Elsevier

Neurocomputing

Volume 338, 21 April 2019, Pages 372-380
Neurocomputing

Automatic wireless mapping and tracking system for indoor location

https://doi.org/10.1016/j.neucom.2018.07.084Get rights and content

Abstract

Automatic vehicle tracking systems ease the completion of numerous tasks in different fields. Moreover they can automatically capture information, this feature allows to perform location tasks. These systems can be implemented at airports, in shopping centers and in other large buildings; in this way, wireless network scans will serve as a basis for the creation of signal maps that can be used in indoor location systems. This work proposes an automatic people tracking system which also allows to map Wi-Fi networks in order to localize people indoor. In order to operate the system, information on vehicle movement was used to capture signal maps, with the aim of reducing the need to perform manual calibration and thus, improving the updating of information. The final location is determined by combining information provided by wireless networks, Bayesian networks are employed for this task.

Introduction

In the last years, considerable developments have been made in the field of indoor location systems; they have become more precise and can now locate users in real-time. These advances have allowed to leverage location systems in different fields; they can be found in case studies in the areas of medicine [1], [2], [10], employee monitoring [3], robots [4], [5] etc. It is necessary to consider both the technology and the algorithms that can be used in location systems. The current trend in location systems is that they should operate, not one, but various technologies, in order to be able to apply information fusion techniques in the final calculation of a users’ position. The purpose of this work is to create a system that will allow to combine the information provided by the different location mechanisms. Moreover, it intends to facilitate the process of location by means of the automatic mapping of signal levels.

Advances in system location have include both hardware and software. In the hardware part, systems with different technologies have been developed, such as Wi-Fi, RFID, Bluetooth, ZigBee or the analysis of electromagnetic fields. These systems have often been combined with other technologies, such as inertial systems. The main problem presented by these systems is low precision, in order to improve it, new algorithms based on fingerprint [12], [16] have been applied. These algorithms have been chosen because they allow to locate users more precisely on the basis of the measurement of changes in the signal levels [12], [20], [21]. The main problem presented by the fingerprint technique is the need to make calibrations of the environment; this means that before the system can be used, a lot of time and efforts are required in order to maintain the information on signal levels updated [12].

This work proposes a multi-agent system which allows to locate devices in indoor spaces by operating iBeacon and Wi-Fi signals. The location system integrates iBeacon in order to determine scanning points, these scanning points are prefixed beforehand as the key points for calibration. At the points at which Bluetooth is detected, automatic scanning of the levels of Wi-Fi and iBeacon signals is carried out with the aim of creating a signal map and in order to make the use of fingerprint feasible. Moreover, the information on signal levels obtained from the iBeacons will be used in order to establish the aisle in the supermarket at which the user is located. After determining the aisle, the device is placed inside it thanks to the measurements made. Once we have these measurements, an application of the Bayesian signal distribution model is performed, obtaining a series of probabilities of stay at a calibration point. The system uses these probabilities to triangulate and calculate the final position.

In addition, time series will be applied in order to reduce the oscillation in the location of users. The proposed system has been used in a supermarket, for this purpose a shopping trolley was developed; the trolley tracks the users path and also guides the users to the place they want to reach. Without leaving out the traditional functions of this device, it is an element that is perfectly integrated with its environment, capable of detecting users and their gestures, this allows the trolley to always follow the correct user. The trolley independently follows the user and helps him in his shopping tasks. Moreover, it tells the system where the user is located at all times. All the movements made by the trolley are calculated autonomously, with the data provided to the Tablet by diverse sensors. This tablet is in charge of calculating the movements; these calculations are passed on to the microcontrollers which change these values into the energy that is provided to the motors for movement. This article is structured in the following way: related work is reviewed in Section 2, Section 3 describes the proposal, Section 4 shows the case study and finally in Section 5 the results and conclusions obtained are outlined.

Section snippets

Heuristics applied to optimization

Currently, location systems are widespread and their function varies depending on the case study. Outdoor location systems are usually based on the use of the Global Navigation Satellite System (GNSS), such as GPS, GLONASS, Galileo etc. [6], these systems often use different satellite networks to establish the final position. In initial works, we can see how sensor networks, consisting of beacons and tags, are deployed in order to locate objects in enclosed spaces. Along with the evolution of

Proposal

The Project was carried out using a multi-agent system which allows to control both the hardware of the trolley as well as the functionality of the location system. A multi-agent system has been chosen due to the possibilities it gives when including new functionalities dynamically. Fig. 1 shows the multi-agent system with its organizations. The trolley organization is in charge of operating the vehicle and allows for its movement, it includes the following agents for this purpose: The User

Case study

The prototype made for the case study is equipped with a Microsoft Kinect 360 camera, which allows to capture the position of the user and to obtain a cloud of points which is used to detect obstacles that are within its reach. In order to detect the obstacles that are not within the camera's view, HC-SR04 distance sensors have been used, these are installed on the sides of the back of the device, given that there are more probabilities that the trolley will crash with these sides when making a

Results and conclusions

The system evaluation method consists of mapping and locating an area of a supermarket in order to analyze the performance of the system. In the tests, different phases have been carried out to verify the operation of the different characteristics. Some features, such as user synchronization or motion control using a PID controller, are evaluated using heuristic techniques and progressive refinements are made. However, in other areas such as position calculation, an extensive statistical

Acknowledgments

This work has been supported by project Diseño y desarrollo de un carro autónomo de transporte de mercancías para la asistencia a personas con movilidad reducida en centros comerciales. Project co-financed with Junta de Castilla y León y Fondo Europeo de Desarrollo Regional (FEDER) funds. Fundación General de la Universidad de Salamanca PC_TCUE15-17_F2_027.

André Sales Mendes PhD Student on Computer Engineer. He is a researcher at the BISITE Research Group (http://bisite.usal.es). He has graduated in Computer Engineering at the University of Salamanca. In the same university he also has obtained a master's degree in Intelligent Systems. During the past years I have won several awards and scholarships related to innovation and artificial intelligence.

References (21)

There are more references available in the full text version of this article.

Cited by (8)

View all citing articles on Scopus

André Sales Mendes PhD Student on Computer Engineer. He is a researcher at the BISITE Research Group (http://bisite.usal.es). He has graduated in Computer Engineering at the University of Salamanca. In the same university he also has obtained a master's degree in Intelligent Systems. During the past years I have won several awards and scholarships related to innovation and artificial intelligence.

Gabriel Villarrubia González PhD Student on Computer Engineer. Currently, he is Assistant Professor at the Department of Computer Science at the University of Salamanca and member of the BISITE Research group. He obtained a Technical Engineering in Management of Computer in 2009 at the Pontifical University of Salamanca, then an Engineering in Computer Sciences degree in 2010 at the same university and postgraduate in Intelligent Systems at the University of Salamanca in 2011. He is co-author of more than 50 papers published in recognized journal, workshops and symposiums, most of them related to Intelligent Systems, Ambient Intelligence, and Distributed Systems. He has been member of the organizing and scientific committee of several international symposiums such as FUSION, PAAMS, MIS4TEL, ISAMI, PACBB.

Javier Caridad Hernández He is finishing his degree in Physical Sciences. He has developed his professional career developing sensors and electronic devices for industry and IoT applications.

Daniel Hernández de la Iglesia (PhD candidate) He has finished his studies in Technical Engineering in Computer Systems at the University of Salamanca (2013), and he has a degree in Computer Engineering from the same University (2014). He completed his studies with the Master in Intelligent Systems of the University of Salamanca (2015). He is currently a member of the computer and automatic department as a PhD student.

Juan Francisco De Paz (PhD.) Received a PhD in Computer Science from the University of Salamanca (Spain) in 2010. He is Assistant Professor at the University of Salamanca and researcher at the BISITE research group (). He obtained a Technical Engineering in Systems Computer Sciences degree in 2003, an Engineering in Computer Sciences degree in 2005 at the University of Salamanca and Statistic degree in 2007 in the same University. He has been co-author of published papers in several journals, workshops and symposiums.

View full text