Prof. Zehavi's Lab
Prof. Ephraim Zehavi is Deputy Dean of the Faculty of Engineering. Work in his lab focuses on coded communication systems in general, and in the areas of wideband wireless communication system in particular.
Research topics include wireless communications, advanced coding technology, MIMO channel tracking, the application of game theory to resource allocation, the coexistance of wireless networks, applications of analog coding for interference channels, and wide band communication systems (CDMA, OFDM).
Utilization of Game Theory for Resource Allocation
Zehavi and his group utilize game theory techniques to resolve conflict situations of sharing radio resources. Due to the high penetration rate of wireless communication systems for public and security usage, radio resources are set to become a very active area.
In contrast to previous work that was based on finding the Nash Equilibrium for the competition situation, Zehavi and his team adopted a cooperative approach (NBS, Nash Bargaining Solution) which increased the capacity gain for each of the players. In a series of papers they found the NBS for various constraints (mask constraint, average power constraint, MIMO, partial side information, and more) as well as a simple algorithm for computing NBS.
Zehavi and his group are currently working on extensions of their previous results on cooperation strategy for resource allocation. The extension is done in several directions, including partial FDM/TDM over frequency selective interference channel, Nash Barraging Solutio for frequency selective interference channel with partial side information, and other alternatives for cooperation. Future research in this area will provide an efficient cross layer approach for increasing the efficiency of future wireless system (Ad-Hoc networks, Cognitive Radio, etc).
Weighted max-min fairness with applications to UWB, WiMax and 3GPP-LTE
Orthogonal Frequency Division Multiple Access (OFDMA) is becoming the dominant technique for wireless multi access systems such as UWB, WLAN, WiMAX and LTE, due to its high spectral efficiency. OFDM waveforms provide the flexibility of allocating the subcarriers to the user in combating frequency selective fading. The total capacity of OFDMA can be optimized by dynamically allocating subcarriers among users according to channel condition. However, the operator must satisfy the subscribers’ demands for providing a reasonable level of Quality of Service (QOS).
The major challenges when considering QOS in wireless networks are the dynamic fluctuation of the channels, efficient usage of the spectrum, bandwidth allocation, and handoff support. It is important to guarantee QOS in each layer so that the network is more flexible and tolerant to QOS issues. Bandwidth allocation plays a major role in this respect; in some systems data services and voice services have to be supported simultaneously. On one hand voice services are very delay sensitive and require real-time service; on the other hand data services are less delay sensitive but are very sensitive to loss of data and expect error-free packets.
In their research, Zehavi and his team address the allocation of subcarriers based on a weighted max-min approach that sets the priority of the users according to a pre-set weight. They aim to extend this approach to guarantee a pre-set data rate for voice services and to allocate the residual capacity to data services.
In previous research, the power adaptation method was suggested to maximize the total data rate of multiuser in downlink of OFDM system. However, this approach does not allow for the fair sharing of resources. Zehavi and his group aim to introduce mechanisms necessary to enable explicit subcarrier allocation for multiple users in wireless systems, in order to show that weighted max-min fairness can assist in network optimization for multiple target rates and with a low computation load.
Tracking of unknown channel dynamics for MIMO applications
Zehavi’s group develops techniques for tracking the channel for MIMO applications and bounding the performance using gene aided Kalman ?lters. In this model a bidirectional multiple-input multiple-output (MIMO) time varying channel is considered.
The projection approximation subspace tracking (PAST) algorithm is used on both terminals in order to track the singular value decomposition of the channel matrix. Simulations using an autoregressive channel model and also a sampled MIMO indoor channel are performed, and the expected capacity degradation due to the estimation error is evaluated. Simulations were performed using an auto-regressive (AR) fading model for the channel matrix. The average capacity loss of this scheme relative to a scheme with full CSI was calculated for different AR parameters and different values of Q, the ”forgetting factor” of the PASTd algorithm.
The group also develops performance bounds for tracking time-varying OFDM MIMO communication channels in the presence of additive white Gaussian noise (AWGN).
Handoff between cellular and WLAN networks
The integration of cellular networks and WLAN networks enables them to increase overall capacity and services. The two networks possess complementary features. On one hand, cellular networks provide wide area coverage at medium data rates and enable connectivity to mobile devices while providing poor coverage for indoor subscribers. On the other hand, WLAN networks provide local coverage indoors with high data rates.
Two approaches to integrating the networks have been proposed in literature: the tightly coupled approach and the loosely coupled approach. Zehavi and his group proposed a third approach, the dual coupling approach, which simpli?es the integration of the two networks and provides a seamless handoff. This approach increases the capacity of the cellular network by four times for most typical situations, due to the use of WLAN resources for voice services. In addition, the QOS of the voice service increases.