Error Performance Optimization Using Logarithmic Barrier Function in Molecular Nanonetworks

Abstract

The performance of a molecular signal detector relies on selected value of the detection threshold. In fact, arbitrary choice of detection threshold would increase either the probability of miss detection or the probability of false alarm. Thereby, an appropriate value of detection threshold is required in order to optimize the error performance of a realistic diffusion-based molecular communication (DbMC) system. For this, we come up with a new approach that yields optimal value for convex optimization problem in DbMC using logarithmic barrier followed by modified Karush-Kuhn-Tucker conditions, Newton Raphson method, and finally rounding the solution to the nearest integer value. Previous works on threshold optimization in DbMC either give suboptimal solution or optimal one whereof the convergence depends on choosing the smallest interval that contains the optimal value. Numerical and simulation results reveal that our proposed methodology can provide optimal solution with utmost accuracy in few iterations. © 1997-2012 IEEE.