Boris Makarevitch

Low complexity adaptive power and subcarrier allocation for OFDMA

The problem of low complexity sub-optimal power and subcarrier allocation for OFDMA systems is addressed in the paper. We propose a heuristic non-iterative method that is an extension of the ordered subcarrier selection algorithm recently proposed for a single user case to the OFDMA systems. The suggested method improves the average BER-performance of the system and it also provides fairness between the users in terms of the bit rates. Both the theoretical analysis and simulation results confirm efficiency of the proposed technique. Based on the proven lemma about inversion of power gains, we strengthen the suggested method by low complexity power loading that consists in equalization of the received signal-to-noise ratios of each user

Jamming Resistant Architecture for WiMAX Mesh Network

The paper presents a jamming resistant architecture for military applications of WiMAX (802.16) mesh network. The main idea is to use multiple base stations (BS), access points to the fixed core network. This will facilitate network survivability both in case of a BS destruction and in case of jamming. When the nodes are affected by jamming they can reroute to another base station if possible. The multiple BS architecture requires distributed scheduling as there is no single point of control that can execute centralised scheduling. A distributed scheduling algorithm, performing well in both jammed and non-jammed environment, is a key issue in implementation of the architecture. Several scheduling algorithms are considered in the paper: a static algorithm using the number of traffic flows in the node, dynamic load-dependent algorithms using two- and three-hop neighbour information. Finally, an algorithm based on finite field initial slot assignment is proposed. This algorithm uses two-hop information and is also load-dependent. Performance of different algorithms is evaluated with and without jamming. The finite field algorithm shows the best performance in both cases

Distributed Scheduling for WIMAX Mesh Network

This paper addresses the problem of distributed scheduling for the WiMAX mesh network. A mesh with multiple base stations increases significantly the network capacity but requires distributed scheduling for its operation. A scheduling algorithm with initial allocation based on finite fields and traffic dependent conflict resolution is proposed. Its performance is evaluated and compared with traffic dependent scheduling not using finite fields

Adaptive Resource Allocation for WiMAX

The mobile WiMAX (standard 802.16e) provides capabilities for adaptive radio resource allocation for users with a slowly changing channel. To achieve a performance gain possible under adaptive allocation it is important to choose an appropriate practical algorithm feasible for real-time implementation. In the paper the allocation task is considered as a two-step process and several combinations of algorithms for each step are proposed. Performance of algorithms is evaluated on a WiMAX model with a variable bit rate traffic.

Application of reinforcement learning to admission control in CDMA network

The paper describes an admission control algorithm for the CDMA networks which is able to adapt to the operating environment. The algorithm is based on the principle of reinforcement learning and it achieves near-optimal performance for various radio propagation conditions and network operators objectives. The performance evaluation results for different state space alternatives and algorithm parameters are presented and compared with the conventional admission control based on the power thresholds.

Practical approaches to adaptive resource allocation in OFDM systems

Whenever a communication system operates in a time-frequency dispersive radio channel, the link adaptation provides a benefit in terms of any system performance metric by employing time, frequency, and, in case of multiple users, multiuser diversities. With respect to an orthogonal frequency division multiplexing (OFDM) system, link adaptation includes bit, power, and subcarrier allocations. While the well-known water-filling principle provides the optimal solution for both margin-maximization and rate-maximization problems, implementation complexity often makes difficult its application in practical systems. This paper presents a few suboptimal (low-complexity) adaptive loading algorithms for both single- and multiuser OFDM systems. We show that the single-user system performance can be improved by suitable power loading and an algorithm based on the incomplete channel state information is derived. At the same time, the power loading in a multiuser system only slightly affects performance while the initial subcarrier allocation has a rather big impact. A number of subcarrier allocation algorithms are discussed and the best one is derived on the basis of the order statistics theory.

On subchannel inversion for improvement of power efficiency in OFDM-based systems

The problem of power loading in Orthogonal Frequency Division Multiplexing (OFDM)-based systems that provides minimization of the bit-error rate (BER) performance under a fixed transmit power (or equivalently minimisation of the transmit power under a fixed value of the BER) is addressed in this work. A lemma about inversion of subchannel power gains is proved and examples that exhibit its utility are presented. As a result, a non-iterative adaptive power allocation technique is derived. The algorithm combines the recently proposed ordered subcarrier selection method with ‘subchannel inversion’. Compared with the original technique, such a combination provides a power gain that depends on the channel statistics.

Scheduling and admission control for 802.11e hybrid coordinator

The paper considers scheduling and admission control algorithms for the 80211e polling-based access. The scheduling problem is formulated as a linear optimisation task that enables to obtain its optimal solution by the simplex algorithm. Performance of the scheduler based on linear programming is evaluated for video calls with variable bit rate video traffic sources and compared with performance of a simple scheduling algorithm with proportional request reduction. An efficient measurement-based admission control is proposed. Finally, the scheduling performance, combined with admission control, is evaluated for both video and voice calls.

Multichannel ad hoc network based on 802.11 standard

The paper proposes a flexible solution to the scalability problem for the 802.11 based military ad hoc networks, which consists in using several frequency channels with different common power levels. By that the network can cover a larger area, and for a small area the amount of interference is reduced by power adjustment. The paper considers different algorithms for power adjustment and evaluates their performance for two channels and various coverage areas and numbers of nodes. Also performance evaluation of a single-channel network is given for comparison. The performance evaluation is done for the protocol model of medium access control. Performance of real-time streaming services is considered with delay and throughput as performance metric. The two-channel network with the proposed adaptive power selection algorithm gives a significant performance gain for the real-time services. This improvement justifies the increase in the amount of used resources

Learning rate control for downlink shared channel in WCDMA

The paper considers bit rate control algorithms for the downlink shared channel in WCDMA (wideband code division multiple access) network. The goal of the rate control is to maximise the data throughput while satisfying quality constraints for the voice users. A learning algorithm, based on Reinforcement learning, is proposed. The algorithm facilitates frame acknowledgments to make decisions on the rate change. The control channel power is used as the state variable and a probabilistic decision policy is applied. Performance of the algorithm is evaluated and compared with performance of a simple algorithm without learning. Data throughput and voice quality cost function are used as performance measures. The learning algorithm achieves higher throughput and lower voice quality cost than the reference algorithm.