[Paper Review] Receiver Antenna Partitioning for Simultaneous Wireless Information and Power Transfer

Rahul Vaze, Jainam Doshi+1 more|arXiv (Cornell University)|Oct 6, 2014

Energy Harvesting in Wireless Networks12 references3 citations

TL;DR

This paper proposes an optimal antenna partitioning scheme for simultaneous wireless information and power transfer (SWIPT) in MIMO systems, where a mobile device dynamically assigns its receive antennas to either information reception or energy harvesting. By modeling the problem as maximizing a submodular function under a matroid constraint, the authors design greedy algorithms that guarantee solutions within (1−1/e) and 1/2 of the optimal capacity, significantly outperforming worst-case bounds in simulations.

ABSTRACT

Powering mobiles using microwave \emph{power transfer} (PT) avoids the inconvenience of battery recharging by cables and ensures uninterrupted mobile operation. The integration of PT and \emph{information transfer} (IT) allows wireless PT to be realized by building on the existing infrastructure for IT and also leads to compact mobile designs. As a result, \emph{simultaneous wireless information and power transfer} (SWIPT) has emerged to be an active research topic that is also the theme of this paper. In this paper, a practical SWIPT system is considered where two multi-antenna stations perform separate PT and IT to a multi-antenna mobile to accommodate their difference in ranges. The mobile dynamically assigns each antenna for either PT or IT. The antenna partitioning results in a tradeoff between the MIMO IT channel capacity and the PT efficiency. The optimal partitioning for maximizing the IT rate under a PT constraint is a NP-hard integer program, and the paper proposes solving it via efficient greedy algorithms with guaranteed performance. To this end, the antenna-partitioning problem is proved to be one that optimizes a sub-modular function over a matroid constraint. This structure allows the application of two well-known greedy algorithms that yield solutions no smaller than the optimal one scaled by factors $(1-1/e)$ and 1/2, respectively.

Motivation & Objective

• To address the challenge of maximizing information transfer (IT) capacity in SWIPT systems under a power transfer (PT) constraint.
• To design a practical antenna partitioning scheme that assigns each mobile receiver antenna exclusively to either IT or PT, avoiding complex power splitting.
• To establish theoretical performance guarantees for the antenna partitioning problem using submodular optimization.
• To extend prior work on single-input-multiple-output (SIMO) SWIPT to the general MIMO case with full channel state information at the receiver (CSIR) and transmitter (CSIT).
• To provide efficient algorithms with provable approximation ratios for a NP-hard integer programming problem in SWIPT.

Proposed method

• Models the SWIPT antenna partitioning problem as an integer program that maximizes MIMO IT capacity under a PT efficiency constraint.
• Proves that the mutual information objective function is submodular under both CSIR and CSIT assumptions, enabling use of greedy optimization techniques.
• Identifies the circuit power constraint as a matroid constraint, allowing application of well-known submodular optimization algorithms.
• Applies two greedy algorithms: continuous greedy with pipage rounding (guaranteed (1−1/e)-approximation) and a simple greedy algorithm (guaranteed 1/2-approximation).
• Uses waterfilling power allocation for optimal power distribution across selected IT antennas under total power constraints.
• Derives theoretical performance bounds and validates them via simulations under Rayleigh fading channels.

Experimental results

Research questions

• RQ1 How can the optimal partitioning of receiver antennas between information and power transfer be achieved in a MIMO SWIPT system?
• RQ2 What is the computational complexity of the antenna partitioning problem, and can it be solved efficiently with performance guarantees?
• RQ3 Can the mutual information expression in SWIPT be modeled as a submodular function under realistic channel assumptions?
• RQ4 What approximation guarantees can be provided for greedy algorithms applied to the SWIPT antenna partitioning problem?
• RQ5 How do the proposed algorithms compare to worst-case theoretical bounds in practical fading environments?

Key findings

• The antenna partitioning problem for SWIPT is proven to be equivalent to maximizing a submodular function under a matroid constraint, enabling efficient approximation.
• The continuous greedy algorithm with pipage rounding guarantees a solution within (1−1/e) ≈ 63% of the optimal IT capacity under both CSIR and CSIT conditions.
• The simple greedy algorithm guarantees a solution within at least 50% of the optimal capacity, with no additional computational overhead.
• Simulations show both algorithms significantly outperform their worst-case theoretical bounds in Rayleigh fading channels, especially the greedy algorithm which outperforms the continuous greedy + pipage rounding method.
• The proposed approach achieves higher spectral efficiency than prior methods that rely on power splitting or relaxed dual-mode antenna operation.
• The performance gain is consistent across varying numbers of receiver antennas, with the circuit power constraint scaled proportionally to maintain fairness in comparison.