2 July 2024 in IEEE Access
Currently, the Message Queuing Telemetry Transport for Sensor Networks (MQTT - SN) protocol has been implemented in operating systems such as TinyOS and Contiki to operate above the network layer, in sensor nodes with low processing capacity, powered by batteries and operating in various topologies. However, minimizing the processes at the node side is necessary to make the MQTT - SN protocol applicable in wireless sensor networks (WSNs) with large-scale linear structures. In this context, this paper presents the development of the MQTT-SN protocol’s finite state machine (FSM) for its operation over IEEE 802.15.4 in linear topologies. For this purpose, an FSM for each of the twelve procedures indicated in the MQTT-SN specification is obtained. Furthermore, each of the FSMs is represented using the specification and description language, and the characteristics of the IEEE 802.15.4 protocol and the sensor node are considered. Through simulation and the exchange of messages between nodes, the operation of the FSMs is verified. Subsequently, implementing the FSMs in WSNs allows for validating the deployment of the MQTT-SN protocol in linear topologies.
26 June 2024 in Enfoque UTE
Passive Optical Networks (PONs) are essential in optical communications to meet the increasing demand for network capacity and connected users, ensuring reliable and adaptable connections for data transmissions. These networks also simplify infrastructures and efficiently utilize energy by eliminating the need for active devices. In this context, Ultra- Dense Wavelength Division Multiplexing (UDWDM) is one of the most prominent solutions for data transmission. This technology uses the narrow separation between channels from 25GHz and even as small as 6.25GHz to increase the transmission capacity. This document analyzes the performance of UDWDM considering the transmission of three simultaneous channels in a PON network for which three different scenarios have been considered for the analysis with the following parameters: i) transmission speed from 10 Gbps to 17 Gbps, ii) the distance from 10 km to 20 km and iii) the separation between channels of 15GHz, 20GHz, and 25GHz. The performance metrics for the analyzed scenario are the bit error rate and the eye diagram. To ensure the reception of the transmitted channel, an analysis of raised cosine and Gauss filters is also discussed. Including these filters allows for verifying whether their use enhances the performance of the channels compared to transmission without a filter. This is crucial for understanding how filters can optimize the quality and reliability of data transmission in a PON, which is of great importance in environments where high efficiency and connection quality are required.
9 February 2022 in bioRxiv
Captive environments trigger the propagation and multiplication of parasites among different reptile species, thus weakening their immune response and causing infections and diseases. Technological advances of convolutional neural networks have opened a new field for detecting and classifying diseases which have shown great potential to overcome the shortcomings of manual detection performed by experts. Therefore, we propose an approach to identify six captive reptiles parasitic agents ( Ophionyssus natricis, Blastocystis sp, Oxiurdo egg, Rhytidoides similis, Strongyloides, Taenia ) from a microscope stool images dataset. Towards this end, we first use an image segmentation stage to detect the parasite within the image, which combines the Contrast Limited Adaptive Histogram Equalization (CLAHE) technique, the OTSU binarization method, and morphological operations. Then, we carry out a classification stage through Mobilenet CNN under a transfer learning scheme. This method was validated on a stool image dataset containing 3616 images data samples and 26 videos from the six parasites mentioned above. The results obtained indicate that our transfer learning-based approach can learn a helpful representation from the dataset. We obtained an average accuracy of 98.66% across the six parasitic agent classes, which statistically outperformed, at a 95% confidence level, a custom CNN trained from scratch.
1 January 2025 in IEEE Access
1 January 2026 in IEEE Transactions on Computational Social Systems