H. Autonomic network modeling and optimization


    H.1 Distributed resource allocation and incentives in peer-to-peer networks

    Peer-to-peer networks constitute the main representative of the class of autonomous networks that already start dominating Internet traffic. Peer-to-peer networks are voluntary resource sharing systems among rational agents that are resource providers and consumers. While altruistic resource sharing is necessary for efficient operation, this can only be imposed by incentive mechanisms, otherwise peers tend to behave selfishly. Selfishness in general terms means only consuming resources in order to absorb maximal utility from them and not providing resources to other peers because this would require effort and would not give any utility. We consider networks where the access technology does not separate upstream and downstream traffic, and these flow through the same capacity-limited access link. We examine such kinds of networks, because (i) they are very likely to proliferate, (ii) they include a clear distinction between the client (download) and server (upload) roles of a peer.


First, we consider a reputation-based mechanism for providing incentives to peers for resource provisioning besides resource consuming.  A separate utility maximization problem is solved by each peer, carried out under a constraint on the level of dissatisfaction the peer intends to cause by not fulfilling others’ requests. This parameter is private information for each peer. The reputation of a peer as a server is updated based on the amount of allocated bandwidth compared to the requested one. Reputation acts towards gradually revealing hidden intentions of

peers and accordingly guiding the resource allocation by rewarding or penalizing peers in subsequent bandwidth allocations.


Further, we consider a network of peers, each with different utility functions which are private information and capture a peer's selfishness or desire for content. The objectiveis to maximize the sum of utilities of peers. We answer the following questions in a peer-to-peer network: (a) what portions of its link capacity does each peer allocate to upload flows from other peers and download flows for itself? (b) How does a peer decide which portion of bandwidth will be allocated to each upload flow and download flow? (c) How can these decisions be taken in a decentralized autonomous fashion? The global link sharing problem of maximizing total utility is hard to solve in a distributed fashion because of coupled utility functions and constraints. That is, the utility of each peer depends on allocation decisions of others. By defining auxiliary variables and constraints, we transform the problem into one that is amenable to "distributization" by dual decomposition. Interestingly, the Lagrange multipliers corresponding to the newly added constraints are interpreted as reciprocals of pairwise reputation metrics. This leads us to a meaningful light-weight reputation-driven protocol with the desirable property that only the amounts of requested and granted bandwidth are circulated, and not reputations.


   Relevant publications:

   [J13] I. Koutsopoulos and G. Iosifidis, “A framework for distributed bandwidth allocation in peer-to-peer networks”, Performance Evaluation, 2010.

   [C24] I. Koutsopoulos and G. G. Iosifidis, ''Reputation-assisted utility maximization algorithms for peer-to-peer networks'', Proceedings of IEEE International Workshop on QoS (IWQoS) 2008, Twente, Netherlands.


   [C25] I. Koutsopoulos and G. Iosifidis, ''Distributed resource allocation algorithms for peer-to-peer networks'', Proceedings of ACM International Conference on Performance Evaluation and Tools (VALUETOOLS), 2008, Athens, Greece.

    H.2 Game theory models for peer-to-peer networks

We consider a content sharing network of non-cooperative peers. The strategy set of each peer comprises: (i) client strategies, namely feasible request load splits to servers, and (ii) server strategies, namely scheduling disciplines on requests. The performance metric is average retrieval delay.


First, we consider the request load splitting game for given server strategies such as First-In-First-Out or given absolute priority policies. A peer splits its request load to servers to optimize its performance objective. We consider the class of best response load splitting policies residing between the following extremes: a truly selfish, or egotistic one, where a peer optimizes its own delay, and a pseudo-selfish or altruistic one, where a peer also considers incurred delays to others. We derive conditions for Nash equilibrium points (NEPs) and discuss convergence to NEP and properties of the NEP. For both the egotistic cases, the NEP is unique. For the altruistic case, each of the multiple NEPs is an optimum, a global one for the FIFO case and a local one otherwise.


Next, we include scheduling in peer strategies. With its scheduling discipline, a peer cannot directly affect its delay, but it can affect the NEP \titafter peers play the load splitting game. The idea is that peer i should offer high priority to (and thus attract traffic from) higher-priority peers that cause large delay to i at other servers. We devise two-stage game models, where, at a first stage, a peer selects a scheduling rule in terms of a convex combination of absolute priorities, and subsequently peers play the load splitting game. In the most sophisticated rule, a peer selects a scheduling discipline that minimizes its delay at equilibrium, after peers play the load splitting game. We also suggest various heuristics for picking the scheduling discipline. Our models and results capture the dual client-server peer role and aim at quantifying the impact of selfish peer interaction on equilibria.

    Relevant publications:

   [C28] I. Koutsopoulos, L. Tassiulas and L. Gatzikis, “Client and server games in peer-to-peer networks”, Proceedings of IEEE International Workshop on QoS (IWQoS), 2009, Charleston, SC, USA.


        H.3. Auction modeling for autonomic networks


        In the autonomic Internet of the future, auction mechanisms arise as key methods for realizing efficient resource allocation. The major asset of auctions is their obliviousness to node utilities, which renders them capable of achieving a desired resource allocation regime without knowledge of the utility functions of involved entities. Auctions can aid in addressing major research challenges in such autonomic settings, such as the need to cope with diverse and conflicting interests of network entities, the need to carry out resource allocation in a decentralized manner, the requirement for matching dynamic spatiotemporal patterns of demand and supply, and the need to operate under limited or no network state and utility information. In this survey paper, we delineate the main trends and challenges associated with auction design. We start from first principles auction design for maximum auctioneer revenue or maximum allocation efficiency for one or multiple indivisible items and for divisible resources. We gradually move to more composite models, those of position auctions for Internet advertisements and those arising in spectrum sharing in cognitive radio networks. We argue that some directions worth pursuing are: (i) the design of advanced auction models that capture multi-level interaction of involved entities, (ii) the employment of double auctions for multiple seller and buyer interaction, and (iii) the design of decentralized negotiation and resource trading mechanisms.


      In autonomic networks with nodes with different needs and utility functions, each entity possesses some resource and can engage in transactions with others to achieve its needs. In fact, efficient network operation relies on node synergy and multi-lateral resource trading. Nodes face the dilemma of devoting their limited resource to their own benefit versus acting altruistically and anticipating to be aided in the future. Wireless ad-hoc networks, peer-to-peer networks and disruption-tolerant networks are instances of autonomic networks where the challenges above arise and the traded resource is energy, bandwidth and storage space respectively. Clearly, the decentralized complex node interactions and the double node role as resource provider and consumer amidst resource constraints cannot be addressed by single-sided auctions and even more by mechanisms with a central controller. We introduce a double-sided auction market framework to address the challenges above. Each node announces one bid for buying and one for selling the resource. We prove that there exist bidding and charging strategies that maximize social welfare and we explicitly compute them. We generalize our result to a generic network objective. Nodes are induced to follow these strategies otherwise they are isolated by the network. Furthermore, we propose a decentralized realization of the double-sided auction with lightweight network feedback. Finally, we introduce a pricing method which does not need a charging infrastructure. Simulation results verify the desirable properties of our approach.


      Relevant publications:

     [C34] I. Koutsopoulos and G. Iosifidis, “Auction mechanisms for Network Resource allocation”, Proceedings of Workshop on Resource Allocation in Wireless Networks (RAWNET),  (in WiOpt 2010), Avignon, France. (invited paper)


     [J14] G. Iosifidis and I. Koutsopoulos, “Double auction mechanisms for resource allocation in autonomous networks”, IEEE Journal on Selected Areas in Communications (Special Issue on Recent Advances on Autonomic Networks), vol.28, no.1, pp.95-102, Jan. 2010.


G. Wireless Network Security issues 


     G.1 Misbehavior detection issues in wireless networks


     The main characteristic of contemporary and envisioned wireless network designs is the easy re-programmability of protocol mechanisms even at run-time across all layers in wireless networks. Misbehavior is the manipulation of protocol parameters, leading to intentional deviation from protocol rules with the goal  to obtain a specific gain. We address the fundamental problem of reliable and fast detection of misbehavior.

    To capture attack uncertainty, we model attack policies by probability tools and focus on a class attacks of interest, based on induced damage by the attacker. We cast the problem within the framework of robust detection that captures the case of an intelligent attacker that can adapt its policy so as to avoid being detected.  With this approach, we establish fundamental performance limits of detection as far as detection delay and probability of detection are concerned. As a case study, we consider Medium Access layer protocol misbehavior by back-off parameter manipulation.  We derive analytic expressions for the worst attack in terms of detection delay and show that the class of sequential detection tests is optimal in the sense of minimizing detection delay.  We show how such a detection test can exploit the open wireless medium and collect observations until significant evidence exists. We extend the basic model to the situation where observations are hindered by interference due to concurrent transmissions. We also present some hints for notifying the rest of the network about the attack. The problem structure calls for a wealth of extensions as well as for an information-theoretic view that is currently under investigation. This will extend our understanding on fundamental performance limits, attack models and counter-measures.  


       Relevant publications:

     [J9] S. Radosavac, G. Moustakides, J.S. Baras and I. Koutsopoulos, "An analytical framework for modeling and detecting access layer misbehavior in wireless networks", ACM Transactions on Information Security, vol.11, no.4, pp.1-26, July 2008.

     [C20] S. Radosavac, J.S. Baras and I. Koutsopoulos, "A framework for MAC layer misbehavior detection in wireless networks", Proceedings of ACM Workshop on Wireless Security (WiSe)  2005, Cologne, Germany.

      G.2. Jamming attacks and defenses in wireless sensor networks


      Sophisticated jamming attacks on wireless sensor networks can have a more detrimental effect than brute-force ones due to the difficulty of being detected. Depending on the level of sophistication, the attacker can have a multi-dimensional attack space, ranging from probability of jamming, transmission power, directionality, multi-channel scanning and jamming and so on. Can the sensor network take a defense action in order to confront the attacker?

      We take a first step towards modeling a jamming attack and attempting to solve the issue of network defense. We consider a single-channel wireless sensor network that operates under a random channel access protocol, and is jammed by a sophisticated jammer. The jammer controls probability of jamming and transmission range and aims at causing maximal damage to the network in terms of corrupted communication links. We show how the jammer can be detected by employing an optimal detection test at a monitor node, based on percentage of incurred collisions. Following detection, we also model the fact that a notification message needs to be transferred out of the jammed area while the attack continues. We also show that it is actually feasible for the network to defend itself (to a certain extent) by adapting channel access probability in an effort to expose the jammer and minimize detection plus notification time.  In order for the jammer to select the optimal policy, it needs to know the network channel access probability and number of neighbors of the monitor node. The network needs to know the jamming probability of the jammer.

      We study the idealized case of perfect knowledge by both the jammer and the network about the strategy of each other and the case where the jammer or the network do not have this knowledge. For the former case, we derive analytic expressions for the optimal attack and defense strategies and show they can often be constrained due to energy limitations.  For the latter case, we formulate and solve optimization problems, the solutions of which give best responses of the attacker or the network to the worst-case strategy of each other. We extend the problem for the case of multiple monitor nodes. We also enrich the attack space by considering adaptable jamming transmission range. Here, we identify the following interesting tradeoff:  Monitors closer to the boundary experience "lighter" attack (hence larger detection delay) but transfer notification message out of the area in fewer expected number of hops.  Ongoing work will allow a better understanding of this class of attacks and of the interaction between the attacker and the network. 


      Relevant publications:

     [J12] M. Li, I. Koutsopoulos and R. Poovendran, “Optimal jamming attacks and network defense strategies in wireless sensor networks”, to appear, IEEE Transactions on  Mobile Computing, 2010

    [C22] M. Li, I. Koutsopoulos and R. Poovendran, "Optimal jamming attacks and defense strategies in wireless sensor networks", Proceedings of IEEE INFOCOM 2007, Anchorage, Alaska.

     G.3. Attacker Localization in Wireless Networks

   Following detection, the attacker needs to be localized. We exploit readily available and measurable quantities on wireless cards to design and implement an attacker localization algorithm. Our scheme is based on the principles of the gradient descent minimization algorithm. The key observation is that the Packet Delivery Ratio (PDR) has lower values as we move closer to the jammer since the receiver power level from the jammer increases. The result is lightweight gradient algorithm operating on the discrete node plane, according to which we are led to (at least) a local optimum, in the sense that we reach the node close enough to the attacker (jammer). We implement our algorithm on a wireless test-bed.

Relevant publications:

     [C30] K. Pelechrinis, I. Koutsopoulos, I. Broustis, and S. Krishnamurthy, “Lightweight Jammer Localization in wireless networks: System design and implementation”, Proceedings of IEEE Global Telecommunications Conference (GLOBECOM), 2009, Hawaii, USA.

F. Cross-layer (PHY/MAC) design and resource allocation issues in wireless networks


     F.1. Resource allocation and transmission adaptation techniques for OFDMA and CDMA networks


     We explore ways by which the performance of wireless networks can be enhanced by multi-layer interaction and information passing. Specifically, we focus on orchestrating mechanisms of the physical and the access layer with the objective to optimize system performance in terms of achievable transmission rate.  Cross-layer coordination is necessitated by the existence of co-channel interference and the need for its most efficient management.  We are primarily concerned with the adaptation of transmission rate and power at the physical layer, in full coordination with channel allocation at the access layer. This problem obtains different (and very interesting) twists when considered in the context of different schemes that are at the forefront of contemporary and envisioned communication systems, such as OFDMA and CDMA. 


             F1.1. Cross-layer adaptive techniques for multi-cell OFDMA networks


             Orthogonal Frequency Division Multiple Access (OFDMA)  is used in wireless LANs and proposed for up-coming wireless mesh network architectures (WiMaX). In the prevalent multi-cell architectures, a central problem is that of optimal cross-layer resource allocation to maximize total delivered system rate.  

            We consider the synergy between the physical and access layers and address the fundamental joint problem of channel allocation, modulation level, and transmit power control. Each Base Station is equipped with one radio, tuned at a spectrum band, which is further divided into orthogonal OFDM sub-carriers and transmits to a set of associated users in the down-link. The spectrum band is reused at different cells. Since performance is solely determined by sub-carrier reuse, it is important to handle co-channel interference appropriately by constructing co-channel user sets and by assigning transmission parameters on a sub-carrier basis so that achievable system rate is maximized. Inherent in the problem is the fundamental trade-off about amount of bits per symbol versus susceptibility to interference. Simply put, you can either have several users reusing the same sub-carrier with low modulation levels or fewer users with high modulation levels, but not both. OFDMA introduces additional novel challenges to resource allocation due to different quality of sub-carriers for users and existing transmission power constraints.

          We derive the structure of the problem and propose two classes of centralized greedy heuristic algorithms. The first one considers each sub-carrier separately and sequentially allocates users from different base stations in the sub-carrier based on different criteria such as minimal induced and received interference, maximal rate increase and SINR balancing.  The second one is based on sequential access point activation and water-filling across sub-carriers based on interference related metrics. The greedy structure of the algorithms can be viewed as stemming from that of the problem for the simple case of two Base Stations: the problem is then reduced to a maximum weighted matching one and is solved optimally with greedy methods. The first class of heuristics with SINR balancing performs better.  The problem opens up directions for future investigation, the most important being : investigating fundamental different between down-link and up-link transmission with regard to co-channel interference, devising distributed intra-cell resource allocation algorithms, taking into consideration packet arrival dynamics, and more.

             F1.2. Carrier assignment algorithms for multi-carrier systems with channel adaptation


             Multi-carrier systems have been around in the last decades, the most prominent example being OFDMA, where a user can use several carriers to transmit traffic. Even 2G FDMA/TDMA systems fall within this category, but with the difference that a user can use only one of the carrier frequencies.  In this activity, we study carrier assignment in a single-cell multi-user multi-carrier system with an underlying TDMA frame, with the objective to satisfy user rate requirements with minimal amount of resources.

              Different users experience different quality in different carriers, due to frequency selectivity of propagation channels and not co-located receivers. A static problem instance is studied, specified by carrier qualities for users and user rate requirements. We employ adaptive modulation to create assignment preferences, such that each user is allocated to good quality carriers and satisfies its rate requirements with few utilized slots. We study integral and fractional user assignment whereby a user is assigned to one or several carriers respectively. The problem of fractional assignment is formulated as a linear programming one. For integral assignment, we introduce two classes of heuristics based on iterative carrier reassignment and user substitution respectively, which can be viewed as emerging from corresponding optimal algorithms for fractional assignment. We extend our approach to allocating groups of carriers. We use Lagrangian relaxation to obtain performance bounds and demonstrate that the two classes of integral assignment heuristics actually emerge naturally from two types of relaxations. The approach performs well in terms of identifying efficient feasible solutions. Moreover, the notion of Lagrangian relaxation makes it amenable to distributed implementation. This observation makes our study a significant prelude in the challenging topic of distributed spectrum coordination and management, which is central to IEEE 802.22 Cognitive Radio standard, currently under development.


             F.1.3. Dynamic resource allocation in CDMA systems with deterministic codes and multi-rate provisioning


            Next generation wireless CDMA systems will rely on advanced techniques such as multi-code structures, controllable code spreading gains, and transmit power adaptation in order to foster delivery of high anticipated data rates. The emerging resource allocation problem in that context obtains a different twist due to cross-channel interference that is caused by code cross-correlation. The problem we address is how to handle available resources so as to deliver maximum rate while satisfying minimum rate requirements of users. 

            We bring into play symbol synchronism and devise symbol-synchronous and asynchronous models that capture different communication scenarios. The synchronous model holds for down-link transmission in one cell. The asynchronous model captures up-link single-cell communication and down-link or up-link scenarios in multi-cell systems. It can also account for multi-path with each path corresponding to a virtual user. The associated resource allocation problem consists of two steps: (i) identification of a subset of admissible codes (i.e. one, where code SINR requirements are satisfied), (ii) code allocation to users to satisfy min-rate requirements. To solve the hard combinatorial problem (i) we use various algorithms based on criteria that capture code cross-correlation, induced interference to the system, and code rates. For the NP-Hard problem (ii), we also use various heuristics. A major finding is that in the synchronous case, the problem structure allows the distinction of the two stages, and hence problems (i), (ii) can be handled separately. On the contrary, in the asynchronous case, this distinction is not feasible: the different user delay profiles perceived at the receiver render code admissibility dependent on code allocation to users. Our models and numerical results indicate interesting trends and lead to useful conclusions and design guidelines for resource allocation algorithms under the aforementioned regimes.

     Relevant publications:

     [J10] I. Koutsopoulos and L. Tassiulas, ''Carrier assignment algorithms for OFDM-based and other multi-carrier wireless networks with channel adaptation'', IEEE Transactions on Communications, vol. 56, no.12, pp.2190-2199, Dec. 2008.

    [J6] I. Koutsopoulos, Ulas C. Kozat and L. Tassiulas, ''Dynamic resource allocation in deterministic-code CDMA systems with multi-rate provisioning'', IEEE Transactions on Mobile Computing, vol.5, no.12, pp.1780-1792, December 2006.

     [J4] I. Koutsopoulos and L. Tassiulas, ''Cross-layer adaptive techniques for throughput enhancement in wireless OFDM-based networks'', IEEE/ACM Transactions on Networking, vol.14,  no.5,  pp.1056-1066,  October 2006.

     [C13] U.C. Kozat, I. Koutsopoulos and L. Tassiulas, ''Dynamic code assignment and spreading gain adaptation for synchronous CDMA wireless networks'', Proceedings of IEEE International Symposium on Spread Spectrum Techniques and Applications (ISSSTA) 2002, vol.2, pp.593-597, Prague, Czech Republic.

     [C9] I. Koutsopoulos and L. Tassiulas, ''Channel-state adaptive techniques for throughput enhancement in wireless broadband networks'', Proceedings of IEEE INFOCOM 2001, vol.2, pp.757-766, Anchorage, Alaska.

     [C8] I. Koutsopoulos and L. Tassiulas, ''Carrier assignment algorithms in wireless broadband networks with channel adaptation'', Proceedings of IEEE International Communications Conference (ICC) 2001, vol.5, pp.1401-1405, Helsinki, Finland.



      F.2. Channel state-aware retransmission and scheduling for maximum throughput in wireless networks


      A fundamental premise in cross-layer design is the ability to optimally respond to varying channel conditions. Channel state should be tracked and appropriate adaptation decisions need to be taken at the scale of channel variations whenever possible. We study the structure of such adaptation policies and derive optimal control policies in the sense of maximizing throughput for: (i) physical layer parameter adaptation, by using a simple channel state monitoring method that exploits available link layer feedback information, (ii) packet scheduling of delay-sensitive traffic.    


             F.2.1. Optimal transmission rate control policies in a  wireless link under partial state information


            We consider the problem of PHY layer transmission rate control to maximize throughput of a wireless link over a finite time horizon. Link quality fluctuation is captured by a 2-state Markov chain with known parameters. A state corresponds to different error probability for given transmission rate. At each state, there exists a most advantageous rate to operate at. The rate controller does not know the actual link state at each time instant, so as to optimally decide which rate to transmit with. Instead, it relies on incomplete state information in the form of feedback of positive and negative acknowledgements (ACKs and NACKs) for transmitted packets, coming from the link-layer automatic repeat request (ARQ) protocol. We cast the problem as

a Partially Observed Markov Decision Process (POMDP) one with a finite horizon and study the class of transmission rate control policies for this type of feedback. For the case when the link state is time-invariant (but unknown to the controller), the optimal control problem is essentially a state estimation problem and the optimal strategy is of threshold structure.  The policy admits a simple interpretation: increase the rate when the number of successive ACKs exceeds a threshold, and decrease the rate when the number of successive NACKs exceeds a threshold.



             F.2.2. Scheduling of real-time delay-sensitive traffic over wireless links


             When scheduling real-time traffic, a basic QoS requirement is delay of delivered traffic. In other words, user packets need to be delivered to their respective destinations within certain deadlines, otherwise they do not fulfill their purpose.  In wire-line networks, the policy that each time transmits the packet that is closer to getting expired (termed Earliest Deadline First, EDF policy) is known to be optimal in the sense of delivering maximum number of packets within their deadlines. However, scheduling real-time traffic over wireless links of time varying quality is a fundamentally different problem and this is precisely the topic of this work. The different instantaneous channel conditions of users, different pattern of channel state transitions, the traffic profiles (e.g. intensity) need to be introduced to the model in addition to deadline constraints. 

              A Base Station schedules down-link real-time traffic to several users. We consider the simple form of Constant Bit Rate (CBR) traffic where packet deadlines of successive packets of a user differ by a constant amount, specific to the user. The optimization criterion we consider is long-term packet loss due to deadline expiration. We cast the problem within a Markov Decision Process (MDP) framework that can be used to derive the optimal policy. Meanwhile, we study two simple suboptimal policies, Earliest Deadline First (EDF) and Best Channel First (BCF) and compare their performance to that of the optimal policy computed by the MDP approach. Their performance depends on traffic load, channel modeling and channel switching rate. For instance, EDF policy performs better for light traffic load and low channel switching rates, whereas BCF is better when traffic load increases and channel state changes rapidly. Various other interesting observations are stated. For future investigation, the issue that merits consideration is the design of advanced practical scheduling policies that take into account the collective impact of all parameters above.


     Relevant publications:

     [J11] I. Koutsopoulos and L. Tassiulas, ''Optimal transmission rate control policies in a wireless link under partial state information'', IEEE Transactions on Automatic Control, vol.55,no.1, pp.127-131, Jan. 2010.

      [C10] I. Koutsopoulos and L. Tassiulas, ''Link adaptation policies for wireless broadband networks'', Proceedings of IEEE Global Telecommunications Conference (GLOBECOM) 2001, vol.1, pp.572-576, San Antonio, Texas.

      [C11] T. Ren, I. Koutsopoulos and L. Tassiulas, ''QoS provisioning for real-time traffic in wireless packet networks'', Proceedings of IEEE Global Telecommunications Conference (GLOBECOM) 2002, Taipei, Taiwan.

     F.3. Joint Access Point selection and channel allocation in wireless networks


     In wireless cellular networks or other networks with single-hop communication, the fundamental access control problem pertains to access point (AP) selection and channel allocation for each user. For users in the coverage area of one AP, this involves only channel allocation. However, users that belong in the intersection of coverage areas of more than one AP have an interesting freedom of choice that gives rise to a whole new twist to the access problem: they can choose  the appropriate AP to establish connection and thus implicitly affect the channel assignment procedure.  This problem falls squarely within the thrust of problems dealing with user association in wireless LANs and wireless mesh networks.

     We address the joint problem of AP selection and channel assignment with the objective to satisfy a given user load vector with the minimum number of channels. This objective allows the system to use the minimum number of resources and preserve some amount of resources for potential traffic load increase or link quality deterioration of users in the future, in which case additional channels would be needed. Our major finding is that the joint problem reduces to plain channel allocation in a cellular network that emerges from the original one after executing an iterative and provably convergent clique load balancing algorithm. For linear cellular networks, our approach leads to minimum number of required channels to serve a given load vector. For two-dimensional cellular networks, the same approach leads to a heuristic algorithm with a suboptimal solution due to the fact that clique loads cannot be balanced. The structure of the algorithm makes it amenable to distributed implementation, whereby a cell requires only knowledge of a few neighboring cells. Numerical results demonstrate the performance benefits of our approach in terms of blocking probability in a dynamic scenario with time-varying number of connection requests. The presented approach constitutes the basis for addressing more composite resource allocation problems in different context. Various extensions can be considered and solved. Of particular interest are algorithms that can also capture specific components under investigation for inclusion in future standards, such as cell size control.


     Relevant publications:

    [J7] I. Koutsopoulos and L. Tassiulas, ''Joint optimal access point selection and channel assignment in wireless networks'', IEEE/ACM Transactions on Networking, vol. 15, no.3, pp.521-532, June 2007.

    [C6] I. Koutsopoulos and L. Tassiulas, ''Joint base station and channel allocation in mobile cellular networks'', Proceedings of International Communications Conference (ICC) 2000, vol.3, pp.1558-1562, New Orleans.

      F.4. Wireless relay networking


       LTE-Advanced wireless systems will be endowed with the task of performing appropriate relaying in severe interference conditions. We consider an interference-limited wireless network, where multiple source-destination pairs compete for the same pool of relay nodes. In an attempt to maximize the sum rate of the system, we address the joint problem of relay assignment and power control. Initially, we study the scenario where each source greedily selects the strategy (transmission power and relay) that maximizes its individual benefit, leading

to a simple algorithm of linear complexity. Then, we propose an iterative relay assignment and power control algorithm of polynomial complexity that is amenable to distributed implementation through appropriate message passing. We evaluate the sum rate performance of the proposed algorithms and derive conditions for optimality. Our schemes incorporate two of the basic objectives of the 4th generation broadband cellular systems, namely interference management and relaying.


Relevant publications:

     [C32] L. Gatzikis and I. Koutsopoulos, “Low complexity algorithms for Relay Selection and power control in interference-limited environments”, Proceedings of IEEE WiOpt, 2010, Avignon, France.



E. Adaptive antenna arrays, Space Division Multiple Access (SDMA) and impact on higher layer protocols


     E.1. Impact of SDMA on channel allocation and buffer management


     Adaptive antenna arrays that are capable of dynamically forming beams of different shapes and orientations are at the forefront of recent advances in signal processing systems. The expectation for delivering data rates that are several multiples higher than current ones stems from their ability to serve several users within the same conventional channel. This is achieved by forming several directional beams at the same time, one towards a specific user, so as to create as little interference as possible to others. Essentially this procedure creates virtual spatial channels which are however not orthogonal due to unavoidable impairments of beam patterns. However, the existence of adaptive antenna arrays at the physical layer raises significant issues at protocols operating at higher layers. This line of work investigates the impact of adaptive antenna arrays on (i) channel allocation, (ii) buffer management, (iii) media access control.  The IEEE 802.11n WLAN standard for high throughput and the envisioned future enhanced IEEE 802.16x standards  (currently under development) will rely a lot on such structures and cross-layer control algorithms in the presence of adaptive antenna arrays.


          E.1.1. Joint channel allocation, beam-forming and power control : a unified approach for OFDMA, CDMA, TDMA


          We capture the impact of SDMA on channel allocation, a mechanism operating at the access layer. This impact is reflected differently into channel reuse for TDMA, CDMA and OFDMA systems, due to different emerging instances of co-channel and cross-channel interference and different effect of user spatial channel characteristics on system channels, namely time slots, codes and sub-carriers respectively. For instance, OFDMA is the most general case with regard to different perceived quality of users in different channels: different users experience different quality in different sub-carriers due to frequency selectivity and there are also temporal channel quality variations. On the other hand, CDMA is the most general case with regard to the regime of interference, since non-orthogonal channels (codes) give rise both to co-channel and cross-channel interference.

          We treat these three schemes in a unified manner. We show that performance is determined by constructing appropriate spatially separable co-channel user sets, i.e. user sets with a large enough aggregate data rate and a minimum SINR requirement for each user. Spatial separability depends non-trivially on spatial user channel characteristics (reflected in user spatial signatures and spatial covariance matrices), beam-forming vectors, transmission powers and utilized transmission rates.  Starting from devising spatial separability conditions, we introduce heuristic algorithms for joint channel allocation, downlink beam-forming and transmission power control with the objective to increase achievable system rate and provide QoS to users in the form of minimum rate guarantees. We study the class of greedy assignment algorithms based on criteria such as induced or received interference and SINR, and the class of SINR balancing algorithms. Our results show the superiority of SIR balancing resource allocation algorithms and demonstrate the performance benefits of cross-layer considerations. Clearly the procedure above (which is presented for a single-cell system) can be extended to realistic multi-cell architectures with a view toward autonomous cell operation based on local information and measurements.


          E.1.2. Resource allocation algorithms in SDMA / OFDMA Wireless LANs with transceiver resource limitations


          The benefits of SDMA are attributed to the ability of an adaptive-antenna-array equipped base station to form multiple beams towards different users, each beam by a dedicated transceiver. In C.1.1 we argued that adaptive antenna arrays at the physical layer mandate significant modifications for higher layer mechanisms  in order to realize  the benefits of SDMA.  In addition, OFDM in conjunction with SDMA creates additional challenges: for optimal operation of a system with N sub-carriers and an antenna array with M elements, NM beams need to be formed. Depending on the values of N and M, this number can be several hundreds or even thousands.  However, oftentimes the number of beams C that can be formed at the transmitter can be less than NM. This is attributed to various reasons: transceiver hardware resource limitations, circuit space constraints, interference specs, etc.

          Therefore, a coupled resource allocation problem emerges, namely assigning transceiver hardware units and OFDM sub-carriers for transmission to users. The problem is coupled, in the sense that the choice of transceivers influences the allocation of sub-carriers and vice versa:  different users can be served either with the same beam from a transceiver and different sub-carriers or with different beams and the same sub-carriers. We characterize the problem structure and propose meaningful heuristic algorithms for beam-forming and assignment of sub-carriers and transceivers to users. The objective is to increase achievable system rate and ensure QoS in the form of minimum rate guarantees.  The proposed algorithms consist of two stages. One, where the allocation is performed under no transceiver limitations. A second one where beams are unified.  The criteria for resource assignment and beam formation are based on spatial separability properties of users, beam vector cross-correlations and induced interference. Numerical results quantify the performance benefits of these cross-layer techniques and provide useful insights and design guidelines for realistic systems. Most importantly, results show there exists a maximum number of transceivers C*, beyond which no further rate benefits are incurred. The approach is promising and opens up several directions for future investigation.


          E.1.3. Spatial Processing techniques and maximum throughput scheduling


          Scheduling in a multi-user system in the down-link of a base station with an antenna array amounts to choosing a subset of users, transmission rates and beams and powers to realize the transmission. This decision needs to be taken at each time slot. In a system with user packet arrivals at user queues, the objective is to devise scheduling techniques that attain maximum throughput. This is essentially equivalent to maintaining stability in the system for as high packet arrival rates as possible, through keeping user queue lengths stable. However, deriving a throughput-optimal policy involves significant computational overhead at each slot, which in our case implies selecting one among an exponentially large number of user subsets, and checking over numerous feasible combinations of data rates (i.e., rates that guarantee a spatially separable user subset).

          We introduce scheduling algorithms that employ randomization to achieve maximum throughput at low computational complexity. The rationale is the following: pick a user subset uniformly at random from the set of currently feasible subsets. Compare the sum of incurred rates (weighted by user queue lengths) with the weighted sum of rates corresponding to the rate vector used at a previous scheduling instant. Then, choose the rate vector that corresponds to the largest one of the two cases. The algorithms operate together with physical layer spatial processing techniques, such as transmit beam-forming and Costa pre-coding. These techniques lead to different spatial separability conditions. Simulation results show that the proposed randomized scheduling with Costa pre-coding algorithms carry the superior performance of Costa pre-coding at the physical, over to throughput benefits at higher layers. As a first step, we show that the algorithms lead to maximum throughput when the set of feasible user subsets does not change with time. Clearly, it is of interest to modify the policy to include the most realistic case of time-varying set of feasible user subsets. This approach could be the cornerstone for addressing and solving more composite problems in the same context.


E.1.4. Space Division Multiplexing for IEEE 802.11n

        We take into account the limitations of the IEEE 802.11n standard on beam formation over many OFDM subcarriers and presence low complexity beam-forming techniques to increase SINR at the receivers. 

     Relevant publications:

     [J8] I. Koutsopoulos and L. Tassiulas, "The impact of space division multiplexing on resource allocation: a unified treatment of TDMA, OFDMA and CDMA'', IEEE Transactions on Communications, vol.56, no.2, Feb. 2008.

     [J3] I. Koutsopoulos and L. Tassiulas, ''Adaptive channel assignment in SDMA-based wireless LANs with transceiver resource limitations'', EURASIP Journal on Signal Processing, Special Issue on signal processing-assisted cross-layer designs, vol.86, no.8, pp.1879-1895, August 2006.

     [J2] I. Koutsopoulos, K.P. Tsoukatos and K. Aggelis, ''Physical layer techniques and maximum throughput scheduling with antenna arrays'', IEEE Communications Letters,  vol.10,  no.6,  pp.465-467, June 2006.

     [C16] I. Koutsopoulos and L. Tassiulas, ''Adaptive channel allocation in OFDM/SDMA wireless LANs with limited transceiver resources'', Third IFIP-TC6 Networking Conference 2004, Athens, Greece and Lecture Notes in Computer Science, vol.3042, pp.1384-1389, 2004

     [C15] I. Koutsopoulos, T. Ren and L. Tassiulas, ''The impact of Space Division Multiplexing οn resource allocation: Α unified approach'', Proceedings of IEEE INFOCOM 2003, San Francisco.

     [C14]  I. Koutsopoulos and L. Tassiulas, ''Adaptive resource allocation in SDMA-based wireless broadband networks with OFDM signaling'', Proceedings of IEEE INFOCOM 2002, vol.3, pp.1376 -1385, New York.

      [C27] C. Papathanasiou, I. Koutsopoulos and L. Tassiulas, “Interference mitigation in co-existing 802.11n WLANs”, Proceedings of Wireless Telecommunications Symposium (WTS), 2009, Prague, Czech Republic.

     [C29] C. Papathanasiou, I. Koutsopoulos and L. Tassiulas, “Low complexity beam-forming techniques for IEEE 802.11 WLANs”, Proceedings of IEEE Vehicular Technology Conference (VTC) 2009 Fall, Anchorage, Alaska, 2009.

     E.2. Media access control protocols for wireless LANs with directional antennas


     In near-future wireless access point (AP) networks, with each AP equipped with an adaptive antenna array, a central problem faced by each AP will be to identify and localize the set of users that are associated with it. The adaptive antenna array can efficiently serve this purpose and aid in accurate user localization decisions, i.e. identifying the specific orientation that the user is located with respect to the AP. In this problem, a significant figure of merit is localization delay.

     In order to take user localization decisions with small delay, the AP should actually address users with their IDs and users should employ an appropriate medium access protocol to respond to the AP request. We introduce and compare the class of of protocols that employ directional beam-forming and use contention-based or contention-free polling methods to locate users residing in or out of coverage range of the AP. In the former ones, the AP scans the area around it, searching for a specific user with successive directional beams. In the latter case, the AP addresses all users in a sector, which then need to employ a collision resolution protocol.  In a follow-up work we enhance the performance of protocols in terms of delay and make them capable of coping with user mobility by adding techniques of user location caching and multiple transceivers for forming several beams simultaneously.


     Relevant publications:

    [C12] I. Koutsopoulos, T. Ren and L. Tassiulas, ''Efficient media access protocols for wireless LANs with smart antennas'', Proceedings of IEEE Wireless Communications and Networking Conference 2003, New Orleans.

     [C21] C. Verikoukis, L. Alonso and I. Koutsopoulos, ''Performance evaluation of directional antenna-assisted MAC protocols in the presence of mobility'', Proceedings of IST Mobile Summit 2006, Mykonos, Greece.

D. QoS provisioning and energy efficiency issues in wireless ad-hoc networks


     D.1. Energy-efficient joint scheduling and power control of end-to-end sessions in wireless ad-hoc networks


     It has become common consensus that independent consideration of communication layers often turns out to be inadequate in terms of providing the desired quality of service (QoS) and power efficiency in wireless networks. The need for a synergistic, cross-layer design framework has already been identified. In that respect, our work constitutes an important step towards a better understanding of the cross-layer paradigm by simultaneously targeting both the power efficiency and the end-to-end QoS in multi-hop wireless networks. We build our approach on the following fundamental question in a wireless ad-hoc network setting: consider a set of end-to-end sessions, each with specific end-to-end rate requirements, to be transferred from a given source node to a specific destination. We are given a certain frame length. How can we satisfy the user session requirements by consuming as small amount of power as possible by employing appropriate scheduling and power control?

     We address the joint problem of power control and scheduling with the objective of minimizing the total transmission power subject to end-to-end bandwidth guarantees and bit error rate constraints of each communication session. After identifying the inherent difficulty of the problem, we propose two classes of heuristics that rely on graph theory principles as well as on derived metrics such as effective interference. The first heuristic follows a top-down design strategy by solving the schedule feasibility problem as the initial step and then targeting the total power efficiency. The idea here is to start at each slot with a link set that is maximal matching of the network graph and remove links until a feasible (from the point of view of SINR requirements) set of links is found. On the other hand, the second heuristic follows a bottom-up approach that schedules one wireless link at a time by greedily filling up free the available time slots so as to evenly distribute interference, and in that sense it is reminiscent of a water-filling-like approach. The simulation results reveal valuable insights about the performance of each strategy. The top-down design strategy turns out to address power efficiency issues better, whereas the bottom-up design strategy with a properly selected cost function for link scheduling shows better performance in finding a feasible solution, namely one that satisfies both the QoS and the transmit power constraints. Significant open issues emerge when extending our approach to capture routing layer decisions as well.


     Relevant publications:

      [J5] Ulas C. Kozat, I. Koutsopoulos and L. Tassiulas, ''Cross-layer design and power-efficiency considerations for QoS provisioning in multi-hop wireless networks'',  IEEE Transactions on Wireless Communications, vol.5, no.11, pp.3306-3315, November 2006

     [C17] U. C. Kozat,  I. Koutsopoulos and L. Tassiulas,  ''A framework for cross-layer design of  Energy-Efficient Communication with QoS provisioning in Multi-hop wireless networks'',  Proceedings of IEEE INFOCOM 2004, Hong Kong.

     D.2. Low-complexity distributed approximate max-min fair scheduling for wireless networks


     Max-min fair bandwidth allocation is a meaningful objective whenever the level of user satisfaction cannot be clearly expressed as a function of the allocated bandwidth. In such cases, an intuitive objective is to split the available resources equally among users. If a user cannot utilize a portion of its allocated resource due to a constraint, this should be evenly distributed among other users. A bandwidth allocation is said to be max-min fair, if it is not possible to increase the allocation of any user without hurting another user with a lower service rate. Max-min fairness can be addressed at the network layer or at the link layer. Our work focuses on the latter one.

     In a distributed and resource-constrained environment such as a wireless ad-hoc network, one would like to devise an algorithm that combines conflict-free scheduling with minimal node coordination and achieves a good approximation of max-min fair rates. The method should ideally rely on readily available local information and should involve minimal signaling load in the network. This is precisely the subject of this work. Namely, we address the issue of approximating max-min fairness in a wireless network without the requirement for network-wide node coordination and we present a low-overhead greedy distributed algorithm for reaching this goal. The algorithm is based on distributed computation of a maximum weighted matching based on appropriately defined link weights that capture number of packets awaiting transmission from each flow, and subsequent scheduling of link flows in an effort to provide max-min rates to them. An inherent feature of our approach is its immunity to topology changes as well as to flow traffic variations. Our method is shown to outperform significantly the centralized (yet, conservative) algorithm of max-min fair rate computation in general topologies in terms of total resulting throughput, minimum shares and node resource utilization. Various extensions can be considered to this model, that would encompass physical layer transmission peculiarities and more advanced resource allocation schemes such as OFDMA. 


     Relevant publications:

    [C18] A. Penttinen, I. Koutsopoulos and L. Tassiulas, "Low-complexity distributed fair scheduling for wireless multi-hop networks", First Workshop on Resource Allocation in Wireless Networks (RAWNET), WiOpt 2005, Riva del Garda, Italy.

     D.3. Broadcasting of signaling information through controlled flooding algorithms in wireless networks


     A problem that arises often in unstructured wireless networks is that of fast dissemination of control information packets from a node to the network. These control packets can carry over to the entire network information about topology and channel state, thus aiding in realizing resource allocation or routing decisions.  A commonly used protocol is that of flooding, which constitutes the fastest way of disseminating packets and is particularly encouraged in cases of rapidly changing topology scenarios. Nevertheless, in a network of almost static nodes, this can lead to significant waste of resources and overhead in terms of bandwidth consumption.

      We propose two alternative protocols for forwarding control packets to a network. These protocols in some sense perform a kind of controlled flooding. The first protocol is based on constructing a minimum spanning tree of the network graph and forwarding packet on tree links. Different cases of weights of tree links are examined, that reflect different topological characteristics of the network, such as node degree, proximity from minimum and maximum degree nodes, network diameter. The second protocol uses a modified method of minimum node cover. The protocols are more efficient than flooding in terms of bandwidth consumption and have satisfactory performance in terms of average packet delivery delay. They are also quite robust in cases of low to average node mobility and are amenable to distributed implementation.


     Relevant publications:

    [C5] I. Koutsopoulos, D. Connors, A. Savvides and S.K. Dao, ''Intra-team multi-hop broadcasting (ITMB): a MAC layer protocol for efficient control signaling in wireless ad-hoc networks'', Proceedings of International Communications Conference (ICC) 2000, vol.3, pp.1723-1727, New Orleans.



C. IEEE 802.11 MAC issues


     We ask the question: how can we improve the throughput of medium access control protocols through minimal and backwards compatible changes? To this end, we propose simple rules by which the contention window of the IEEE 802.11 MAC protocol depends on two factors that have not been to date present in the protocol: queue lengths and channel quality (through PHY layer rate). Thus, nodes with larger backlog and/or better channel quality are implicitly obtaining the channel with higher chances. We show that the capacity region is significantly enlarged compared to the IEEE 802.11 protocol.

     In [C23] we identify the problem of service time estimation in 802.11 MAC. An estimate of the service time is important to have as input to higher layer algorithms and mechanisms. However this is challenging to obtaining given the significant degree of randomness that is present in the protocol.


Relevant publications:

[J15] R. Oliveira and I. Koutsopoulos, “Maximum throughput Access control in wireless LANs through Max-weight inspired policies”, to appear, IEEE Transactions on Vehicular Technology.


[C26] R. Oliveira and I. Koutsopoulos, “Queue and channel state awareness for maximum throughput Access control in CSMA/CA based wireless LANs”, Proceedings of IEEE Wireless Communications and Networking Conference (WCNC), 2009, Budapest, Hungary.  


[C23] R. Oliveira, L. Bernardo, P. Pinto and I. Koutsopoulos, ''On service time estimation in 802.11 WLANs with heterogeneous sources'', Proceedings of IEEE International Symposium on Wireless Communication Systems (ISWCS) 2008, Reykjavik, Iceland.

B. Wireless Sensor Networks

     B.1 Distributed inference (estimation) in wireless sensor networks

The primary tasks that are expected to be fulfilled by wireless sensor networks with certain performance guarantees are the estimation of an unknown parameter or parameter process, or the detection of an event in the deployment area. Performance guarantees can be low estimation error, or small probabilities of false alarm and missed detection. Such operational tasks are mapped onto objectives that are different from conventional ones such as end-to-end throughput maximization or delay minimization for which networks are traditionally deployed. Network control actions should take into account these different objectives so that the network performs the tasks it is entitled to with best performance.

A fundamental distinguishing characteristic in sensor networks is the spatial correlation of generated traffic. We study the problem of maximum lifetime in wireless sensor networks that are entitled with the task of estimating an unknown parameter or process. Sensors take measurements and transfer them in multi-hop fashion to a fusion center (FC) for Maximum Likelihood (ML) estimation. To engineer the network for lifetime maximization while adhering to estimation error specifications, the number of measurements by each sensor per unit of time (namely, sensor measurement rate) and the routes to the FC are controlled. Sensor spatial correlation, measurement accuracies, link qualities and energy reserves affect sensor measurement rates and the routes to the FC, and, in turn, measurement rates and sensor characteristics impact estimation error. We show that the problem can be decomposed into separate optimization problems where each sensor autonomously takes its measurement rate and routing decisions, and we propose an iterative primal-dual algorithm with low-overhead signaling for solving it. Our work optimally captures the fundamental tradeoff between network lifetime and estimation quality and yields a solution based on distributed sensor coordination.


We investigate optimal endogenous sensor measurement rate control, in-network data aggregation and routing for achieving the goal above. Sensors take measurements and aggregate incoming data from neighbors in a single outgoing flow by applying appropriate aggregation weights. By doing so, they control the variance of outgoing flow. Each sensor controls its measurement rate and aggregation weights, and aggregated measurement data are routed to the FC for Maximum Likelihood (ML) estimation. The challenge is to find an optimal compromise between eliminating data redundancy and maintaining data representation accuracy so as to adhere to estimation quality constraints and reduce the volume of transported data, thus improving network lifetime.


In [C35], we study the impact of physical layer (PHY) transmit rate control on energy efficient estimation in wireless sensor networks. A sensor network collects measurements and transmits them to a Fusion Center (FC) with controllable PHY transmission rates. The FC performs estimation of an unknown parameter process based on sensor measurements, and it needs to adhere to an estimation error constraint. The objective is to maximize network lifetime. High transmission rates consume more energy per transmitted bit, however they convey larger amount of data per unit time and thus can aid in satisfying the estimation error constraint. We identify basic structural properties of the optimal solution, and we propose an iterative algorithm for reaching a solution based on light-weight feedback from the FC.


     Relevant publications:

    [C33] I. Koutsopoulos and M. Halkidi, “Measurement aggregation and routing techniques for Energy-efficient Estimation in Wireless Sensor Networks”, Proceedings of IEEE WiOpt, 2010, Avignon, France.

    [C31] I. Koutsopoulos and M. Halkidi, “Lifetime maximization in wireless sensor networks with an Estimation mission”, Proceedings of IEEE Global Telecommunications Conference (GLOBECOM), 2009, Hawaii, USA.

     [C35] I. Koutsopoulos, S. Stanczak and A. Feistel, “Transmit rate control for energy efficient estimation in wireless sensor networks”, Proceedings of IEEE Global Telecommunications Conference (GLOBECOM) 2010, Miami. FL.

     B.2 Sensor Networks Architectural design issues


     This piece of work attempts to shed some light from a different angle on fundamental problems that arise in the context of wireless sensor networks. Specifically, we identify analogies between the themes of source and channel coding and some problems that arise in the context of wireless sensor networks. Our aim is to establish a framework that leverages well-known, established methods and results from the former two areas to tackle problems associated to sensor networks.

      First, we address an important tradeoff between required precision and transmission rate in the class of sensor networks with flat architecture by using rate-distortion theory. Next, we address the problem of optimal placement of super-sensors in an area covered by sensors with the objective to minimize a generic cost factor that captures several special cases, and we show the analogy to a vector quantization problem. Based on the analogy of the resulting system with a discrete source emitting symbols, we show that a traffic load balancing problem in the sensor network can be reduced to an entropy maximization problem. Finally, we consider a hierarchical sensor coverage problem that involves deploying a set of sensors with sophisticated sensing capabilities over a grid of ordinary sensors. We cast the problem in the framework of channel coding by defining appropriate analogies under the common denominator of redundancy. With this work we started scratching the surface of an area that holds significant promise in solving fundamental problems pertaining to wireless sensor networks, establishing fundamental operational performance limits and so on.  


     Relevant publications:

     [C19] I. Koutsopoulos, S. Toumpis and L. Tassiulas, ''On the relation between source and channel coding and sensor network deployment", International Workshop on Wireless Ad-hoc networks (IWWAN) 2005, London UK. 

A. Resource allocation and channel estimation for mobile satellite channels


     A.1. Channel allocation and handover issues for non-geostationary mobile satellite networks


     In the near future, terrestrial radio networks are envisioned to integrate with satellite systems in order to provide global coverage and ubiquitous connectivity. In order to establish communication for non-handheld and hand-held user terminals, the radio link design must allow full- and half-duplex operation respectively, the latter being desirable when radiation power restrictions are imposed. Sophisticated resource management is a prerequisite for enhancing system capacity. However, a major inherent problem of the satellite link is packet propagation delay, which may lead to inefficient resource allocation and reduced spectral efficiency.

      We address the resource allocation problem that arises in the context of a medium-earth-orbit (MEO) satellite system with half-duplex communication capabilities. MEO satellite systems are characterized by large propagation delays and large intra-beam delay variations due to their high altitude and relatively large beam sizes. These factors lead to significant unnecessary channel consumption sometimes. We propose a carrier frequency classification scheme, in which the set of available carriers are partitioned into classes and each class is associated with a range of propagation delays to the satellite and is used for communication to and from users located in positions within that delay range. Essentially, users in areas with similar time delays use carriers of the same delay class.  We compute the position of rings that correspond to different delay classes on the satellite footprint. We investigate the phenomenon of elongated cells in the outmost coverage area of the satellite due to earth curvature. For efficient resource allocation, we recognize that a required feature (in addition to identifying the serving satellite and beam) is that of delay class determination.  The method leads to better channel utilization and reduced call blocking rate.

     We also study handover issues pertaining to mobile satellite systems. Handovers in such systems occur even for no user mobility due to the fast relative movement of satellite beams on the earth surface. We define criteria for selection of the serving satellite and beam for a user based on link quality and user position in the footprint. We propose methods for reliable prediction of the time instant of satellite-to-satellite and beam-to-beam handover. These algorithms can be used for channel reservation purposes. Specifically for beam handover, we propose a method for beam selection during beam handover based on the maximum anticipated time spent in a beam, which essentially minimizes the number of transitions and reduces communication overhead.


     Relevant publications:

    [J1] I. Koutsopoulos and L. Tassiulas,  ''Efficient resource utilization through carrier grouping for half-duplex communication in GSM-based MEO mobile satellite networks'', IEEE Transactions on Wireless Communications, vol.1, no.2, pp.342-352, April 2002.

   [C7] I. Koutsopoulos and L. Tassiulas, ''A synchronization-based scheme for simultaneous full- and half-duplex communication in GSM-based MEO mobile satellite networks'', Proceedings of IEEE Global Telecommunications Conference (GLOBECOM) 1999, vol.1A, pp.276-280, Rio de Janeiro, Brazil.

    [C4] I. Koutsopoulos and L. Tassiulas, ''Reliable handover prediction and resource allocation in MEO mobile satellite networks'', Proceedings of IEEE/IMACS Circuits, Systems and Communications Conference 1999, Athens, Greece.

   [C3] I. Koutsopoulos and L. Tassiulas, ''A unified framework for handover prediction and resource allocation in non-geostationary mobile satellite networks'', Proceedings of IEEE Vehicular Technology Conference (VTC) 1999 Fall, vol.4, pp.2106-2110, Amsterdam, Netherlands.

     A.2. Channel estimation of mobile satellite channels in urban environments with the ray tracing method


Ray tracing is a method that is used for deriving an estimate of channel quality and relies on approximating the electromagnetic wave as a ray. Two extra valid simplifying assumptions in the case of mobile satellite systems are: (i)  the received signal in an urban environment is determined to a large extent through interaction with the surrounding environment of the receiver and, (ii) due to the large distance from the satellite transmitter, the wave at the receiver can be considered a plane wave.

The model derives the possible directions and angles through which the signal reaches the receiver. We determine conditions for existence of a line-of-sight (LOS) signal, as well as signals due to reflection ground and surrounding buildings, diffraction from horizontal and vertical building edges and a combination thereof.  The basic advantage of the model is its simplicity and low computational cost. Through comparison with values from real measurements, the proposed model is found to achieve good approximation of the received signal strength. Ray tracing is also used for feeding back channel state information at the transmitter, which in turn can control transmission power.  We classify satellite beams based on the environment they cover (urban, suburban, rural, etc) and formulate a problem of splitting the available transmission power in beams so as to maximize a probabilistic metric of providing acceptable link quality to users.


     Relevant publications:

    [C1]  T. Sofos, I. Koutsopoulos and P. Konstantinou, "A deterministic ray-tracing based model for land mobile satellite channel in urban environment'', Proceedings of IEEE Vehicular Technology Conference (VTC) 1998, vol.1, pp. 658-660, Vancouver, Canada.

    [C2] I. Koutsopoulos and P. Constantinou,  ''Joint channel estimation and satellite antenna power control in mobile satellite networks using ray tracing'', Proceedings of IEEE Vehicular Technology Conference (VTC) 2000 Spring, vol.2, pp.1586-1590, Tokyo, Japan.


©1998-2010 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works, must be obtained from the IEEE.


Nedstat Basic - Free web site statistics
Personal homepage website counter
Free counter