DESI - Trabajos de fin de Especialidad en Sistemas Embebidos
https://hdl.handle.net/11117/3428
2024-03-28T13:51:16ZDevelopment of a Cost-Efficient Video Acquisition System for Medical Applications Using a Dual-Core Cortex-M Microcontroller
https://hdl.handle.net/11117/10265
Development of a Cost-Efficient Video Acquisition System for Medical Applications Using a Dual-Core Cortex-M Microcontroller
Bautista-Guerrero, Snaider A.
Video acquisition systems, consisting of video capture, encoding, transmission and display, are widely used in medical applications. However, current systems present challenges such as higher costs, power consumption, and reduced availability. Therefore, this paper proposes a video acquisition system using a dual-core Cortex-M microcontroller. Tests showed the system is suitable for video acquisition, processing, and display. The storage component presented limitations, by only achieving the correct saving of 32.36 % of the acquired video. Finally, it is shown that a Dual-Core Cortex-M microcontroller could be used for low-cost medical applications that do not require high quality video storage.
2023-07-01T00:00:00ZDeployment of Machine Learning Algorithm to Predict Battery Behavior
https://hdl.handle.net/11117/10264
Deployment of Machine Learning Algorithm to Predict Battery Behavior
Flores-Triana, Jorge A.; Cinco-Ahumada, Jesus A.
The growth of the electric car industry has increased in recent years, along with the trend of green energy around the world. For this reason, automotive companies have invested in finding different solutions to monitor lithium batteries that power vehicles. These applications include State of Charge (SoC) and State of Health (SoH) analysis of the battery cells by monitoring key variables such as temperature, current, and voltage to predict the behavior of the system and apply preventive maintenance.
In this paper, a deep neural network using the Deep Learning MATLAB Toolbox was designed to predict the SoC from an emulated battery in Simulink. The model was then compiled and deployed in an NXP S32K344 microcontroller using the NXP Model-Based Design Toolbox. The results obtained showed a network with up to 90% accuracy and an execution time of 2.6 ms when running the core at 160 MHz.
2023-07-01T00:00:00ZGeolocalization and Diagnostic Status Over LEOs and 4G Networks
https://hdl.handle.net/11117/10263
Geolocalization and Diagnostic Status Over LEOs and 4G Networks
MartÃnez-Torres, Josue J.; MartÃnez-Preciado, Dorian R.
In modern times, communication between remote devices in the field is critical for monitoring the correct functionality, unexpected events, faults, and automated processes. This ensures system stability, recovery, and alert message transmission. This paper describes a system that integrates two modems working in parallel with a microcontroller to create a monitoring device. This device was used and tested to diagnose and control an external device in real time.
2023-07-01T00:00:00ZCAN-Ethernet Gateway for Visualizing CAN Network Packets in Wireshark
https://hdl.handle.net/11117/10262
CAN-Ethernet Gateway for Visualizing CAN Network Packets in Wireshark
Ruiz-Ochoa, Juan C.
In automotive control systems, the analysis of CAN messages is key to identify and correct errors. This paper presents a CAN-Ethernet Gateway, which allows knowledge acquisition from the analysis CAN systems. The FRDM-K64F board is utilized as it supports both CAN and UDP communication protocols. In addition, a CAN transceiver is used to establish the connection with the Vector CAN interface. Therefore, all gateway functionality is visualized in the computer. For testing, CANoe with an ITESO license is used to enable the simulation of the CAN system. The Gateway successfully captures CAN messages of different sizes and facilitates their analysis in Wireshark by displaying each data in a structured format, which enables enhanced and efficient monitoring of automotive systems.
2023-07-01T00:00:00Z