Optimal Deployment of Phased Array Antennas for RFID Network Planning Based on an Improved Chicken Swarm Optimization

Weiguang Shi,Wei Wang,Yang Yu,Shanshan Zhang,Yu Cao,Shuxia Yan,Junchao Gao

DOI: https://doi.org/10.1109/jiot.2021.3067013

IF: 10.6

2021-10-01

IEEE Internet of Things Journal

Abstract:Effective network planning improves performance in the radio-frequency identification (RFID) system. This article proposes an optimal deployment of phased array reader antennas for RFID network planning (RNP). For practical considerations, the RNP problem is analyzed and formulated based on a multipath propagation model, where each reader is equipped with a phased array antenna. The gains and radiation directions of the antennas are adjusted by voltage states instead of gestures, reducing the cumbersome antennas redeployment. An indicator called amplitude fluctuation under narrowband (AFN) is proposed to reflect the disturbance of frequency selective fading caused by the multipath effect. To effectively address the RNP problem, an improved chicken swarm optimization algorithm with two targeted strategies is developed. Simulation and experiment comparisons with the existing algorithms demonstrate the superiority of the proposed approach.

computer science, information systems,telecommunications,engineering, electrical & electronic

What problem does this paper attempt to address?

The paper primarily addresses the network planning problem (RNP) in ultra-high frequency radio frequency identification (UHF RFID) systems, aiming to propose an optimization scheme to improve system performance. Specifically, the paper tackles the following key issues:

1. **Reducing the need for physical adjustments**: Traditional RNP methods often consider only dipole antennas, which leads to frequent physical adjustments (such as changing the direction or position of the antenna) during actual deployment. This paper introduces phased array antennas, which can change the direction and gain of the antenna by adjusting the voltage state rather than physical gestures, thereby simplifying the deployment process.

2. **Simulating link states under multipath effects**: In indoor environments, multipath effects significantly impact signal transmission. Existing RNP solutions rarely consider these effects, resulting in optimization outcomes that do not match actual conditions. This paper establishes a simulation environment that includes a multipath propagation model to more accurately simulate link states.

3. **Mitigating the impact of frequency-selective fading**: Commercial UHF RFID systems typically operate within a narrow bandwidth and have multiple frequency channels. Multipath effects cause signals on different frequency channels to suffer varying degrees of attenuation or enhancement, known as frequency-selective fading (FSF). Existing methods usually assume a fixed frequency rather than a varying frequency, ignoring the impact of FSF. This paper proposes an indicator—the amplitude fluctuation of the narrowband channel transmission function (AFN)—to assess the impact of FSF on system performance and incorporates it into the optimization objectives.

4. **Improving the performance of optimization algorithms**: The RNP problem is a non-deterministic polynomial-time problem, and traditional deterministic algorithms are difficult to solve within an acceptable time frame. Therefore, most existing solutions are based on evolutionary computation (EC) or swarm intelligence (SI) algorithms. This paper proposes an improved chicken swarm optimization algorithm (CSO), including two strategies—monomer perturbation in the chicken swarm (MTR) and particle update in hen movement (PRH)—to further enhance optimization effectiveness.

In summary, the main contributions of this paper are the establishment of a propagation model that combines the characteristics of phased array antennas and multipath effects, the proposal of a new RNP problem model that considers frequency-selective fading, and the development of an improved chicken swarm optimization algorithm to solve the RNP problem. The effectiveness and superiority of the proposed methods are comprehensively validated through simulations and applied in a practical case.