This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
Utilization of electromagnetic radio spectrum is one of the key challenges faced by the wireless networks. Cognitive radio is viewed as the wireless device which is used to improve the spectrum efficiency. This report discusses the different channel allocation methods used in cognitive radio, an algorithm to reduce co channel interference's for a multi cognitive radio system and also about a channel assignment scheme called intelligent channel allocation through which optimum utilization of spectrum can be derived without much change to the infrastructure. It also explains how spectrum utilization is done during the normal call traffic conditions and responses during the emergency situations.
Over the last few years demand for several problems is being faced by governmental organisations around the world with respect to the scarcity utilization of the electromagnetic radio spectrum. Recently Federal communications commission came up with a report which says that at any time only 10% of the frequency spectrum is actively used leaving 90% unused. Sometimes some frequency bands are heavily used and some other bands were being partly used. Due to the large demand in wireless communication like mobile telephony the band allocated for 2G band spectrum and 3G band spectrum has been seriously limited. The increased users are needed to be accommodated within the given band. These adjustments are made by decreasing the cell size at the equipment cost. But the disadvantage of this is below a certain cell size; handoff increases and again limits the use of bandwidth. The only ideal way is through dynamism in the spectrum allocation. Spatial and temporal variations are the factors associated with the call traffic in cluster of cells leading to some being congested and rest being idle. These factors can be considered for efficient use of the spectrum whenever and wherever there is a heavy traffic. The other way to ensure spectrum efficiency is by allocation frequency bands to secondary users (to whom frequency band is not allocated) to get access of frequency band allocated to primary users. Cognitive radio is the means by which this can be done i.e. to improve the spectrum utilisation by exploiting unused spectrum.
In mobile communications, Frequency reuse and Co channel interference's have become key issues, and they have to be considered while considering channel allocation/borrowing. These issues are better addressed by a well-designed scheme. A balanced flexibility to the network can be made to deal during the emergency situations which may lead to higher demand in the spectrum leading to traffic congestion.one of the good examples is during a disaster in Enscheda ,Netherlands a part of the city was destroyed by an explosion caused in a fireworks depot. Many people were injured, and the fire brigade, relief workers experienced a breakdown in communication internally. The factors that caused this difficulty were overloading of the frequency bands. By studying the behaviour of the call attempts to the affected cells such situations can be examined. To cater to this kind of emergency situation's the frequency allocation has to be improved.
In the following sections the report explains different type of channel allocation schemes, cognition in radio networks and a channel allocation algorithm for multi cell cognitive radio and an intelligent channel allocation scheme capable of handling emergency situations.
The main issue with cellular system is limited transmission spectrum, which has to be shared by many users. Each cell is allotted a frequency spectrum, when a user moves into a cell then they are permitted to utilize the channel allocated .the benefit factor is that different cells can use the same channel such that the cells are supposed to be separated by a distance according to system propagation characteristics or else co channel interference occurs .channel allocation deals with the allocation of channels to cells in a network.
In this paper several channel assignment strategies are described. The channel assignment strategies are classified as hybrid channel assignment, fixed channel assignment and dynamic channel assignment. In fixed channel assignment the allocation is like specific channels to specific cells, and they are permanently allocated i.e., static and they allocate in such a way that frequency reuse is maximised but a major drawback of this allocation scheme is the channels in the cell remain the same irrespective of the traffic or customers which might result in traffic congestion. There are several borrowing schemes such as borrow first available, simple borrowing scheme, borrow from richest and basic algorithm with reassignment. The next channel allocation is dynamic channel assignment. This channel assignment reduces the problem mentioned in fixed channel assignment when traffic is non-uniform , in this allocation scheme there is no relationship between channels and cells and all the channels are part of central pool , and the channels are dynamically assigned to the cells as new calls arrive and this ensures that frequency reuse is not violated. Once the call terminates the channel is returned to the pool. These allocation schemes can be used either as centralised or a distributed. But dynamic channel assignment have their own problems too this method has a degree of randomness and due to this frequency reuse is not utilised to maximum extent unlike the Fixed channel assignment systems where cells using the same channel have minimum reuse distance, the next issue with dynamic channel assignment is this methodology uses complex algorithms to decide which channel is efficient , these algorithms can be computationally intensive and require large computing resources to be in order during real time. But the advantage with dynamic channel allocation is they handle busy cell traffic and radio resources efficiently, this allocation scheme allows number of channels per cell to vary with the traffic and hence channel capacity is maximised. The next category of the channel allocation scheme combines the advantage of both dynamic channel assignment and fixed channel assignment and hence it is called hybrid channel assignment. In hybrid channel assignment the channels are categorised into two sets one is borrowable and the other is fixed. channels are assigned as in fixed allocation scheme to each cell and if a cell needs excess of channel which was assigned to it previously then the cell can borrow a channel from its neighbouring cells as long it is available for use and using this channel doesn't violate the requirements of frequency reuse, but there is one major issue in channel borrowing i.e. when a channel is borrowed by the cell from its neighbouring cell, other cells are not supposed to use the channel borrowed because of co channel inference, this leads to over time call blocking. To minimize this issue algorithms are used to ensure call borrowing happens with the most available neighbouring cells. The commonly used channel borrowing methods are borrowing with channel ordering, simple channel hybrid borrowing and borrowing with direction channel locking.
The borrowing with channel ordering was designed to improve the simple channel borrowing approach, the borrowing with channel ordering have two special features one is fixed to dynamic channels ratio varies with the traffic and the second feature is in a cell the first nominal channel will have the highest priority in being applied to a call in the given cell. The last nominal channel is borrowed from the neighbouring cells, once a channel is borrowed by the cell the channel is locked in the cells within the reuse distance. The term locked here means channel cannot be borrowed or used from the neighbouring cells. But the channel can be borrowed only if it's free, it is a strict criteria. The next type is borrowing with directional channel locking in this method channels which are borrowed are only locked in the nearby cells that are affected by borrowing. This differs from borrow channel ordering scheme in such a way that the borrowed channel in every cell is locked within the cell's frequency reuse distance. The benefit of borrowing with directional channel locking is that maximum number of channels is available in the presence of borrowing and call blocking is reduced. Another extension of channel borrowing is to set a portion of channels as dynamic channels with other channels fixed to specific cells. Whenever a cell requires a channel, instead of borrowing from the neighbouring cells, the channel is borrowed from a bank of dynamic channels.
Finally the efficient channel allocation scheme should follow these conditions: - channel assignment schemes must not violate frequency reuse conditions, should adapt to the traffic conditions and utilize the available transmission conditions efficiently. Most of the channel assignment strategies meet the following conditions addressed above to an extent.
To ensure the spectrum utilization in an efficient way the users of various services are supposed to be given access to a spectrum with the limits. And this is called dynamic spectrum access. Cognitive radios are the class of radios which make DSA a possible option. By sensing and adapting to an environment these devices fill voids in the spectrum and increase the spectral efficiency. These radio devices sense the spectrum, detect the primary users and accommodate the secondary users in the free band without creating interference. The main properties of cognitive radios are ; sensing radio frequency technology, cognition capable enough to distinguish the primary users in the spectrum and adaptability capacity to change parameters for better utilisation of unused spectrum. They are capable of conforming to spatial variations and temporal variations in the traffic and increase the overall throughput by accommodating secondary users. The base station exchanges the channel occupancy details and the traffic status with the mobile switching centre time to time. The mobile switching centre is the controller which is capable of taking decisions based on past and current traffic status in order to decide which cell should be allocated with the channel under the given conditions. Once the mobile switching centre allocates the channel it's the base station channel which tune themselves to the corresponding frequencies to serve the corresponding users. Decentralising the control to base station from the mobile switching centre leads to increase in the inter base station traffic as cognition requires the past traffic statics of all the cells with minimum frequency reuse distance .this may also lead to conflict of demand for available channels.
Till now many research has taken place over the allocation of channel in cognitive radio networks, but for channel allocation two factors i.e., channel co interference and frequency reuse factor needs to be addressed when deciding which type of channel allocation to be used.
In multi cell cognitive radio systems due to co channel interferences among the cells in the cluster they cannot select the same channel at a given time, the channel allocation mechanism decides the number of channels to be allocated to a cell in each cluster while for subcarriers cell allocation is responsible for categorizing the allocated channel to users in a given cell hence cell allocation depends on the access networks and the assumption is that cognitive radio systems are orthogonal frequency division multiple access based .With Orthogonal frequency division multiplexing system assumption, channels are split into a set of orthogonal narrow band subcarriers .In a multi cell cognitive radio systems the channel allocation algorithm considers two factors during allocation of resources to cells, the first one is data rate and the other is the degree of interference, this algorithm selects the cell with the largest gap to obtain required data rate, while for the second factor it ensures overlapping channels are allocated, so that less channel interfering is given. In the algorithm, channels are assigned to each cell in the cluster and necessary to assign the sub carriers of channel to secondary cognitive radio users. The rest of channels can be used for other clusters and hence can reduce the time for channel reallocation. Here a proportional fairness algorithm is used and this algorithm is divided into 2 steps, firstly assigns the smallest number of subcarriers to deal the minimum required rate of users, and next is to allocate the rest.
The intelligent channel allocation scheme adapts hybrid channel allocation, the system adapts to temporal variations by working on a slot to slot basis by varying the fixed to dynamic channel ratio. A central dynamic pool will control all the dynamic channels, the system running in the first time slot will predict the traffic of the next time slot. Dynamic channels are efficiently utilised by the reassignment schemes. For the prediction of the traffic in each cell a two stage predictor is used. The first one is a long term predictor and then a short term predictor and copies of each run to all cells independently. The call statics is made as number of calls per time slot, by utilising the traffic data at same time on the previous days the long term predictor will estimate the traffic level of given time. Assuming data to be stationary with a good auto correlation, Auto regressive model of prediction is used for Long term predictor. This model is one of the linear prediction formulas that predict an output of a system based on previous outputs and inputs. A first order auto regressive model is chosen as the long term predictor. Hence the input fed is the previous day traffic at the same time slot. The long term predictor output error values similar to adjoining time slots of the same day is assumed to be stationary with good auto correlation properties. These values are later fed into short term predictor which is of the third order auto regressive predictor. While compared to long term predictor output error, short term predictor output error is very low. The call traffic estimation is made more accurate by a two stage prediction process using the input data characteristics.
Initially each cell in the system is allocated a group of channels which is referred as primary channels belonging to the cell.at the start of each day the number of primary channels allotted is altered based upon the average traffic predicted ie the long term predictor output. Since the output of short term predictor is only known before the time slot. And the output of the short term predictor is used at the beginning of each time slot to prepare each cell for expected traffic which is done by assigning additional channel requirement for every cell. A simple strategy used for channel transfer i.e. richest to poorest to ensure fairness in channel distribution and satisfying constraints due to frequency reuse and co channel interference.
The highest priority is given to primary channels during channel allocation for an incoming call, if primary channel slots are not available then free slots are searched in secondary channels, and in any case if these are not available too then neighbouring cells are searched. With these three stages call allocation process each incoming call is given a fair chance to get accepted and channel prioritization is done only at the logical level. At physical level logical channel levels are mapped on to their corresponding counterparts and are done to achieve better adjacent channel and co channel interference. A channel reassignment system which involves transfer of calls to free slots in primary channels from secondary channels is used. This reassignment is practiced to reduce co channel interference and also for efficient dynamic channel utilization. In this system the algorithm used directional channel locking scheme to mitigate co channel interference. This method reduces the interference by restricting inter cellular movements of a channel and also ensures that in a single channel no two adjacent cells are present.
The predictor functions well because the autocorrelation is perfect for the long term predictor output, and slight change in the autocorrelation leads to failure of the predictor. This failure of correlation happens whenever call traffic increases over a specified area. But the intelligent channel allocation algorithm has been designed to make sure that this doesn't happen. Whenever a call traffic level increase suddenly to a higher level than those predicted by Long term predictor, the two stage predictor use is discontinued and a simple predictor which uses only a previous traffic samples is used. This simple predictor is used to predict the call traffic. When sum of the calls is more than the threshold level an emergency is declared. And emergency is revoked as soon as the above condition ceases. The system keeps switching back and forth as long as the sum value fluctuates around the threshold value. To stop this from occurring again and again the threshold value used for emergency revoking is kept lower than the value used to declare one. For a call allocation architecture the fact is threshold is measured as the boundary for the complete change and any change across the boundary in both directions would lead to instability of the system. To increase the throughput of the affected cells the number of channel borrowed must be as large as possible. This is achieved automatically by using a predictor like a simple short term predictor which samples only a few past traffic values which in turn raises additional channel requirement value and ensures maximum throughput is obtained. In case of emergency services high priority will be accorded and a set of dedicated channels will be allotted to them which differ from those available for regular commercial services. Due to this all the calls which are being supported by secondary channels will be forcibly terminated after a gap of 120 seconds. These terminations will be preceded by a warning. This would gradually increase the overall throughput. In cellular networks emergency services are given a high priority and separate channels are allotted to them than those available for the regular services. In the foregoing algorithm these channels are not included in the available channels set, hence they are exempted from forcible termination scheme.
With cognitive radio technology, new challenges appear leading to interest in research in this technology. The main challenges in cognitive radio technology are cross layer adaptation, frequency reuse, co channel interference and spectrum sensing.
In multi cell cognitive radio systems, Orthogonal frequency division multiplexing system is assumed, since cell allocation is dependent on access networks, these OFDM spectrum shaping capability, its flexibility along with its adaptivity are the strengths that makes OFDM a best transmission technique, but it also involves its own challenges such as phase noise and frequency offset sensitivity on which much should be researched.
In the intelligent channel allocation scheme which is based on long term and short term call traffic statistics supported by cognitive radio techniques for efficient usage of spectrum holds the promise of providing an effective and quick response during emergency situations. This algorithm provides intelligence to the existing radio networks in form of traffic prediction which is in turn used to achieve maximum spectrum utilization under normal traffic conditions as well as during the emergency situations with an effective and quick response. It proposes to be an efficient channel allocation system which gives the incoming calls a greatest possible chance to be accepted in turn reducing the amount of complex computations required by using a slot wise basics. A better throughput is obtained by using this algorithm, as compared to any other channel allocation schemes by utilizing the non-uniformity in cell traffic levels of adjacent cells. Although this scheme explains a better throughput it doesn't explain the power allocation that occurs between the channels.