnès Enangue1 1Laboratory of Energy, Materials, Modeling and Method (E3M), National Higher Polytechnic
School of Douala, University Douala, Cameroon ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - The aim of this work is to study a lightweight
Arrival (TOA), Time Difference Of Arrival (TDOA), Frequency Difference Of Arrival (FDOA), Angle Of Arrival (AOA), Power Difference Of Arrival (PDOA), Frequency Difference of Arrival (FDOA) and Differential Doppler Rate (DDR). Depending on the degree of criticality of the information provided by the position of the object to be geolocated, the choice of one technique or another is open. In recent years, the aim has no longer been simply to geolocate an object or provide it with a means of navigation, but to find the best compromise between the complexity of implementing the hardware and software, the cost of the process and the results obtained, given that the results depend on all these parameters.
Key Words: indoor localisation, fingerprinting technique, power levels, M-WKNN algorithm, residual method.
Several works studying different algorithms for indoor positioning have been done. Zhou et al. [10] proposed an RSS transform–based weighted k-nearest neighbor (WKNN) indoor positioning algorithm to improve the positioning accuracy and real-time performance of Wi-Fi fingerprint–based indoor positioning. From the results, the proposed approach achieved a localization error of 1.52 m. Hu et al. [11], studied a self-adaptive WKNN algorithm with a dynamic K for indoor positioning based on Wi-Fi. In [12], the research leverages the use of a deep neural network-based ensemble classifier to perform indoor localization with heterogeneous devices. Among many indoor position calculation algorithms, those using machine learning are most studied such as deep learning and neural networks (DNN and MLNN) [12-14]. Works carried out in indoor positioning system [14-16] generally deals with near-perfect environments, free of human traffic (empty corridors, houses with no furniture). With such perfect process, the results can easily be duplicated. However, these results are far from reflecting the reality of locating or tracking an object. The presence of an object in the path of the wave amplifies the multipath phenomenon, reduces the power picked up and increases the propagation time [8]. The presence of a human body randomly distorts the signal by reducing its energy [17].
system for predicting indoor location in an ordinary dwelling using the fingerprinting technique and the MWKNN (Modified Weighted K-Nearest Neighbors) algorithm based on listening to the power levels of Wi-Fi terminals. To this end, an error reduction procedure has been developed and implemented in the M-WKNN algorithm. The residual calculation technique and sample weighting are used for this error reduction. In order to assess the performance of the proposed approach, a comparison is made with two traditional algorithms, KNN and WKNN. The results reported in this work evaluate the performance of the MWKNN algorithm at 92% with respect to the input data
1. INTRODUCTION Geolocation is essential today for many of the applications embedded in electronic devices. In fact, the technological boom that the world has been experiencing for the last two decades or so means that we need to have a perfect command and understanding of the space around us. Geolocation has two main applications, depending on the environment: outdoor geolocation and indoor geolocation. Today, because of the ever-improving results of satellite systems [1-2] used or dedicated to positioning, the scientific community is focusing mainly on the indoor environment, especially as some studies show that the human being spend most of the time in enclosed spaces [34] (indoor environment). The Global Positioning System and other satellite systems can no longer provide the same reliability here as they do outdoors. Firstly, it is difficult for waves to propagate and penetrate the indoor environment, and secondly, the multipath phenomenon that characterises this system in particular clouds the results and distorts them. Another factor is the reference frame, which can be physical, relative, absolute or symbolic [5-6]. To overcome these disadvantages, a number of indoor methods have been proposed, particularly those involving fingerprinting.
This work is a contribution to the search for the best compromise in smartphone positioning in a real indoor domestic environment. The first objective is to propose a lightweight and most effective method, and the second is to present the phenomena that reduce the efficiency of the used algorithm, working in real indoor location conditions. This work is based on an existing technique; the
To date, however, there is no standard method for either indoor or outdoor positioning [7-9]. In order to find the position of an RF device, different measurements can be used such as Received Signal Strength (RSS), Time Of
© 2023, IRJET
|
Impact Factor value: 8.226
|
ISO 9001:2008 Certified Journal
|
Page 650