The radio frequency spectrum is not being efficiently utilized partly due to the current policy of allocating the frequency bands to specific services and users. In opportunistic spectrum access (OSA), the ``white spaces'' that are not occupied by primary users (a.k.a. incumbent users) can be opportunistically utilized by secondary users. To achieve this, we need to solve two problems: (i) primary-secondary incumbent protection, i.e., prevention of harmful interference from secondary users to primary users; (ii) secondary-secondary network coexistence, i.e., mitigation of mutual interference among secondary users. The first problem has been addressed by spectrum sensing techniques in cognitive radio (CR) networks and geolocation database services in database-driven spectrum sharing. The second problem is the main focus of this dissertation. To obtain a clear picture of coexistence issues, we propose a taxonomy of heterogeneous coexistence mechanisms for shared spectrum access. Based on the taxonomy, we choose to focus on four typical coexistence scenarios in this dissertation. Firstly, we study sensing-based OSA, when secondary users are capable of employing the channel aggregation technique. However, channel aggregation is not always beneficial due to dynamic spectrum availability and limited radio capability. We propose a channel usage model to analyze the impact of both primary and secondary user behaviors on the efficiency of channel aggregation. Our simulation results show that user demands in both the frequency and time domains should be carefully chosen to minimize expected cumulative delay. Secondly, we study the coexistence of homogeneous CR networks, termed as self-coexistence, when co-channel networks do not rely on inter-network coordination. We propose an uplink soft frequency reuse technique to enable globally power-efficient and locally fair spectrum sharing. We frame the self-coexistence problem as a non-cooperative game, and design a local heuristic algorithm that achieves the Nash equilibrium in a distributed manner. Our simulation results show that the proposed technique is mostly near-optimal and improves self-coexistence in spectrum utilization, power consumption, and intra-cell fairness. Thirdly, we study the coexistence of heterogeneous CR networks, when co-channel networks use different air interface standards. We propose a credit-token-based spectrum etiquette framework that enables spectrum sharing via inter-network coordination. Specifically, we propose a game-auction coexistence framework, and prove that the framework is stable. Our simulation results show that the proposed framework always converges to a near-optimal distributed solution and improves coexistence fairness and spectrum utilization. Fourthly, we study database-driven OSA, when secondary users are mobile. The use of geolocation databases is inadequate in supporting location-aided spectrum sharing if the users are mobile. We propose a probabilistic coexistence framework that supports mobile users by locally adapting their location uncertainty levels in order to find an appropriate trade-off between interference mitigation effectiveness and location update cost. Our simulation results show that the proposed framework can determine and adapt the database query intervals of mobile users to achieve near-optimal interference mitigation with minimal location updates. / Ph. D.

With the coming growth in Internet of Things (IoT) applications, we can expect environ-ments with many independent networks operating in nearby locations. Wireless Sensor Networks (WSN), which have become popular during the last few years, are the main type of networks used in IoT. The IEEE 802.15.4 protocol designed for low-rate wireless personal area networks has been widely adopted for this kind of network. Together with ZigBee, this protocol is gaining increasing interest from the industry, as they are con-sidered a universal solution for low-cost, low-power, wireless connected monitoring and control devices. Internetwork interference issues in IEEE 802.15.4 networks can be a ma-jor problem because of the extensive use of wireless channels. In this thesis, an in-depth simulation study of the internetwork interferences is performed using Castalia, a widely used network simulator. We focus on the beacon collision problem, as it has been proved to be the main cause of performance degradation for coexisting networks. We carry out a prestudy of the main node simulation parameters to setup the different scenarios. Then, we evaluate how the overlap of the active periods and the location of the nodes affect the network performance. We continue with a network coexistence analysis to study the inter-action of two networks of two nodes and their performance regarding the beacon reception rate. We show that there are significantly different operation regions, depending on the network location. Following this, a probabilistic analysis is carried out in order to obtain an average beacon reception rate depending on the size of the area considered. Finally, we discuss available beacon collisions avoidance methods, taking into account the detailed simulation results. Our conclusions have theoretical and practical implications for the design of wireless sensor networks, and for the evaluation of beacon collisions avoidance schemes. / Med den kommande tillväxten i sakernas Internet (IoT) applikationer är miljöer med många oberoende nätverk som verkar på närliggande ställen väntat. Trådlösa sensornätverk (WSN), som har blivit populära under de senaste åren, är den vanligaste typen av nät som används i sakernas Internet. IEEE 802.15.4 protokollen, konstruerad för låghastighet trådlösa personlig area nätverk, har fått stor spridning för WSNs. Tillsammans med ZigBee, får de en snabb ökat intresse från industrin, eftersom de betraktas som en universallösning för låg kostnad låg energi trådlös anslutning för övervaknings- och kontrollinstrument. Internätverk interferens i IEEE 802.15.4 nätverk kan vara ett stort problem på grund av den omfattande användningen av trådlösa kanaler. I denna avhandling är en djupgående simulation studie utfört med hjälp av Castalia, ett utbrett använt nätverk simulator. Vi fokuserar på beacon kollisionsproblem, eftersom det har visat sig vara den främsta orsaken till prestandaförsämring för samexisterande nätverk. Vi utför en förstudie av den viktigaste simulation parametrarna för att bestemma de olika scenarierna. Sedan utvärderar vi hur överlappningen av de aktiva perioderna och placeringen av noderna påverkar nätverkets prestanda. Vi fortsätter med en nätverksamexistens analys för att studera interaktionen mellan två nätverk av två noder, och deras prestanda avseende beacon mottagnings kvot. Vi visar att det finns betydligt olika operationsområdena, beroende på nätverksens placering. Därefter är en sannolikhetsanalys utfört för att erhålla en genomsnittlig beacon mottagnings kvot, beroende på storleken på betraktade områden. Slutligen diskuterar vi tillgängliga metoder för att undvika beacon kollision, med hänsyn till de detaljerade simuleringsresultaten. Våra slutsatser har teoretiska och praktiska kon-sekvenser för utformningen av trådlösa sensornätverk, och för utvärderingen a metoder för att undvika beacon kollision.