Se ofrecen los siguientes trabajos de fin de titulación:
TFM-IAS-2526-001: Explainable AI in orchestration frameworks: Understanding Framework Behavior in Energy Use Cases
DESCRIPTION:
This thesis investigates how an established framework behaves within an energy-focused use case, analyzing its actions and decisions through explainable AI methods to improve transparency and interpretability.
REQUIREMENTS:
- Basic knowledge of AI/ML concepts
- Programming skills in Python
- Interest in explainability and case study evaluation
TUTOR:
David Jiménez Bermejo (david.jimenezb@upm.es) / Giuseppe Conti (g.conti@upm.es)
TFM-IAS-2526-002: Design of an Energy Probe for Dataset Creation and Anomaly Detection
DESCRIPTION:
The work focuses on building an energy probe capable of generating datasets by recreating general conditions and forcing scenarios that reveal energy anomalies, supporting future predictive models.
REQUIREMENTS:
- Understanding of networks and energy systems
- Programming and data analysis (Python, R, MATLAB)
- Motivation for anomaly detection techniques
TUTOR:
David Jiménez Bermejo (david.jimenezb@upm.es) / Alberto del Rio Ponce (a.delriop@upm.es)
TFM-IAS-2526-003: Reinforcement Learning in MuJoCo: Replication and Enhancement of Prior Research
DESCRIPTION:
This thesis replicates and extends previous reinforcement learning experiments in MuJoCo environments, aiming to refine algorithms and improve performance in simulated control and robotics tasks.
REQUIREMENTS:
- Knowledge of reinforcement learning algorithms
- Experience with Python and RL libraries (e.g., Stable Baselines3)
- Interest in simulation and robotics
TUTOR:
Alberto del RÍo Ponce (a.delriop@upm.es) / Javier Serrano Romero (javier.serrano@upm.es)
TFM-IAS-2526-004: Benchmarking Multimodal Models: Extending Framework Functionalities with State-of-the-Art Approaches
DESCRIPTION:
The work extends an existing framework to evaluate different state-of-the-art multimodal models across tasks, benchmarking their performance and analyzing trade-offs of integration.
REQUIREMENTS:
- Familiarity with ML model evaluation
- Experience running ML experiments
- Basic understanding of multimodal or XR applications
TUTOR:
Alberto del Rio Ponce (a.delriop@upm.es) / Verónica Ruiz Bejerano (veronica.ruiz@upm.es) / Javier Serrano Romero (javier.serrano@upm.es)
TFM-IAS-2526-005: Orchestration of Large Language Models Based on Request Complexity
DESCRIPTION:
This thesis develops an orchestrator capable of selecting and deploying large language models dynamically, matching model selection to the complexity and requirements of user requests.
REQUIREMENTS:
- Solid programming skills (Python, APIs)
- Knowledge of LLM frameworks and tools
- Interest in orchestration and automation systems
TUTOR:
Alberto del Rio Ponce (a.delriop@upm.es) / Javier Serrano Romero (javier.serrano@upm.es)
TFM-IAS-2526-006: Knowledge Graph Integration with LLMs for Human–Machine Understanding of Technical Documentation
DESCRIPTION:
The work builds a reasoning layer on top of an existing knowledge graph derived from technical documentation, using large language models to enhance human–machine understanding and improve query responses
REQUIREMENTS:
- Background in semantic web and knowledge graphs
- Experience with NLP and text processing
- Familiarity with Python-based graph/NLP libraries
TUTOR:
Alberto del Rio Ponce (a.delriop@upm.es) / Verónica Ruiz Bejerano (veronica.ruiz@upm.es) / Javier Serrano Romero (javier.serrano@upm.es)
TFG-EES-2526-001: Energy Consumption Analysis: Probe-Based Estimations vs. Hardware Solutions
DESCRIPTION:
This thesis develops a network probe to estimate energy consumption, comparing the accuracy and efficiency of software-based estimations with results obtained from existing hardware solutions.
REQUIREMENTS:
- Background in computer networks
- Skills in statistical analysis and modeling
- Interest in energy efficiency and measurement
TUTOR:
Giuseppe Conti (g.conti@upm.es) / David Jiménez Bermejo (david.jimenezb@upm.es)
TFM-EES-2526-002: PV and Storage Solution Design for Electric Mobility at the A5 Interchange (Campamento)
DESCRIPTION:
This thesis designs and optimizes a photovoltaic and energy storage solution for a mobility hub supporting 100% electric vehicles in an urban renewal context, including energy needs assessment, system sizing, and integration strategies.
REQUIREMENTS:
- Background in renewable energy and PV systems
- Skills in simulation tools (PVSyst, HOMER, MATLAB)
- Interest in electromobility and sustainable urban planning
TUTOR:
David Jiménez (david.jimenezb@upm.es)
TFM-EES-2526-003: Design and Simulation of Additional Computing Nodes Deployment Powered by Energy Communities
DESCRIPTION:
This thesis evaluates the technical and energetic feasibility of deploying extra edge computing nodes powered by Local Energy Communities, modeling interconnected renewable sources and demand for optimal performance and sustainability.
REQUIREMENTS:
- Energy systems and distributed computing skills
- Experience with simulation and modeling
- Interest in edge computing and renewables
TUTOR:
David Jiménez (david.jimenezb@upm.es) / Javier Serrano (javier.serrano@upm.es) / Giuseppe Conti (g.conti@upm.es)
TFM-EES-2526-004: Exploratory Analysis of Large Language Model (LLM) Integration for Local Energy Community Management
DESCRIPTION:
This thesis explores and analyses the integration of LLMs for energy management in Local Energy Communities, including use cases, technical feasibility, proposed architectures, and simulation of their optimization impact.
REQUIREMENTS:
- AI and machine learning skills
- Programming (ideally Python)
- Interest in sustainable energy
TUTOR:
David Jiménez (david.jimenezb@upm.es)
TFG-EES-2526-005: Energy Community Generation Capacity Assessment Tool Using Aerial Image Analysis
DESCRIPTION:
Development and validation of a software tool to determine the renewable generation potential in urban areas for prospective energy communities by processing aerial and drone images.
REQUIREMENTS:
- Python programming, image processing (OpenCV, scikit-learn)
- ML knowledge, GIS basics
TUTOR:
David Jiménez (david.jimenezb@upm.es) / Giuseppe Conti (g.conti@upm.es)
TFG-EES-2526-006: PV Installation Protection Model Trainer for Self-Consumption Installations
DESCRIPTION:
Design and assembly of a physical trainer panel reproducing all necessary protection elements for residential/commercial PV self-consumption systems, including demonstration of real protection actions.
REQUIREMENTS:
- Electrical installations, circuit protection
- Practical assembly skills, safety focus
TUTOR:
David Jiménez (david.jimenezb@upm.es) / Giuseppe Conti (g.conti@upm.es)
TFG-EES-2526-007: Functional Demonstrator of Differential and Thermal-Magnetic Protection Devices
DESCRIPTION:
Construction of a demonstrator allowing controlled faults and visualization of the trip of differential and magneto-thermal protections for teaching and training purposes.
REQUIREMENTS:
- Electrical safety, components handling
- Skills in assembly and instructional materials
TUTOR:
David Jiménez (david.jimenezb@upm.es) / Giuseppe Conti (g.conti@upm.es)
TFG-EES-2526-008: Single-Phase Transformer Trainer: Demonstration and Basic Testing
DESCRIPTION:
Design and build a didactic trainer for single-phase transformers, enabling demonstration of principles, parameters, and standards-based testing for academic labs.
REQUIREMENTS:
- Electrotechnics and instrumentation
- Workshop and assembly abilities
TUTOR:
David Jiménez (david.jimenezb@upm.es) / Giuseppe Conti (g.conti@upm.es)
TFG-EES-2526-009: Local Environmental Data Acquisition for PV Generation Model Calibration
DESCRIPTION:
Construction and deployment of an environmental sensor system feeding local data (irradiance, temperature, etc.) into PV production forecast models, improving accuracy and adaptability.
REQUIREMENTS:
- Sensors, microcontrollers (Arduino, Raspberry Pi)
- Data acquisition, calibration, programming
TUTOR:
David Jiménez (david.jimenezb@upm.es) / Giuseppe Conti (g.conti@upm.es)
TFG-EES-2526-010: Energy Metric for Quantifying Impact of Training and Inference of One-Step and Two-Step AI Models on Edge Computing
DESCRIPTION:
Design and implementation of an energy metric to model and compare the impact of training and inference of single-step and dual-step AI models running in user-proximate networked computing systems.
REQUIREMENTS:
- AI, data analysis, programming
- Energy efficiency in distributed systems
TUTOR:
David Jiménez (david.jimenezb@upm.es) / Alberto del Rio (a.delriop@upm.es)
TFG-EES-2526-011: Electrical Experiment Escape Room for the ETSIT Electrotechnics Laboratory
DESCRIPTION:
Gamified lab experience: design and implementation of an escape room with sequential electrical experiments and safety-based challenges to boost student motivation and teamwork.
REQUIREMENTS:
- Basic electrotechnics, creativity, educational design
- Teamwork and organizational skills
TUTOR:
David Jiménez (david.jimenezb@upm.es)
TFG-EES-2526-012: Energy Consumption Monitoring Framework for Edge Computing Nodes
DESCRIPTION:
Implementation of a hardware/software framework for real-time energy monitoring in edge computing nodes, supporting analytics and facilitating system optimization.
REQUIREMENTS:
- Network and hardware programming (Python), IoT basics
- Interest in data analytics and energy management
TUTOR:
David Jiménez (david.jimenezb@upm.es) / Javier Serrano (javier.serrano@upm.es) / Alberto del Rio (a.delriop@upm.es)
TFG-TAV-2526-001: LLM-Assisted Multimedia Classification for QoE Estimation in Video Probes
DESCRIPTION:
This thesis applies basic large language models to classify multimedia content by genre (e.g., sports vs. cartoons), enabling quality probes to better adapt QoE estimation according to content-specific characteristics.
REQUIREMENTS:
- Familiarity with LLMs and NLP basics
- Knowledge of multimedia content types and formats
- Programming experience in PyTorch or TensorFlow
TUTOR:
Alberto del Rio Ponce (a.delriop@upm.es) / Álvaro Llorente Gómez (alvaro.llorente@upm.es) / David Jiménez Bermejo (david.jimenezb@upm.es)
TFG-TAV-2526-002: Advancing Remote Production: Evaluation of UHD Content with Voctomix 2.0
DESCRIPTION:
The research evolves existing remote production workflows by testing and evaluating Voctomix version 2.0 with UHD content, assessing its improvements and limitations compared to earlier deployments.
REQUIREMENTS:
- Background in multimedia networks or broadcasting
- Interest in video production tools and workflows
- Hands-on system integration skills
TUTOR:
Javier Serrano Romero (javier.serrano@upm.es) / Álvaro Llorente Gómez (alvaro.llorente@upm.es)
TFM-TAV-2526-003: Intelligent HD-to-UHD Upscaling System for Audiovisual Content
DESCRIPTION:
The project designs and implements a deep learning-based upscaler to convert HD video content to UHD quality, optimizing perceptual and objective metrics, and delivering a functional software tool.
REQUIREMENTS:
- Image/video processing and neural networks
- Python and deep learning frameworks
- Knowledge of video quality assessment
TUTOR:
David Jiménez (david.jimenezb@upm.es) / Álvaro Llorente (alvaro.llorente@upm.es) / Alberto del Rio (a.delriop@upm.es)
TFM-TAV-2526-004: Framework for Automated Verification of Age Rating Compliance in Audiovisual Content
DESCRIPTION:
Development of a modular system to automatically verify and audit content age classification using EPG extraction, logo recognition, and audio/text content analysis to ensure protection of minors on TV and streaming.
REQUIREMENTS:
- Audio/image processing and machine learning
- Python programming, media analysis libraries
- Interest in digital law and content regulation
TUTOR:
David Jiménez (david.jimenezb@upm.es) / Álvaro Llorente (alvaro.llorente@upm.es)
TFG-TAV-2526-005: Extension of Bjøntegaard Delta Metric with a Third Analysis Parameter
DESCRIPTION:
Implementation and evaluation of a new version of the BD metric for compression evaluation, incorporating a third parameter (e.g., spatial/global complexity) to improve discrimination in image/video assessment.
REQUIREMENTS:
- Image processing, Python or MATLAB
- Knowledge in metric analysis and statistics
TUTOR:
David Jiménez (david.jimenezb@upm.es) / Álvaro Llorente (alvaro.llorente@upm.es)
TFG-RSI-2526-001: Passive Network Probes with Kubernetes-Orchestrated AI for Bandwidth Estimation
DESCRIPTION:
The research designs AI-driven passive probes that estimate available bandwidth without disrupting active traffic, leveraging Kubernetes to orchestrate deployment, scalability, and management.
REQUIREMENTS:
- Understanding of network protocols and measurements
- Basic Kubernetes knowledge
- Programming and AI/ML fundamentals
TUTOR:
Alberto del Rio Ponce (a.delriop@upm.es) / Javier Serrano Romero (javier.serrano@upm.es)