Wireless Channel Allocation

Get Complete Project Material File(s) Now! »

wireless channel allocation: the background

The process for solving the channel allocation problem usually involves three parts. In the first are defined the criteria to chan-nel allocation. The criteria represent the factors to be considered in the channel allocation process, such as, occupation or inter-ference on channel, traffic in the links, mobility, or fairness. In the second part are defined the approaches to be adopted, e.g., centralized or distributed, and with/without control channel. In the third part is defined the technique used to model and solve the channel allocation problem.
Some of the techniques may be simple. For example, a node locally monitors the occupation of the channels in its transmis-sion radius and selects the channel with lower occupation. In this case, if several nodes select the same channel with lower occupation, they compete to the channel at the time of trans-mission.
Other techniques may be more complex because they involve different channel allocation criteria associated simultaneously, or a larger volume of data processing. In the last case, for ex-ample, the adoption of heuristic or evolutionary algorithms can be more efficient to find a feasible solution to solve the problem.
Figure 3 shows the parts that represent the process for solv-ing the channel allocation problem and which will be detailed in the next subsection. The check mark (in green) indicates the criteria, approaches and the techniques that we will use in our proposal of channel allocation.
Figure 3: Criteria, approaches, and techniques used to solve the chan-nel allocation problem

Criteria to channel allocation

Although there are several types of criteria used in channel al-location, they can be used with more or less emphasis depend-ing on the type of optimization sought in the network. In this subsection, we address some criteria such as based on channel interference, network traffic, fairness, and user mobility.

Selection criterion based on channel interference

In this criterion, the channels are periodically scanned and the least used or less interfering channel is selected. The use of the channel is represented by the signal level or the number of frames transported by a channel.
The criterion based on channel interference is used by both operation modes: infrastructure and ad hoc. In infrastructure mode, an Access Point (AP) is used to scan the channels, eval-uate their use or the interference level and allocate the best channel, as shown in [6] and [7].
In ad hoc networks, the channel allocation is not performed by a central unit, as an AP. Instead, the decisions of channel allocation are made by the nodes independently, selecting the less interfering channel.
Most ad hoc networks operate as a cluster. A cluster is a set of nodes consisting of a cluster head (responsible for allocating resources) and cluster members (that determine the boundary of the cluster). The cluster head usually is elected dynamically by the cluster members. In the work presented by [8], the clus-ter head is represented by a channel coordinator that continu-ously monitor the power level in all channels. If the load on the channel increases beyond the capacity, the channel coordinator change to an additional channel with the lowest power level measurement.
The problem of the channel allocation considering only the monitoring and selection of channels with less activity/interfer-ence is that a device receives the signals only from others de-vices within its communication range. They do not detect the interference at 2 hops (interference among neighboring links). For example, in Figure 4, the AP receives beacons of device1. Device3 receives beacons of device2. Since AP does not see de-vice3 (and vice versa), the channel allocation mechanism can allocate the same channel to the two links (AP-device1 and device2-device3). This will cause interference between the two links and this problem is known as hidden terminal problem.
Figure 4: Hidden terminal problem
To mitigate the hidden terminal problem, [9] proposed a cen-tralized client-driven channel allocation approach. In that proposal, the client that is under the transmission range of several APs is associated to an AP that has a channel with fewer users.

Selection criterion based on network traffic

The solution proposed by [9] is known as traffic-agnostic and the channel allocation process occurs without to consider the traffic in the links. Thus, a non-interfering channel can be al-located to an inactive link (without traffic), underutilizing the channel. A channel allocation mechanism using traffic as se-lection criterion allows allocating non-interfering channels (or with lower interference level) to links with higher traffic. Thus, shared channels may be allocated to links with lower traffic load. Several authors in the literature have proposed channel allocation mechanisms considering the traffic as a decision cri-terion, as presented in [10], [11], [12], and [13].

Selection criterion based on fairness

In the selection criterion based on fairness, the objective is to allocate channels or increase the network performance in a fair way. In [14] is proposed a method to maximize the network throughput and, at the same time, to enhance fairness. In [15] is presented a mechanism for maximizing the use of per-flow bandwidth with fairness. In [16], the authors present a strat-egy to maximize the minimal channel gain to achieve relative fairness.

Selection criterion based on user mobility

The channel allocation related to user mobility has been stud-ied mainly in cellular systems. In [17] and [18], the authors pro-pose mobility-aware resource allocation schemes for femtocell networks. The objective is to associate the best spectrum set of frequency/time while considering the user’s mobility.
In [19] and [20] the authors present strategies for cognitive and VANET networks. In [19], the channels change when a node joins or leaves the network. In [20], the nodes use the relative velocity of nodes to determine the cluster centroid and the po-sition of the nodes. The nearest node of the cluster centroid is designed as the cluster head and allocates the channels for cluster members.
Table 1 summarizes the criteria previously seen and which are addressed in the literature.
In channel allocation, the approaches can be classified as cen-tralized or distributed and with common control channel or without common control channel.

READ  Thermomechanical assessment of empirical potentials 

Centralized and Distributed Approach

In a centralized approach, a central unit is responsible for the channel allocation for each node/link in the network. In the literature, several authors have used a centralized approach to the channel allocation in wireless networks, as presented in [21], [22], [23], and [24].
In the distributed approach, the decision about channel al-location is taken locally in a manner fast, adaptative, but not always optimally. The distributed approach has been used in several types of networks. In [25] and [26] are presented dis-tributed channel allocation mechanisms for IEEE 802.11 net-works and in [27] and [28] for cognitive radio networks.

Common Control Channel and Not Common Control Chan-nel

The Common Control Channel (CCC) is a dedicated channel which allows maintaining a minimum connectivity among nodes and to exchange control messages.
Several previous works in the literature, such as in [29], [30], [31], and [28], address the channel allocation problem using a dedicated CCC. In that cases, the network nodes use two inter-faces, one operating as CCC interface and other as data inter-face, simplifying the coordination process between the nodes.
Others works, such as [32] and [33], do not use a CCC in channel allocation process. In [32], for example, is used a single interface and the channels with less occupancy are selected. The information about the channel is passed to all other devices
through the channel that is being shared by control and data plane.
Table 2 summarizes the approaches used in the channel allo-cation process.

Techniques used to model and solve the channel allocation problem

In this subsection, we address some techniques used to model and solve the channel allocation problem, such as linear pro-gramming, heuristics, evolutionary algorithms, and graph the-ory.

Linear programming

Linear Programming is a technique very used in channel allo-cation in wireless networks. Usually, the channel allocation is described as an NP-hard problem, meaning that the optimal solution grows exponentially with the size of the network. To simplify the resolution of the problem, the channel allocation mechanism can be modeled as a Binary Linear Program (BLP) that can be solved in polynomial time [34]. Several authors ad-dress the channel allocation problem as a linear programming problem as in [35], [36], [37]. A survey about techniques based on Linear Programming for resource allocation in wireless net-works can be found in [38].

Heuristics

In problems involving linear programming, it is common the difficulty of quickly finding a feasible solution in computational time. A heuristic method allows finding a near-optimal solution quickly in cases where an exhaustive search is impractical. In a heuristic method, the algorithm is interactive and in each in-teraction, it searches a good solution, as, the channels with less interference level. Several heuristics techniques are presented in the literature to solve the channel allocation problem.
In [39] and [40] is used Simulated Annealing. In [41], [42], and [43] are proposed Greedy heuristic. In [44] and [45] are presented TABU heuristic.
TABU heuristic [46] starts with an initial solution according to some criterion (e.g., less interfering channels) and at each it-eration, the best solution in the neighborhood is sought. Each selected solution is stored in a list (TABU list) and it is not allowed to repeat moves that lead to an already selected solu-tion. The list remains in memory for a certain amount of time or number of interactions. The final result is expected to be a global optimum or the nearest solution. In our work, we com-pared our channel allocation strategy with TABU heuristic.

Evolutionary Algorithms

Evolutionary Algorithm (EA) is a term used to describe population-based stochastic search algorithms that in some sense mimic natural evolution [47].
Several authors in the literature address the channel alloca-tion using evolutionary algorithms. In [48], Genetic Algorithm (GA) is applied to channel allocation in OFDMA system. In [49], authors propose an algorithm based on ant colony intelligence to solve the problem of channel allocation in Peer to Peer (P2P) links. In [50], authors present a particle swarm optimization method to assign conflict free channel in mobile wireless net-works. In [51], neural network is used for dynamic channel allo-cation in mobile multimedia networks, considering the handoff and traffic mobility.
Although the Evolutionary algorithm is a technique widely used in channel allocation, it is an extensive topic that is not part of our scope. Surveys about Evolutionary algorithms are presented by [52], [53], and [54].

Table of contents :

1 introduction 
1.1 Context and Motivation
1.2 Problem Statement
1.2.1 Can the frequency spectrum be better allocated according to the user behavior?
1.2.2 Can a channel allocation mechanism be easily deployable?
1.3 Contribution of this thesis
1.4 Organization
2 background and related work 
2.1 Wireless Channel Allocation: The Background
2.1.1 Criteria to channel allocation
2.1.2 Approaches in Channel Allocation
2.1.3 Techniques used to model and solve the channel allocation problem
2.1.4 Interference models
2.1.5 Mobility model
2.2 Wireless Channel Allocation: The Related Work
2.2.1 Related Work in Ad Hoc network
2.2.2 Related Work in VANET network
2.3 Conclusion
3 channel allocation strategy in ad hoc networks 
3.1 Formulation of the problem
3.1.1 Channel model
3.1.2 Network model
3.1.3 Interference model
3.1.4 Problem Statement
3.2 The Channel Allocation Mechanism
3.2.1 State1 (Scheduler)
3.2.2 State2 (Topology Manager)
3.2.3 State3 (Allocation Mechanism)
3.2.4 State4 (Interaction Mechanism)
3.3 Computational Complexity Analysis of the Algorithm
3.4 Evaluation Scenarios
3.4.1 Scenario1: Grid topology
3.4.2 Scenario2: SLAW mobility model
3.5 Conclusion
4 channel allocation strategy in vehicular networks 
4.1 Formulation of the problem
4.1.1 Channel model
4.1.2 Network and interference model
4.1.3 Problem Statement
4.2 The Channel Allocation Mechanism
4.3 Performance Evaluation
4.3.1 Scenario1: Manhattan Grid Scenario
4.3.2 Scenario2: Traces of the Cologne City with DTN protocol
4.4 Conclusion
5 conclusions and future horizons 
5.1 Summary of the Thesis
5.2 Future Horizons
bibliography

GET THE COMPLETE PROJECT

Related Posts