arXiv subjects

Arun Padakandla

Publications and source records attributed to Arun Padakandla.

24 records · Page 2Linked to original sources

arXiv2015

Achievable rate region for three user discrete broadcast channel based on coset codes

We present an achievable rate region for the general three user discrete memoryless broadcast channel, based on nested coset codes. We characterize 3-to-1 discrete broadcast channels, a class of broadcast channels for which the best known coding technique\footnote{We henceforth refer to this as Marton's coding for three user discrete broadcast channel.}, which is obtained by a natural generalization of that proposed by Marton for the general two user discrete broadcast channel, is strictly sub-optimal. In particular, we identify a novel 3-to-1 discrete broadcast channel for which Marton's coding is \textit{analytically} proved to be strictly suboptimal. We present achievable rate regions for the general 3-to-1 discrete broadcast channels, based on nested coset codes, that strictly enlarge Marton's rate region for the aforementioned channel. We generalize this to present achievable rate region for the general three user discrete broadcast channel. Combining together Marton's coding and that proposed herein, we propose the best known coding technique, for a general three user discrete broadcast channel.

cs.IT↗

arXiv2015

An Achievable rate region for the $3-$user interference channel based on coset codes

We consider the problem of communication over a three user discrete memoryless interference channel ($3-$IC). The current known coding techniques for communicating over an arbitrary $3-$IC are based on message splitting, superposition coding and binning using independent and identically distributed (iid) random codebooks. In this work, we propose a new ensemble of codes - partitioned coset codes (PCC) - that possess an appropriate mix of empirical and algebraic closure properties. We develop coding techniques that exploit algebraic closure property of PCC to enable efficient communication over $3-$IC. We analyze the performance of the proposed coding technique to derive an achievable rate region for the general discrete $3-$IC. Additive and non-additive examples are identified for which the derived achievable rate region is the capacity, and moreover, strictly larger than current known largest achievable rate regions based on iid random codebooks.

cs.IT↗

arXiv2013

Computing sum of sources over an arbitrary multiple access channel

The problem of computing sum of sources over a multiple access channel (MAC) is considered. Building on the technique of linear computation coding (LCC) proposed by Nazer and Gastpar [2007], we employ the ensemble of nested coset codes to derive a new set of sufficient conditions for computing the sum of sources over an \textit{arbitrary} MAC. The optimality of nested coset codes [Padakandla, Pradhan 2011] enables this technique outperform LCC even for linear MAC with a structural match. Examples of nonadditive MAC for which the technique proposed herein outperforms separation and systematic based computation are also presented. Finally, this technique is enhanced by incorporating separation based strategy, leading to a new set of sufficient conditions for computing the sum over a MAC.

cs.IT↗

arXiv2013

Achievable rate region based on coset codes for multiple access channel with states

We prove that the ensemble the nested coset codes built on finite fields achieves the capacity of arbitrary discrete memoryless point-to-point channels. Exploiting it's algebraic structure, we develop a coding technique for communication over general discrete multiple access channel with channel state information distributed at the transmitters. We build an algebraic coding framework for this problem using the ensemble of Abelian group codes and thereby derive a new achievable rate region. We identify non-additive and non-symmteric examples for which the proposed achievable rate region is strictly larger than the one achievable using random unstructured codes.

cs.IT↗

arXiv2008

The Size of Optimal Sequence Sets for Synchronous CDMA Systems

The sum capacity on a symbol-synchronous CDMA system having processing gain $N$ and supporting $K$ power constrained users is achieved by employing at most $2N-1$ sequences. Analogously, the minimum received power (energy-per-chip) on the symbol-synchronous CDMA system supporting $K$ users that demand specified data rates is attained by employing at most $2N-1$ sequences. If there are $L$ oversized users in the system, at most $2N-L-1$ sequences are needed. $2N-1$ is the minimum number of sequences needed to guarantee optimal allocation for single dimensional signaling. $N$ orthogonal sequences are sufficient if a few users (at most $N-1$) are allowed to signal in multiple dimensions. If there are no oversized users, these split users need to signal only in two dimensions each. The above results are shown by proving a converse to a well-known result of Weyl on the interlacing eigenvalues of the sum of two Hermitian matrices, one of which is of rank 1. The converse is analogous to Mirsky's converse to the interlacing eigenvalues theorem for bordering matrices.

cs.IT↗

arXiv2007

Separable convex optimization problems with linear ascending constraints

Separable convex optimization problems with linear ascending inequality and equality constraints are addressed in this paper. Under an ordering condition on the slopes of the functions at the origin, an algorithm that determines the optimum point in a finite number of steps is described. The optimum value is shown to be monotone with respect to a partial order on the constraint parameters. Moreover, the optimum value is convex with respect to these parameters. Examples motivated by optimizations for communication systems are used to illustrate the algorithm.

cs.IT↗