Accessible Capacity of Secondary Users

TL;DR

This paper investigates the maximum reliable transmission rate for secondary users in Gaussian interference channels without disrupting primary users' error performance, modeling primary encoders as generalized trellis codes and deriving bounds on accessible capacity.

Contribution

It introduces a new problem formulation for secondary access in interference channels, modeling primary encoders as GTCs and deriving bounds on accessible capacity considering primary error constraints.

Findings

Derived upper and lower bounds on accessible capacity.

Analyzed the impact of primary error performance on secondary access.

Numerical results illustrate the effects of GTC and interference margin.

Abstract

A new problem formulation is presented for the Gaussian interference channels (GIFC) with two pairs of users, which are distinguished as primary users and secondary users, respectively. The primary users employ a pair of encoder and decoder that were originally designed to satisfy a given error performance requirement under the assumption that no interference exists from other users. In the scenario when the secondary users attempt to access the same medium, we are interested in the maximum transmission rate (defined as {\em accessible capacity}) at which secondary users can communicate reliably without affecting the error performance requirement by the primary users under the constraint that the primary encoder (not the decoder) is kept unchanged. By modeling the primary encoder as a generalized trellis code (GTC), we are then able to treat the secondary link and the cross link from the secondary transmitter to the primary receiver as finite state channels (FSCs). Based on this, upper and lower bounds on the accessible capacity are derived. The impact of the error performance requirement by the primary users on the accessible capacity is analyzed by using the concept of interference margin. In the case of non-trivial interference margin, the secondary message is split into common and private parts and then encoded by superposition coding, which delivers a lower bound on the accessible capacity. For some special cases, these bounds can be computed numerically by using the BCJR algorithm. Numerical results are also provided to gain insight into the impacts of the GTC and the error performance requirement on the accessible capacity.