Section: New Results

Wireless Networks

Participants : Eitan Altman, Abdulhalim Dandoush, Majed Haddad, Jithin Kazhuthuveettil Sreedharan.

Range-based localization algorithms in wireless sensor networks are more accurate but also more computationally complex than the range-free algorithms. In collaboration with M. S. Elgamel (Univ. of Louisiana, USA), A. Dandoush has revised the Trigonometric based Ad-hoc Localization System (TALS) proposed in the literature. In [83] , they propose a new technique to optimize the system: by eliminating the need of solving a linear system of equations via least square methods or its variants or the need for any square root operations, the computational overhead is reduced. Also, a novel modified Manhattan distance is proposed and used in the elimination process ensuring thereby a very good accuracy with less complexity than the basic TALS. Through a mathematical analysis and intensive simulations, the optimized TALS is shown to present superior performance and accuracy results compared to other localization techniques.

The enhanced Inter Cell Interference Coordination (eICIC) feature has been introduced to solve the interference problem in small cells. It involves two parameters which need to be optimized, namely the Cell Range Extension (CRE) of the small cells and the ABS ratio (ABSr) which defines a mute ratio for the macro cell to reduce the interference it produces. In [72] , A. Tall, Z. Altman (Orange Labs, Issy les Moulineaux) and E. Altman propose self-optimizing algorithms for the eICIC. The CRE is adjusted by means of a load balancing algorithm. The ABSr parameter is optimized by maximizing a proportional fair utility of user throughputs. The convergence of the algorithms is proven using Stochastic Approximation theorems. Numerical simulations illustrate the important performance gain brought about by the different algorithms.

Cognitive Radios are proposed as a solution to scarcity of wireless spectrum and one of the main challenges here is to gain knowledge about the spectrum usage by the licensed users, termed as spectrum sensing. In [29] , Vinod Sharma (Indian Institute of Science, Bangalore, India) and J. K. Sreedharan study novel algorithms for spectrum sensing which minimize the expected time for spectrum sensing with stringent constraints on the probability of wrong detection. Algorithms are distributed in nature and the work proves that the algorithms are asymptotically optimal distributed sequential hypothesis tests. Along with theoretical guarantees, many practical scenarios in Cognitive Radios are also investigated.

The fast development of SON technology in mobile networks renders critical the problem of coordinating SON functionalities operating simultaneously. SON functionalities can be viewed as control loops that may need to be coordinated to guarantee conflict free operation, to enforce stability of the network and to achieve performance gain. In [30] , A. Tall and Z. Altman (Orange Labs, Issy les Moulineaux), R. Combes (Supelec ), and E. Altman propose a distributed solution for coordinating SON functionalities. It uses Rosen's concave games framework in conjunction with convex optimization. The SON functionalities are modeled as linear Ordinary Differential Equation (ODE)s. The stability of the system is first evaluated using a basic control theory approach together with strict diagonal concavity notion that originates from game theory. The coordination solution consists in finding a linear map (called coordination matrix) that stabilizes the system of SON functionalities. It is proven that the solution remains valid in a noisy environment using Stochastic Approximation.