Deep Neural Networks (DNNs) have been proposed to solve many difficult problems within the context of wireless sensing. Indoor localization and human activity recognition (HAR) are two major applications of wireless sensing. However, current fingerprint-based localization methods require massive amounts of labeled data and suffer severe performance degradation in NLOS environments. To address this challenge, we first apply DNNs to multi-modal wireless signals, including Wi-Fi, an inertial measurement unit (IMU), and ultra-wideband (UWB). By formulating localization as a multi-modal sequence regression problem, a multi- stream recurrent fusion method is developed to combine the current hidden state of each modality. This is done in the context of recurrent neural networks while accounting for the modality uncertainty directly learned from its immediate past states. The proposed method was evaluated on a large-scale open dataset and compared with a wide range of baseline methods. It is shown that the proposed approach has an average error below 20 centimeters, which is nearly three times better than classic methods. Second, in the context of activity recognition, we propose a multi-band WiFi fusion frame- work that hierarchically combines the features of sub-6 GHz channel state information (CSI) and the beam signal-to-noise ratio (SNR) at 60 GHz at different granularity levels. Specifically, we introduce three fusion methods: simple input fusion, feature fusion, and a more customized feature permutation that accounts for the granularity correspondence between the CSI and beam SNR measurements for task-specific sensing. To mitigate the problem of limited labeled training data, we further propose an autoencoder-based unsupervised fusion network consisting of separate encoders and decoders for the CSI and beam SNR. The effectiveness of the framework is thoroughly validated using an in-house experimental platform which includes indoor localization, pose recognition, and occupancy sensing. Finally, in the context of array processing, we solve the Model order estimation (MOE) problem, a prerequisite for Direction of Arrival (DoA) estimation in the presence of correlated multipath, a well-known difficult problem. Due to the limits imposed by array geometry, it is not possible to estimate spatial parameters for an arbitrary number of sources; an estimate of the signal model is required. While classic methods fail at MOE in the presence of correlated multi-path interference, we show that data-driven supervised learning models can meet this challenge. In particular, we propose the application of Residual Neural Net- works (ResNets), with grouped symmetric kernel filters to provide an accuracy over 95%, and a weighted loss function to eliminate the underestimation error of model order. The improved MOE is shown improve subsequent array processing tasks such as reducing the overhead needed for temporal smoothing, reducing the search space for signal association, and improving DoA estimation. / Doctor of Philosophy / Radio Frequency (RF) signals are used not only for wireless communication (its most well-known application), but is also commonly used to sense the environment. One specific application, localization and navigation, can require accuracy of 0.5 meters or below, which is a significant challenge indoors. To address this problem, we apply deep learning (a technique that has gains significant attention in recent years) to fuse types of RF signals, including signals and devices commonly used in smart phones (e.g., UWB, WiFi and IMUs). The result is a technique that can achieve 20cm accuracy in indoor location applications. In addition to localization, commercial WiFi signals can also be used to sense/determine human activity. The received signals from a WiFi transmitter contain sensing information about the environment, including geometric information (angles, distance and velocity) about objects. We specifically show that our proposed approach can successfully recognize human pose, whether or not a specific seat is occupied, and a person's location. Moreover, we show that this can be done with relatively little labelled data using a technique known as transfer learning. Finally, we apply the another neural network structure to solve a particular problem in multi-antenna processing, model order estimation in the presence of coherent multipath. The resulting system can deliver a 95% accuracy in complex environments greatly improving overall array processing.

WiFi is becoming common in households and digital devices needs to support it. At the same time the devices are getting smaller and the Ethernet port may seem superfluous. When testing these devices the test environment needs to be able to provide WiFi connectivity. The tests may be focused on testing WiFi but it could also be the only network connectivity and thus needs to be very reliable. With a large number of devices in a small physical area a normal WiFi setup would have a density of devices that is too high for today’s1 WiFi standards. A combination of wired physical medium and physical isolation was considered.

Le trafic de données mobiles augmente de façon permanente. Afin d'éviter une saturation, les opérateurs doivent décharger le réseau cellulaire vers des réseaux d'accès alternatifs. WiFi se trouve être une bonne solution qui permet à l'opérateur de tirer parti de bandes de fréquence sans licence ainsi que du très grand nombre de points d'accès déjà déployés. Dans cette thèse, nous présentons tout d'abord un état de l'art des différentes solutions de couplage entre LTE et WiFi. Nous montrons que la plupart de ces solutions ne garantissent pas la continuité de session ou sont sujettes à une duplication des procédures de sécurité. Ceci a conduit à la proposition du Very Tight Coupling entre LTE et WiFi. Dans ce type d'architecture, les points d'accès WiFi sont connectés à une station de base LTE et les mécanismes de sécurité LTE sont réutilisées afin de permettre un accès rapide au réseau WiFi. Ceci permet également d'avoir une double connectivité et de garder le trafic de signalisation sur le réseau LTE, ce qui donne la possibilité d'avoir des procédures de sélection optimisées. Nous étudions comment le Very Tight Coupling peut être implémenté et comment les points d'accès WiFi intégrés dans les passerelles résidentielles peuvent être connectées aux stations de base LTE dans le cas d'un réseau fixe/cellulaire convergent. Nous évaluons ensuite par des outils mathématiques, les performances de différents schémas de couplage et calculons le taux de capacité pouvant être économisée. Ensuite, nous présentons une implémentation du Very Tight Coupling sur une plateforme utilisant une interface radio LTE réelle basée sur Open Air Interface. Nous effectuons plusieurs expérimentations afin de trouver la meilleure configuration du protocole de la couche liaison de données. Nous démontrons que le fait d'utiliser WiFi et LTE en parallèle n'augmente pas systématiquement le débit. / The mobile data traffic has been continuously increasing. To avoid saturation of cellular network, operators need to use alternative access networks for offloading purpose. WiFi is a good solution as the operator can take advantage of its unlicensed spectrum as well as the large number of deployed WiFi access points.In this thesis, we first provide a state-of-the-art of the different coupling solutions between LTE and WiFi. We show that most solutions cannot guarantee session continuity or duplicate the security procedures. This leads to propose "Very Tight Coupling" between LTE and WiFi. In this architecture, WiFi access points are connected to the LTE base stations and the security mechanisms of LTE are reused to ensure fast access to WiFi. It allows dual connectivity and to keep control signalling in the LTE network, which gives the possibility to have optimized interface selection procedures.We study how very tight coupling can be implemented and how WiFi APs that integrated in customer residential gateways can be connected to LTE base stations in a converged fixed/cellular network. We then mathematically evaluate the performance of different deployment schemes and compute how much capacity can be saved on the LTE network. Furthermore, we implement the solution on a platform with a real LTE radio interface based on the Open Air Interface framework as a proof-of-concept. We perform several experiments to find the configuration of the link-layer protocols that gives the highest bit rate. In particular, we show that using WiFi and LTE simultaneously does not always increase the bit rate.

No description available.

Förstudien behandlar den tekniska delen av ett teoretiskt stadstäckande öppet WiFi i Trollhättan, Sverige. Trollhättans kommun beställde en förstudie från Högskolan Väst som skulle agera beslutsunderlag till hur de skulle ta ärendet vidare. Studien är begränsad till den tekniska biten, där bland annat val av utrustning, trådlösa tekniker och riskanalys av nätverket ingår. Arbetet varit väldigt teoretiskt inriktat med en metod som till största del består utav att samla information och kunskaper från vetenskapliga rapporter, teknikorienterade hemsidor och andra relaterade arbeten.De trådlösa standarder som jämfördes teoretiskt i förstudien var 802.11g och 802.11n. 802.11n valdes som den standarden som bäst passade just detta nätverk, främst för dess positiva egenskaper när det gäller täckning, hastighet och bakåtkompatibilitet. 802.11n seräven ut att garantera en säker framtid när det gäller implementering i mobila enheter. Föreslaget på utrustning består av två olika accesspunkter, antingen 2602e eller 3602e beroende på den slutgiltiga budgeten. Dessa produkter är från Cisco och klarade olikastresstester med bra resultat samtidigt som de stödde den föreslagna standarden 802.11n. WLAN controllern (Ciscos 5500-serie) som valdes ut stödjer det krav på antal accesspunkter som bestämts, samtidigt som den lämpar sig bra för eventuella framtidakrav. Routern till detta nätverk är en Cisco 891-router. Anledningen till att denna router valdes var för att den har 4 portar som kan ge ström över Ethernet till accesspunkterna. Säkerhetsanalysen gjordes med hjälp utav CORAS-metoden där det diskuteras vilka riskernätverket kan ställas inför och vad dessa risker kan få för konsekvenser. Försök till att hitta lösningar som var direkt applicerbara på det stadstäckande öppna nätverket gjordes även.Förstudien avslutas med en kostnadsberäkning av nätverket, baserad på andras tidigare arbeten och deras kostnader, speciellt från Jönköping, Helsingborg och Mountain View.

This thesis aims to draft technology and its practical verification for online monitoring of people in closed buildings (hospitals, hospices, ...). The technology will be based on WiFi triangulation and evaluated self-designed device. In the practical part will be tested reliability, accuracy and economic comparison with alternative technologies.

Objective of this master thesis solves possible way of WiFi Honeypot realisation, which is constructed to detecet malicious network activity and attacks in radio environment that uses a set of IEEE 802.11 protocols. A specific configuration was created on the mITX format motherboard and contains scripts and software for data collection, analysis and its evaluation. Based on information and knowledge about specific network attacks it is possible to identify data traffic leading to anomalies and detect possible network attack. The final device was tested in real use for long-term data collection and evaluation of network activity in the given location. This fulfills the main goal of this work, which is implementation of WiFi Honeypot with support for IEEE 802.11 protocols and with possible deployment for real use.

No / A novel printed diversity monopole antenna is presented for WiFi/WiMAX applications. The antenna comprises two crescent shaped radiators placed symmetrically with respect to a defected ground plane and a wideband zigzag neutralization line is connected between them to achieve good impedance matching and low mutual coupling. Theoretical and experimental characteristics are illustrated for this antenna, which achieves an impedance bandwidth of 54.5% (over 2.4 – 4.2 GHz), at a reflection coefficient < −10 dB, mutual coupling < −16 dB. An acceptable agreement is obtained for the computed and measured gain, radiation patterns, and envelope correlation coefficient. These characteristics demonstrate that the proposed antenna is an attractive candidate for the multiple input multiple output (MIMO) portable mobile devices.

Les réseaux sans fil créent des changements importants dans le comportement des utilisateurs par rapport aux réseaux dits câblés ou filaires. Ainsi, les utilisateurs peuvent être mobiles, ce qui les amène à évoluer d’un point d’accès (PA) à un autre. Lors du changement de PA, l’utilisateur mobile (UM) peut expérimenter des délais de transmission occasionnés par les étapes de recherches du prochain PA. Un délai de quelques dixièmes de secondes est désagréable lors d’utilisation d’applications à flux constant de données (comme la téléphonie IP). En général, les gens utilisent souvent les mêmes trajets pour les déplacements courants, comme aller au travail ou faire les courses. L’objectif de ce travail de recherche est de démontrer qu’il est possible d’ajouter des algorithmes à l’appareil mobile pour lui permettre, de façon autonome, d’apprendre les habitudes de l’UM pour améliorer la sélection du prochain PA. En utilisant les temps libres du module de communication WiFi pour sonder l’environnement et accumuler des informations sur les habitudes, trois algorithmes ont été développés et sont présentés dans ce travail de recherche. Le premier algorithme, « Optimisation par courbes comparatives de Puissance sans le Temps » (OPT), mémorise les variations des niveaux de puissances des signaux WiFi. En couplant les mesures de variations, nous avons développé un mécanisme qui détermine les meilleures paires de PA pour les transitions. OPT donne de bonnes performances lorsque l’UM peut se relier à tous les PA qu’il rencontre, ce qui n’est pas le cas lorsque certains PA sont sécurisés. Une deuxième approche a été testée en utilisant un mécanisme d’apprentissage par renforcement (nommé APP). Cet algorithme permet de catégoriser les PA et permet de déterminer les meilleurs candidats aux transferts. En environnement sécurisé, cet algorithme présente des performances nettement supérieures à OPT. Par contre, dans un environnement présentant beaucoup de PA, le temps d’apprentissage est non négligeable puisque APP doit tenter de se relier à tous les PA accessibles pour les caractériser. Pour réduire la période d’apprentissage d’APP, nous avons combiné les deux algorithmes pour mettre en commun leurs avantages. Cette troisième solution (nommée APPOPT) a produit un mécanisme de transfert intercellulaire efficient pour la qualité de sélection du prochain PA comparativement au mécanisme standard de transfert du WiFi.

The global mobile data usage has grown nearly 70\% annually in recent years. The huge mobile data usage requirement drives the mobile industry to brace the formidable challenge and invent next-generation mobile technologies. LTE, as a successful cellular technology, has gained tremendous importance in recent years due to its high data-rates and improved data access method for mobile devices. Even though LTE still may not be able to meet the mobile data challenge due to current spectrum scarcity in licensed bands. Thus, cellular network faces serious challenges to provide high performance mobile service to end users in the near future. In order to sustain the possible increase in mobile capacity demand, utilizing the unlicensed band as a supplementary band for LTE is being considered as a promising solution to expand the capacity of mobile systems. Based on the innovation of carrier aggregation, 3GPP has approved a study item on LAA-LTE, which will assist LTE by offloading mobile data in unlicensed band. Thus, LAA-LTE will operate in the spectrum that overlaps with WiFi, which is another popular unlicensed band technology. The concern is that LAA-LTE and WiFi are unlikely to have mechanisms to directly coordinate with each other, considering different core networks, backhauls and deployment plans of LAA-LTE and WiFi networks. The overarching goal of my research is to investigate the following two aspects: 1) Investigate how LTE will impact on WiFi using experimental analysis when both of them share the same channel, 2) Develop a possible coexistence algorithm to trigger the coexistence between LAA-LTE and WiFi in unlicensed band.