Reducing transceiver complexity and power consumption are important design factors of next-generation radio communication systems. In this paper, a dual-band single-chain radio system is demonstrated. The Digital Front-End (DFE) is based on the AMD RFSoC$\mathbf{4}\times \mathbf{2}$platform, and the Radio Frequency Front-End (RF-FE) is based on a connectorised architecture using commercial-off-the-shelf (COTS) products. Measured results characterise the error vector magnitude (EVM) and block error rate (BLER) of a dual-band, single-chain radio system across four modulation schemes, demonstrating comparable performance in both frequency bands. In addition, single-band measurements were conducted for comparison, revealing that the dual-band configuration maintains performance levels similar to those of the single-band setup. This highlights the advantages of the dual-band architecture, which supports multiple frequency bands without compromising performance. It significantly reduces transceiver complexity and power consumption, presenting a compelling alternative to implementing dual-band transceivers by duplicating radio chains.

Reducing transceiver complexity and power consumption are critical considerations in the design of next-generation radio communication systems. This work demonstrates a dual-band, single-chain radio unit for neutral host Open RAN applications. The radio unit features a Digital Front-End (DFE) implemented on the AMD RFSoC$4\mathrm{x}2$platform and a Radio Frequency Front-End (RF-FE) constructed using connectorised commercial-off-the-shelf (COTS) components. The error vector magnitude (EVM) and block error rate (BLER) performance of the dual-band radio system are measured under customisable test conditions, highlighting the radio's flexibility.

Abstract Automatic Modulation Classification (AMC) is a rapidly evolving technology, which can be employed in software defined radio structures, especially in 5G and 6G technology. Machine Learning (ML) can provide novel and efficient technology for modulation classification, especially for systems working in low signal to noise ratio (SNR). In this article, two dynamic systems not reliant on received signal phase lock and frequency lock are presented, with both employing ML to classify the modulation types for different received SNR. The first model is developed from the previous existing literatures, which utilises constellation images (CI) and image classification technology. Here, modulation types can be detected in a dynamic way without phase lock and frequency lock. In the second model, a new method named Graphic Representation of Features (GRF) is proposed, which represents the statistical features as a spider graph for ML. The concepts are tested and verified using simulations and RF data using a lab software defined radio (SDR). The results from the two models are compared. With the GRF techniques an overall classification accuracy of 59% is observed for 0 dB SNR and 86% at 10 dB SNR, compared to a random guess accuracy of 25%.

This article presents a distributed and decentralized architecture for the implementation of distributed artificial intelligence (DAI) using hardware platforms provided by the Internet of Things (IoT). A trained DAI system has been implemented over IoT, where each IoT device acts as one or more of the neurons within the DAI layers. This is accomplished by the utilization of decentralized, self-managed blockchain technologies that allow trusted interactions and information to be exchanged between distributed neurons. The platform was built and customized to be used within the IoT system, and it is capable of handling DAI-related tasks. A new consensus mechanism based on Proof of Authority (PoA) and Proof of Work (PoW) has been designed and implemented, along with bespoke block and transaction formats. The proposed architecture was analyzed, implemented, and tested using a dedicated testbed with low-cost IoT devices. A quantitative measurement and performance evaluation of the system based on a real-world IoT application was conducted. The implemented DAI is found to have an accuracy of 92%–98%, with an energy cost of 0.12 joules (J) when utilizing a Raspberry Pi to run one neuron. The measured hash per joule (h/J) when using a Raspberry Pi for mining is 13.8 Kh/J compared to 54 Kh/J using an ESP32. The results showed that it is feasible to implement a DAI system utilizing the IoT hardware platform while maintaining the system’s accuracy. The integration of the blockchain has added an element of security and trust to the data and the interaction between system components.

To reduce energy consumption, a device in a long range wide area network (LoRaWAN) needs to adjust its transmit power according to the distance from its tagged gateway. It is important to measure the performance of the LoRaWANs uplink with power control. In this article, we focus on the analysis of the coverage probability and the meta distribution of the signal-to-interference ratio (SIR) for a LoRaWAN uplink with fractional power control (FPC). The LoRaWAN uplink is analyzed based on the Poisson point process. We present the possible reductions in transmit power of devices whilst ensuring that the received signal power is greater than the receiver sensitivity. We derive the coverage probability of a LoRaWAN uplink and show how power control influences it. Finally, utilizing the meta distribution of SIR, the fine-grained information of the LoRaWAN is revealed. The results show that the power control greatly increases the successful probability of edge-devices with little effect on the probability of inner-devices if an appropriate FPC coefficient is chosen. This is because the LoRa signal can be demodulated at a very low required SIR threshold.