TFM: DESIGN AND IMPLEMENTATION OF AN ADAPTER FOR COMMUNICATIONS THROUGH COGNTIVE RADIO

This work is part of the ROBIM project in which the working group B105 Electronic Systems Lab of the University Universidad Politécnica de Madrid collaborates. The ROBIM project takes part in the program Programa Estratégico CIEN with the support of the CDTI (Centro para el desarrollo tecnológico Industrial) and the RDF (Regional Development Forum) for Europe.

The ROBIM project seeks to automate technical inspections of buildings, reducing costs and execution times associated with these processes. The system makes use of a drone for inspection work, thus avoiding the installation of scaffolding and all the security measures that the process requires, which is costly in time and money. Currently, the drone has a communication channel that allows users to obtain information on the process, as well as direct the drone whenever necessary.
The main objective of this work is to create a secondary, safe and effective communication channel, for situations where communication with the main system is not possible. To achieve this, the project stablish the following requierements:

– The device must allow radiocommunication in ISM bands.
– The device has an USB interface to connect with the computer/drone.
– The communication must be reliable by allowing communication throwgh various channels and implementing software-defined radio and cognitive radio.

Therefore, to achieve these objectives, this work proposes the design of a 2-channel device for radiocommunication in the 433 MHz and 868 MHz bands, using two SPIRIT1 transceivers and an ARM Cortex-M4 microcontroller.

Picture of the device’s high-level design

The Hardware design has been made usign the Altium Designer PCB design layout tool . The designed PCB is divided into three parts: the power/communication stage, the control stage with the microcontroller and the radiofrecuency stage with both SPIRIT1 trasnceivers.

Picture of the 3D reconstruction of the board designed in Altium Design tool

The software design has been developed in 2 stages: software design of an application for evaluation boards during the PCB manufacturating process and software design of a final application for the designed PCB.
For the software design of evaluation board, the NUCLEO – L053R8 with the X-NUCLEO-IDS01A4 radio frequency module has been chosen, which allows radio communication in the 868 MHz band. The final design of the software is based on the software of the evaluation board but improving its functionality by adding communication through two channels with a cognitive procedure based on the CSMA / CA protocol and implementing serial communication with the user.

The application designed for the device allows, then, a cognitive communication based on CSMA/CA protocol in bands 433 MHz and 868 MHz in addition to communication with the user and the drone enabling the possibility of the implementation of the second channel for the communication with the drone.

Gated Recurrent Unit Neural Networks for Automatic Modulation Classification With Resource-Constrained End-Devices

The article “Gated Recurrent Unit Neural Networks for Automatic Modulation Classification With Resource-Constrained End-Devices” by our lab member Ramiro Utrilla has just been published in the IEEE Access, a high-impact open-access journal.

This work has been carried out in collaboration with researchers from the CONNECT – Centre for Future Networks and Communications in Dublin (Ireland), where Ramiro carried out a research stay of 3 months.

In this article, they focus on the Automatic Modulation Classification (AMC). AMC is essential to carry out multiple CR techniques, such as dynamic spectrum access, link adaptation and interference detection, aimed at improving communications throughput and reliability and, in turn, spectral efficiency. In recent years, multiple Deep Learning (DL) techniques have been proposed to address the AMC problem. These DL techniques have demonstrated better generalization, scalability and robustness capabilities compared to previous solutions. However, most of these techniques require high processing and storage capabilities that limit their applicability to energy- and computation-constrained end-devices.

In this work, they propose a new gated recurrent unit neural network solution for AMC that has been specifically designed for resource-constrained IoT devices.


The proposed GRU network model for AMC.

They trained and tested their solution with over-the-air measurements of real radio signals, which were acquired with the MIGOU platform.


Dataset generation scenario set up.


Comparison of signals recorded at (a) 1 and (b) 6 meters. The signals in the bottom row are the normalized version of those in the top row.

Their results show that the proposed solution has a memory footprint of 73.5 kBytes, 51.74% less than the reference model, and achieves a classification accuracy of 92.4%.

Increasing the training set can lead to improvements in the performance of a model without increasing its complexity. These improvements allow developers to reduce the complexity of the model and, therefore, the device resources it requires. However, longer training processes can lead to fitting and gradient problems. These tradeoffs should be explored when developing neural network-based solutions for resource-constrained end-devices.