On Energy Allocation and Data Scheduling in Backscatter Networks with Multi-antenna Readers

TL;DR

This paper develops online policies for data admission and link control in multi-antenna backscatter networks, optimizing throughput utility while stabilizing buffers, using the MDPP method and advanced optimization techniques.

Contribution

It introduces novel online control policies for backscatter networks that jointly optimize throughput utility and buffer stability, employing the MDPP framework and solving complex non-convex problems.

Findings

Proposed policies achieve optimal throughput utility.

Buffer stability is maintained under the policies.

System performance with short packets is effectively analyzed.

Abstract

In this paper, we study the throughput utility functions in buffer-equipped monostatic backscatter communication networks with multi-antenna Readers. In the considered model, the backscatter nodes (BNs) store the data in their buffers before transmission to the Reader. We investigate three utility functions, namely, the sum, the proportional and the common throughput. We design online admission policies, corresponding to each utility function, to determine how much data can be admitted in the buffers. Moreover, we propose an online data link control policy for jointly controlling the transmit and receive beamforming vectors as well as the reflection coefficients of the BNs. The proposed policies for data admission and data link control jointly optimize the throughput utility, while stabilizing the buffers. We adopt the min-drift-plus-penalty (MDPP) method in designing the control policies. Following the MDPP method, we cast the optimal data link control and the data admission policies as solutions of two independent optimization problems which should be solved in each time slot. The optimization problem corresponding to the data link control is non-convex and does not have a trivial solution. Using Lagrangian dual and quadratic transforms, we find a closed-form iterative solution. Finally, we use the results on the achievable rates of finite blocklength codes to study the system performance in the cases with short packets. As demonstrated, the proposed policies achieve optimal utility and stabilize the data buffers in the BNs.