Oportunidades de Parceiros

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos (cristina@dei.uminho.pt) and João Lopes

Descrição:

In robotics-based rehabilitation, fostering effective human-robot cooperation and interaction is essential to achieve successful therapeutic outcomes. Effective rehabilitation involves early, intensive, and repetitive practice, but it is equally important to ensure that therapies are engaging and motivating to prevent patient discouragement. Health-oriented games, commonly referred to as serious games, have the potential to captivate and motivate patients, encouraging their active participation and potentially enhancing motor recovery. By leveraging Unity, it is possible to design diverse and immersive therapeutic experiences through multimodal games combined with biofeedback strategies for gait, posture, and balance training. This proposal aims to develop a serious game with a practical clinical context, integrated into a robotic walker, to facilitate intuitive and enjoyable rehabilitation. Innovative solutions include to introduce reinforcement learning for exploring learning according to the evolution of the person, such that the difficulty increases according to the profile of the person. An important aspect of this work is the opportunity to test the developed serious game with real ataxic patients, allowing for valuable insights and continuous improvements based on real-world feedback. This approach seeks to foster active patient involvement, strengthen human-robot interaction, and ultimately contribute to a faster and more effective recovery process.

This opportunity may be adapted for a master dissertation or for an extracurricular internship.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos (cristina@dei.uminho.pt)

Descrição:

Resting tremors are the most apparent and wellknown symptom of Parkinson’s Disease (PD), that affects the daily living of these patients. The use of electroencephalography (EEG)-based Brain Computer Interface systems provide the opportunity to create a connection between the user's imagination and the command for functional electrical stimulation (FES), providing physical feedback during the therapy when the movement is correctly imagined. This creates the possibility of neurorehabilitation for neurologic patients as PD. This proposal aims to develop an BCI system for the reduction of tremors in PD patients. The patients will be instructed to perform motor imagery focusing on a movement without tremor and they will receive FES for tremor reduction if the movement is correctly imagined.

Objectives:

1. State-of-the-art of the most recently developed EEG-based BCI-FES system for neurorehabilitation.

2. Development of the EEG-based BCI-FES system for the reduction of tremors.

3. Design and implementation of an experimental validation protocol to evaluate the developed system's effects on patients' motor performance and EEG signals.

4. Processing of the acquired data and discussion by comparison with literature.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos (cristina@dei.uminho.pt); Joana Almeida (pg50453@alunos.uminho.pt)

Descrição:

This dissertation proposes an adaptive shared control framework between an upper-limb exoskeleton and a robotic arm to reduce muscular load and manage fatigue during repetitive tasks. By integrating reflex-inspired control with physical assistance, the system dynamically adjusts support based on user effort and fatigue levels. This hybrid approach aims to enhance user endurance and efficiency in industrial and rehabilitation settings, where sustained physical activity can lead to muscle strain or fatigue. The proposed system not only optimizes force distribution between the human and robotic components but also adapts to varying task demands, ensuring ergonomic and efficient assistance.

Objectives:

i. Review on upper-limb exoskeletons, reflex-based control and fatigue detection methods;

ii. Integrate biologically inspired reflex mechanisms into an upper-limb exoskeleton;

iii. Develop coordinated control strategies to distribute physical effort effectively between the exoskeleton and the robotic arm;

iv. Implement algorithms to detect muscle fatigue using physiological signals, enabling dynamic adaptation of assistance levels;

v. Validate the system’s performance in real-world environments, assessing user comfort, fatigue reduction, and task efficiency.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos (cristina@dei.uminho.pt), Diogo Martins (id11648@alunos.uminho.pt)

Descrição:

This project aims to detect muscular fatigue using non-invasive sensors like NIRS, IMUs, or skin conductance, eliminating the need for EMG. The system focuses on continuous, comfortable monitoring across applications.

Objectives:

- Identify physiological markers of fatigue from alternative sensors

- Develop and train AI models for fatigue estimation

- Test system in controlled and applied environments

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Professora Cristina Santos (cristina.santos@dei.uminho.pt), Professor Nuno Ribeiro (nuno.fribeiro@dei.uminho.pt)

Descrição:

Este projeto visa desenvolver uma estratégia inovadora de prevenção de quedas em idosos, com foco específico na deteção e resposta a escorregamentos — a principal causa identificada de quedas acidentais em populações envelhecidas. Anualmente, são reportadas cerca de 684.000 quedas fatais e 37,3 milhões de quedas não fatais que requerem cuidados médicos, afetando maioritariamente indivíduos com mais idade. Assim, é crucial identificar precocemente os indivíduos em risco e dotá-los de ferramentas eficazes para contrariar perturbações inesperadas na marcha.

A presente proposta está estruturada em torno de dois objetivos principais interligados:

i) Desenvolvimento de uma estratégia baseada em ortóteses ativas para prevenção de quedas por escorregamento. Isto é, desenvolver uma estratégia de deteção e resposta que utilize um exoesqueleto com capacidade de aplicar torque assistivo, de forma faseada e adaptativa, para contrariar os efeitos de escorregamentos durante a marcha. A abordagem baseia-se em princípios de controlo neuromuscular, nomeadamente Central Pattern Generator (CPGs), permitindo sincronização com a marcha natural do utilizador e resposta em tempo real a perturbações externas;

ii) Exploração de algoritmos baseados em CPGs e algoritmos baseados em Inteligência Artificial para deteção e reação a escorregamento. Ou seja, estudar, simular e aplicar diferentes arquiteturas de CPGs, com o objetivo de detetar pelo menos dois tipos distintos de escorregamento e reagir eficazmente a cada um, garantindo a estabilidade postural e evitando a queda. O sistema deve ser sensível à fase da marcha, ajustando o comportamento da ortótese em conformidade.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Professora Cristina Santos (cristina.santos@dei.uminho.pt), Ricardo Andrade (id11180@uminho.pt)

Descrição:

Seniors and patients with neuromuscular diseases face limitations in performing activities of daily living (ADLs), requiring easy-to-use wearable robots to regain independence. Wearable robotic devices like orthosis and exoskeletons are increasingly seen in the scientific and healthcare community as an avenue for gait and ADL assistance. However, several factors have hindered the widespread adoption of these devices at the clinic, the workplace, and at home. Rigid frames and cuffs and torque-based actuators lead to bulky and heavy devices, kinematic incompatibility, low usability, and lower actuation efficiency. Evidence-based research proposes soft wearable exoskeletons (SWEs), particularly exosuits, as an efficient avenue for ADL assistance at patients’ homes. Compared to rigid exoskeletons, SWEs replace a rigid frame with a soft, textile-based interface and torque-based actuation with force-based actuation. As such, the new design encompasses a soft interface similar to day-to-day clothing and active fibers easily integrated into this interface for force generation. Within these, Shape Memory Alloy (SMA)-based exosuits have shown promise. These materials experience large shape changes when subjected to stimulus, thus generating high forces. Furthermore, this smart material is easily machined into wires and fibers, making it a prime candidate for textile integration. Control of smart materials like SMAs is a novel field in materials and electrical engineering, with most of the breakthrough appearing in the last five years. Due to the material unorthodox inputs (thermally-induced shape change) and highly non-linear and highly hysteretic behaviour, position/force control is an open research question with high potential for innovation. Question like effective sensorization, control architecture, cooling integration and non-linearity correction have yet to be solved. This master thesis aims to design a low-level control architecture for a SMA spring and spring bundles. The work will include the determination of the optimal type of control (position vs force-based), control hardware (including physical sensors vs self-sensing control, processing unit selection, power supply and cooling and heating) and control algorithms for effective heating and cooling, position control and safety. The integration of several SMA actuators in a bundle and in a textile matrix, following the chosen architecture, also poses a significant challenge and promising innovations.

Requerimentos: Basic understanding of control strategies and loops; Experience in STM32 or Arduino programming; MATLab programming

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Professora Cristina Santos (cristina.santos@dei.uminho.pt), Ricardo Andrade (id11180@uminho.pt)

Descrição:

Seniors and patients with neuromuscular diseases face limitations in performing activities of daily living (ADLs), requiring easy-to-use wearable robots to regain independence. Wearable robotic devices like orthosis and exoskeletons are increasingly seen in the scientific and healthcare community as an avenue for gait and ADL assistance. However, several factors have hindered the widespread adoption of these devices at the clinic, the workplace, and at home. Rigid frames and cuffs and torque-based actuators lead to bulky and heavy devices, kinematic incompatibility, low usability, and lower actuation efficiency. Evidence-based research proposes soft wearable exoskeletons (SWEs), particularly exosuits, as an efficient avenue for ADL assistance at patients’ homes. Compared to rigid exoskeletons, SWEs replace a rigid frame with a soft, textile-based interface and torque-based actuation with force-based actuation. As such, the new design encompasses a soft interface similar to day-to-day clothing and active fibers easily integrated into this interface for force generation. Within these, Shape Memory Alloy (SMA)-based exosuits have shown promise. These materials show a shape memory effect when under a stimulus that allows the design of actuators with high-power to weight ratios that are easily inserted into soft interfaces. However, the design of these actuators has been stifled by low bandwidths and non-linear behaviours. Novel research on the field of SMAs has focused on finely tuning micro and nanoscale variables like atomic composition and thermomechanical processing. Through finely tuning these variables, it is possible to tailor the SMA to fit the actuation purpose, something that has yet to been done for biomedical, gait assistance applications. The large design space and the fact that the large majority of SMAs have not yet been manufactured and validated poses a significant challenge that has yet to be solved.

This master thesis aims to fuse computational material engineering, big-data artificial intelligence-based computational methods and strong knowledge of the requirements of gait assistance applications to develop a comprehensive and effective machine-learning based model that can accurately predict SMA actuation metrics from nano and micro-scale parameters. It further encompasses the accurate conception of the design space within the large variety of possible SMAs, the determination of actuation requirements for prediction trough biomedical-informed decisions and validation of the model trough manufacture and experimental validation of novel SMAs. By integrating computational material engineering with big-data artificial intelligence methods, this thesis aims to revolutionize the design and application of SMA-based exosuits for gait assistance, ultimately supporting greater independence and quality of life for those in need.

Requerimentos: Big-data and machine learning programming languages (e.g. python), design models optimization (e.g., ANSYS), sensor-based experimental design.

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Professora Cristina Santos (cristina.santos@dei.uminho.pt), Ricardo Andrade (id11180@uminho.pt)

Descrição:

Seniors and patients with neuromuscular diseases face limitations in performing activities of daily living (ADLs), requiring easy-to-use wearable robots to regain independence. Wearable robotic devices like orthosis and exoskeletons are increasingly seen in the scientific and healthcare community as an avenue for gait and ADL assistance. However, several factors have hindered the widespread adoption of these devices at the clinic, the workplace, and at home. Rigid frames and cuffs and torque-based actuators lead to bulky and heavy devices, kinematic incompatibility, low usability, and lower actuation efficiency. Evidence-based research proposes soft wearable exoskeletons (SWEs), particularly exosuits, as an efficient avenue for ADL assistance at patients’ homes. Compared to rigid exoskeletons, SWEs replace a rigid frame with a soft, textile-based interface and torque- based actuation with force-based actuation. As such, the new design encompasses a soft interface similar to day-to-day clothing and active fibers easily integrated into this interface for force generation. Due to the high compliance of soft interfaces and the integration of active components, the design of this interface is of paramount importance both for comfort and efficient force transmission. Bio-inspired design principles—modeling textiles on the hierarchical, anisotropic behavior of muscles, tendons, and fascia—have recently emerged as a compelling strategy: by emulating tissue mechanics, a garment can transmit assistance more effectively while feeling natural to the user. Overall, a clothing-like interface with characteristics that closely resemble the musculoskeletal system of the individual can provide optimal force transmission and comfort.

This master thesis aims to join musculoskeletal system modelling, mechanical and material engineering and computer-aided design (CAD) to develop a breakthrough textile- based interface for a soft exosuit that effectively mimics the human soft tissues, optimizes mechanical properties and boosts force transmission. The work will comprise four phases:

1. Musculoskeletal and exosuit modelling in OpenSIM, simulating interface stiffness, load paths and assistive force requirements;

2. Optimization of the mechanical properties of the exosuit through optimization algorithms to maximize force transmission and minimize pressure;

3. Evaluation and selection of candidate textile materials against the target properties;

4. Translation of points 2. and 3. in a manufacturable exosuit through CAD modelling.

Requerimentos: Basic understanding of human lower-limb gait kinematics; some prior knowledge of biomechanical modelling; mechanical properties and material behaviour; CAD-based design

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos (cristina@dei.uminho.pt), Joana Almeida (joana.almeida@dei.uminho.pt)

Descrição:

This project combines wearable IMUs and EMG sensors to detect muscular fatigue and posture in real time. AI-based feedback systems will prevent injury and optimize rehabilitation and ergonomic interventions.

Objectives:

- Review fatigue and posture indicators from sensors

- Develop AI models for fatigue and posture classification

- Implement biofeedback strategies

- Validate system in laboratory and real-world settings

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos (cristina@dei.uminho.pt); Joana Almeida (pg50453@alunos.uminho.pt)

Descrição:

Gait rehabilitation for individuals with spinal cord injury (SCI) or stroke demands adaptive, personalized assistance that seamlessly integrates functional electrical stimulation (FES) and exoskeletons (EXO). Current hybrid systems suffer from critical limitations: FES often relies on fixed stimulation patterns that ignore patient-specific muscle coordination, while EXO control strategies frequently depend on electromyography (EMG), which is impractical in real-world scenarios due to signal noise, electrode placement challenges, and interference from FES artifacts. These shortcomings restrict the systems’ ability to adapt to complex, real-life activities of daily living (ADLs), ultimately limiting their clinical efficacy and patient autonomy.

This proposal aims to overcome these challenges by developing a bio-inspired control framework for hybrid FES-EXO systems, integrating methods such as Central Pattern Generators (CPGs), Dynamic Movement Primitives (DMPs), muscle synergies, and reflex-based feedback. This innovative approach will enhance gait rehabilitation through personalized FES and exoskeleton coordination. The FES component will employ user-specific muscle synergies extracted from surface EMG (sEMG) data collected during various ADLs. These synergies will drive dynamic adjustments of FES parameters, including pulse width, frequency, and activation timing, to optimize muscle recruitment. For the EXO, a novel coordination method will be developed by selecting the most suitable bio-inspired strategy to synchronize with FES. The two control strategies will be unified into a hierarchical architecture, where high-level decision-making coordinates FES and EXO actions to ensure smooth, stable motion across all ADLs.

Objectives:

i. Literature Review on bio-Inspired control methods;

ii. Collect motion data from healthy individuals to develop control algorithms based on natural movement patterns;

iii. Extraction of muscle synergies and establishment of reference trajectories for each ADL

iv. Implement and test the hybrid control framework in a simulated environment to assess its efficacy;

v. Integration of the controller on an actual hybrid FES-EXO system, optimizing for real-time performance;

vi. Experimental validation on healthy and clinical subjects

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos (cristina@dei.uminho.pt); Joana Almeida (pg50453@alunos.uminho.pt)

Descrição:

Lower-limb exoskeletons (LLEs) for spinal cord injury (SCI) and stroke rehabilitation often rely on fixed kinematic trajectories, limiting their adaptability to real-world activities of daily living (ADLs) such as walking, stair climbing, and sit-to-stand transitions. These devices are typically bulky, requiring external support like crutches, which increases upper-body exertion and cognitive load, reducing usability and therapeutic benefits. This proposal introduces a bio- inspired hybrid control framework to enhance adaptability and balance in LLEs.

The framework combines Central Pattern Generators (CPGs) with Reinforcement Learning (RL) to achieve synchronized, adaptive movements across diverse ADLs. Nonlinear oscillators, such as Matsuoka or Hopf models, will generate coordinated joint trajectories while allowing smooth transitions through phase resetting and modulation. A lightweight convolutional neural network (CNN) will predict user intent in real time using minimal sensors, emphasizing noise robustness and computational efficiency for embedded systems. Additionally, a deep RL agent will dynamically adjust CPG outputs to maintain balance, evaluated through stability metrics such as the center of mass within the base of support and orbital energy during movement. The RL policy will be trained using human-exoskeleton simulations and fine-tuned on hardware, ensuring safe and adaptive real-world performance.

Objectives:

i. Review of the state of the art on CPG control, balance control, and RL applications in exoskeletons;

ii. Acquire motion data from healthy individuals with an LLE to create a comprehensive dataset for model training and CPG reference creation;

iii. Development of a multi-DoF network with nonlinear oscillators to generate adaptive, synchronized joint trajectories for diverse ADLs;

iv. Familiarize with the team-own CNN for user intention decoding and optimization to the outlined ADLs;

v. Train a deep reinforcement learning agent to modulate CPG outputs for dynamic balance across varying ADL scenarios;

vi. Integration of the hybrid control framework into a lower limb exoskeleton

vii. Validation and assessment of balance and usability and ADL execution through experimental trials with real-world exoskeleton users.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina Santos (cristina@dei.uminho.pt), Joana Figueiredo (joana.figueiredo@dei.uminho.pt )

Descrição:

Functional electrical stimulation (FES) through robotic devices such as the neuroprosthesis (figure) enables that subject who suffer from muscle weakness, hemiparesis, or paralysis can recover their motor functions. Moreover, FES is a promising technique to augment the muscle strength of athletes. The control over electrical stimulation parameters in FES is still challenging due to the nonlinear behavior and unique characteristics of each subject’s muscles. Neuroprosthesis’ control strategies based on the combination of feedback and feedforward components have proven high reliability and performance. Current strategies use a trajectory tracking control, where the system imposes predefined joint angle trajectories.

This dissertation aims to develop an adaptive control for neuroprosthesis to adapt the stimulation according to each subject's physiological response at the minimal fatigue. For this purpose, this work will complement the available PID-based feedback controller with a machine learning-based feedforward controller that models the human muscle function to evoke different motor tasks, such as walking. Moreover, this dissertation aims to carry out clinical studies with neurologically impaired subjects to assess the FES’ motor effects during walking. Objectives:

- Conduct experiments with multiple subjects to collect data to build the machine learning model;

- Develop the feedforward controller of the neuroprosthesis with adaptability to subject's physiological response;

- Integrate the controller into the neuroprosthesis (already including a PID control);

- Test the developed control strategy with healthy subjects;

- Conduct clinical studies with neurologically impaired subjects;

- Evaluate the control response, comparing to existing control strategies;

- Write the dissertation.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina Santos (cristina@dei.uminho.pt), Joana Figueiredo (joana.figueiredo@dei.uminho.pt )

Descrição:

Stroke is the main causes of disability among adults, limiting the patients’ physical ability to walk. Robotic devices able to repetitively provide the assistance that the patient really needs (and when needs) appeared as a promising intervention. There is evidence demonstrating that leveraging assistance from exoskeleton-generated external forces and Functional Electrical Stimulation (FES)-generated internal forces is the optimal intervention for the patient regaining motor independence.

This dissertation aims at developing intelligent controls to manage the assistance of an ankle exoskeleton and FES system (i.e., to control exoskeleton’s motion and FES-driven muscle activation) according to the patients’ motor needs. Moreover, this dissertation aims to carry out clinical studies with neurologically impaired subjects to evaluate the assistance benefits of each device.

Objectives:

- Familiarize with the commercial robotic devices: exoskeleton and FES;

- Conduct experiments with multiple subjects to collect data to build the machine learning model;

- Develop artificial intelligence tool to generate reference trajectories customized to each user (height, body mass, age, motion speed, motion intention);

- Optimize the available impedance control to drive the ankle exoskeleton’s motion according to the generated trajectories;

- Optimize the feedforward controller of the FES with adaptability to subject's physiological response;

- Test the enhanced control strategies with healthy subjects;

- Conduct clinical studies with neurologically impaired subjects;

- Evaluate the control response, comparing to existing control strategies;

- Write the dissertation.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Professora Cristina P. Santos (cristina@dei.uminho.pt ), Joana Figueiredo (joana.figueiredo@dei.uminho.pt )

Descrição:

Stroke is the main causes of disability among adults, limiting the patients’ physical ability to walk. Robotic devices able to repetitively provide the assistance that the patient really needs (and when needs) appeared as a promising intervention. There is evidence demonstrating that leveraging assistance from exoskeleton-generated external forces and Functional Electrical Stimulation (FES)-generated internal forces is the optimal intervention for the patient regaining motor independence.

This dissertation aims at developing an intelligent control to manage the assistance of an ankle exoskeleton and FES system (i.e., to control exoskeleton’s motion and FES-driven muscle activation) according to the patients’ motor needs. Moreover, this dissertation aims to carry out clinical studies with neurologically impaired subjects to evaluate the assistance benefits of each device.

Objectives:

- Familiarize with the commercial robotic devices: exoskeleton and FES;

- Develop artificial intelligence tool to generate reference trajectories customized to each user (height, body mass, age, motion speed, motion intention);

- Optimize the available impedance control to drive the ankle exoskeleton’s motion according to the generated trajectories;

- Optimize the muscle synergy-based control to enable smooth muscle activation using FES;

- Test the enhanced control strategies with healthy subjects;

- Conduct clinical studies with neurologically impaired subjects;

- Write the dissertation.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina Santos (cristina@dei.uminho.pt), Joana Figueiredo (joana.figueiredo@dei.uminho.pt)

Descrição:

The application of smart wearables for tracking human motion and activity in daily life is rising in several applications such as assessing the gait-related injuries, rehabilitation success, and performance of workers and athletes. In particular, the use of SmartShoe (shoe instrumented with compact sensors) is rising for monitoring the gait and balance performance in several scenarios (indoor and outdoor stairs, ramps). This dissertation aims to achieve a SmartShoe as a robust device for repeatable use in daily motion analysis. It aims the development of deep learning methods to estimate ground reaction force, gait parameters (speed, number of steps, step length), and balance parameters (center of pressure, base of support) from the plantar pressure and foot acceleration data. This prototype has been developed in the scope of a national project in collaboration with an industrial consortium.

Objectives:

- Collect multiple data points with the SmartShoe with healthy and injured subjects;

- Develop robust algorithms to monitor the plantar pressure colormap and balance parameters;

- Develop deep neural networks to estimate gait parameters and ground reaction force;

- Compare the accuracy of the developed algorithms with gold-standard tracking systems;

- Design and implement feedback strategies to visualize the parameters estimated by the SmartShoe in real-time;

- Dissertation writing.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos and PhD student Vitor Martins

Descrição:

Humanoid robots operating in hybrid industrial environments must autonomously reason, plan and adapt to complex tasks and unforeseen conditions. Recent advances in Large Language Models (LLMs) provide a promising pathway for enabling semantic task planning based on high-level objectives and environmental context, incorporating object recognition and localization, as well as human-state and activity. This thesis proposes the development of a high-level planning system for humanoid robots powered by LLMs. The system will translate natural language goals or environmental descriptions into executable action plans, leveraging LLMs; capability for reasoning, task decomposition, and flexible adaptation. Plan execution will be validated in simulation by connecting the planner to a basic motion control library, enabling rapid testing without requiring detailed low-level motion implementations. Ground-truth perception will be used to provide accurate environment and task information. System performance will be evaluated based on planning correctness, execution feasibility, adherence to workspace constraints, and compliance with safety standards.

Objectives:

1. Survey of the state-of-the-art in LLMs and their role in robotic reasoning and planning.

2. Design of a planning system leveraging an LLM to generate task-oriented action sequences.

3. Implementation and integration with a high-level simulation control interface (e.g., Omniverse Isaac Sim motion APIs).

4. Evaluation of the system in simulated industrial tasks based on metrics such as task success rate, plan consistency, and adherence to safety constraints.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos and PhD student Vitor Martins

Descrição:

In dynamic and unstructured environments, humanoid robots must go beyond basic reactive control and incorporate advanced perception to interact effectively with humans and their surroundings. Vision-Language Models (VLMs) offer a cutting-edge solution by aligning visual data with semantic, language-based understanding. The main goal of this dissertation is the development of a context-aware perception module for humanoid robots, utilizing VLMs to enable the interpretation of complex, real-world scenes. The module will support essential capabilities such as object detection and localization, task recognition, and the inference of human states and activities. By integrating structured, high-level semantic information into the robot’s decision-making and action-planning pipeline, the system will facilitate safer, more adaptive, and more intelligent interactions. Through multimodal perception, the robot will be able to identify and prioritize objects for manipulation based on the task, as well as infer human intent through activity recognition. The developed system will be validated in simulated scenarios requiring strong contextual awareness, focusing on challenges such as object affordance recognition, human-state estimation, and dynamic task adaptation.

Objectives:

1. Survey of the state-of-the-art in VLMs and their role in robotic perception.

2. Design of a perception system integrating VLMs with the humanoid robot’s sensor suite.

3. Implementation and training of the VLM-based perception module in simulation.

4. Evaluation of the model in scenarios requiring contextual perception, such as task recognition, object identification and localization, and human-state identification.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos and PhD student Vitor Martins

Descrição:

Humanoid robotics have a tremendous potential for industrial tasks that are too demanding on the Human. These robots aim to emulate the wide range of human movements and behaviors suited for human environments. However, despite notable progress, a critical gap persists between human agility and robotic dexterity, which needs to be addressed. The main goal of this dissertation is the development of a motion control strategy that grants robustness to humanoid movement, allowing it to execute complex tasks such as the industrial ones. Methods such as reinforcement learning and Dynamic movement primitives can be explored.

Objectives:

1. Review of the state of art regarding humanoid robots and control strategies.

2. Conceptualization of the control strategy;

3. Implementation and training of the control algorithms;

4. Test and validation of the implemented algorithms.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos (cristina@dei.uminho.pt) and PhD students Sara Cerqueira and Diogo Martins (diogo-martins-9@live.com.pt)

Descrição:

Work-related Musculoskeletal Disorders (WRMSDs) represent 53% of occupational diseases and have an economic impact of €240 billion on Europe. Poor posture and muscle fatigue are two of the main causes of these disorders. The team has been working on reducing the incidence of this causes. At the moment, it already has both an ergonomics-based and fatigue-based control strategies implemented for a UR10e robot. The main goal of this proposal is to develop an impedance controller that integrates, manages the importance, and balances the action of both these controllers, while considering the workers’ productivity. It should receive fatigue, postural ergonomics and productivity scores to model the impedance controller parameters. Meaning that, when the worker is not fatigued, good posture (ergonomics) should be prioritize. When fatigue starts to be felt by the workers, then this controller should take more action than the ergonomics one. To tune the controller reinforcement learning algorithms can be used.

Objectives:

1. Review of the state of art regarding impedance controllers in Human-Robot Collaboration (HRC);

2. Familiarization with team’s HRC framework, namely the UR10e robot and controllers;

3. Design and implementation of the control strategy;

4. Development of the RL environment to train and tune the impedance controller.

5. Test and validation of the implemented control strategy.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos (cristina@dei.uminho.pt) and PhD students Sara Cerqueira and Diogo Martins (diogo-martins-9@live.com.pt)

Descrição:

Work-related Musculoskeletal Disorders (WRMSDs) represent 53% of occupational diseases and have an economic impact of €240 billion on Europe. Muscle fatigue is a significant cause of these disorders. It is a biochemical process that changes muscles’ characteristics and limits worker’s performance, by reducing muscle power/force and inducing discomfort, and, when the effort is high and recurrent, can even cause pain and injuries, which raises the need to study and predict muscle fatigue in a timely manner. Automated fatigue estimation models, driven by the gold-standard surface electromyography (sEMG) data, have been widely used. However, sEMG sensors require direct contact with skin, and their rigorous positioning is challenging and requires anatomic knowledge. Less intrusive methods are still required to monitor upper-body muscle fatigue in industrial context.

The main goal of this proposal is to improve the team’s muscle fatigue estimation model based on sEMG data and Artificial Intelligence (AI). Other muscle fatigue metrics will be pursued. Sensor fusion will be exploited towards alternative methods of assessing muscle fatigue, including kinematics data and other biosignals, such as electrocardiogram (ECG), galvanic skin response (GSR) or VO2. Therefore, this project aims to find the biosignals that correlate the most with muscle fatigue and explore and develop time-efficient AI-based models.

Objectives:

1. Review of the state of art on muscle activity and fatigue assessment, AI models, and identification of the main challenges;

2. Familiarization with team’s sensors, such as Xsens MTw Awinda (kinematics), EMG, ECG, GSR and VO2 sensors, and their signals;

3. Protocol design and experimental sensory data collection for the development of AI algorithms;

4. Exploration of different biosignals for estimating muscle fatigue;

5. Train and test of AI algorithms;

6. Evaluation, selection, and tuning of the best model;

7. Integration of the fatigue estimation model in the team’s framework.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos (cristina@dei.uminho.pt) and PhD students Sara Cerqueira and Diogo Martins (diogo-martins-9@live.com.pt)

Descrição:

Work-related Musculoskeletal Disorders (WRMSDs) represent 53% of occupational diseases and have an economic impact of €240 billion on Europe. They are responsible for 30% of the years lived with disability, causing chronic pain to 60% of the European workers. After an injury, workers’ pain reduces their productivity. Furthermore, serious injuries hinder workers from returning to work, since industrial workstations are not suitable for workers with musculoskeletal disorders. These issues could be addressed by assessing worker’s pain continuously. To replace subjective questionnaires and scales, automated pain detection models have been explored recently, combining Machine Learning with physiological signals, such as electrocardiogram (ECG), galvanic skin response (GSR) or electromyography (EMG) sensors. However, the performance of these models is still limited, and they behave like black boxes, which lowers the human trust on them. Further, industrial applications require an unobtrusive sensor setup, which implies avoiding sensory and signal redundancy.

The main goal of this proposal is to develop an automated and reliable pain detection model based on Deep Learning. To choose only the essential biosignals, dimensionality reduction methods and Explainable Artificial Intelligence (XAI) algorithms can be used. Therefore, this project aims at the development of time-efficient AI-based pain detection models driven by physiological signals.

Objectives:

1. Review of the state of art regarding musculoskeletal pain, Deep Learning models for pain detection, and identification of the main challenges;

2. Familiarization with team’s sensors, such as Xsens MTw Awinda (kinematics), EMG, ECG and GSR sensors, and their signals;

3. Protocol design and experimental sensory data collection for the development of AI algorithms;

4. Train and test of Deep Learning algorithms;

5. Evaluation, selection, and tuning of the best model;

6. Exploration of different XAI algorithms;

7. Integration of the pain detection model in the team’s framework.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos (cristina@dei.uminho.pt) and PhD students Sara Cerqueira and Diogo Martins (diogo-martins-9@live.com.pt)

Descrição:

The upcoming Industry 5.0 focus on mass product customization. Therefore, it demands flexible, agile, and quickly changeable workstations that can adapt to the production changes. However, if a process is continuously changing, different expertise is required from the human workers. As a consequence, this can impact production quality and productivity, since inexperienced workers need time to learn and adapt to different processes, especially if they are complex. Robotics can solve this problematic. Learning from demonstration is a research field inside Human-Robot Collaboration (HRC) where robots learn how to master a skill from human demonstrations. After mastering this skill, a robot can act as a teacher, and, without getting tired, can teach a new worker how to perform the given task, by working with him, and correcting him whenever necessary.

The main goal of this proposal is the development of a task-quality management control strategy for Human-Robot Collaboration that consists of two phases: a phase where the robot learns with a human expert how to perform the task; and a second phase, where the previous algorithm is placed inside the robot’s control loop so that the robot helps the worker to perform the task, teaching and correcting him during its execution, only when is required (assist-as- needed).

Objectives:

1. Review of the state of art regarding learning from demonstration algorithms (e.g. dynamic movement primitives) and assist-as-needed control strategies;

2. Familiarization with team’s HRC framework, namely the UR10e robot;

3. Design and development of the learning from demonstration algorithms;

4. Design and development of the assist-as-needed control strategy;

5. Integration of the control strategy in the team’s framework.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos (cristina@dei.uminho.pt) and PhD students Sara Cerqueira and Diogo Martins (diogo-martins-9@live.com.pt)

Descrição:

Work-related Musculoskeletal Disorders (WRMSDs) represent 53% of occupational diseases and have an economic impact of €240 billion on Europe. These injuries are multifactorial and are influenced by the worker’s mental well-being. Factors such as psychological distress, high demands at work and reduced decision-making authority increase WRMSD’s risk. In the scope of Industry 5.0, collaborative robots are being progressively adopted in repetitive tasks to ease physical labor. However, close physical collaboration can augment worker’s cognitive workload/stress, potentially degrading performance, which is why collaborative robots are still mostly limited to co-existence with humans. Hence, considering the cognitive workload perceived by the worker needs to be addressed.

The main goal of this proposal is to develop an impedance controller that balances worker’s perceived cognitive workload and productivity. This work includes exploring physiological signals to find the most suitable indicators for the cognitive workload assessment. This, together with a productivity score, will be the controller’s inputs, to model its impedance parameters, using Reinforcement Learning or other optimization techniques. Therewith, when the worker is not experiencing a significant cognitive workload, productivity should be prioritized; and, when cognitive overload starts to be felt by the worker, the robot motion should be tailored to mitigate it.

Objectives:

1. Review of the state of art regarding cognitive workload/stress assessment and impedance controllers in Human-Robot Collaboration (HRC);

2. Definition of cognitive workload indicators based on physiological signals;

3. Familiarization with wearable sensory systems (electrocardiogram, galvanic skin response, etc) and team’s HRC framework, namely the UR10e robot and controllers;

4. Design and implementation of the control strategy;

5. Test and validation of the implemented control strategy.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos (cristina@dei.uminho.pt) and PhD students Sara Cerqueira and Diogo Martins (diogo-martins-9@live.com.pt)

Descrição:

The recuperation from musculoskeletal disorders/injuries is heavily reliant on patient involvement in a physiotherapy program. Transition towards telerehabilitation is growing expressively, with digital musculoskeletal care and physiotherapy companies getting a lot of attention of investors. Technology and digital physiotherapists may not replace in- person care, but they can complement it. Monitoring home exercise compliance and performance quality are important factors to ensure a successful rehabilitation program, and its evaluation relying on self-reports or video recordings is subjective, inaccurate, and performed afterwards, not continuously. This highlights the need for a continuous, objective, data-driven method capable of tracking patients’ outcomes and progress in home rehabilitation programs. Real-time guidance while executing rehabilitation exercises is still lacking and wearable motion sensors can be the answer. Moreover, the concept of gamification is becoming very relevant in healthcare to enhance users’ motivation and engagement, showing potential in rehabilitation therapy, where it helps to reduce the associated tedium, but more investigation is necessary.

The main goal of this proposal is the development of a digital physiotherapy tool for tracking the recovery of injured people, through a remote and personalized physiotherapy program that guides the users in the execution of the prescribed rehabilitation exercises. Different biofeedback types (vibrotactile, visual or audio) for patients’ real-time awareness will be explored. Users’ engagement strategies will also be considered to ensure the patients’ adherence.

Objectives:

1. Review of the state of art regarding digital physiotherapy and gamification strategies;

2. Delineation of the physiotherapy exercises and patient’s performance metrics (in collaboration with the physiotherapist Tiago Pinhão);

3. Development and integration of the software and hardware (sensors (commercial) and feedback devices, such as vibrotactile motors or graphical user interface) modules;

4. Evaluation of the tool’s reliability, usability and impact on patient’s recovery.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos (cristina@dei.uminho.pt) and PhD students Sara Cerqueira and Diogo Martins (diogo-martins-9@live.com.pt)

Descrição:

Work-related Musculoskeletal Disorders (WRMSDs) represent 53% of occupational diseases and have an economic impact of €240 billion on Europe. Bad postures are one of their main causes. To reduce WRMSDs’ risk, it should be evaluated through postural assessments. However, these are commonly based on ergonomists’ subjective snapshot observations and are time-consuming. More objective tools are required to expedite assessments. Furthermore, current automated approaches typically rely on joint angles, whose estimation require complex algorithms (e.g., biomechanical models), and provide discrete scores. Artificial Intelligence (AI) may be the key to solve this issue, by allowing a simpler approach that uses raw kinematic data directly to estimate a continuous postural risk. Additionally, industrial applications require an unobtrusive sensor setup, which implies avoiding sensory redundancy.

The main goal of this dissertation is to develop an ergonomic assessment tool. It includes developing differentiable Deep Learning regression models for estimating the ergonomic risk based on kinematic data (acceleration and angular velocity), provided by wearable sensors. To choose only the essential signals, dimensionality reduction methods and Explainable Artificial Intelligence (XAI) algorithms can be used. Further, the presentation of the risk assessment should be intuitive, in an interface, to allow workers and ergonomists to perceive which body segments are aggravating the risk.

Objectives:

1. Review of the state of art on ergonomic assessment methods and AI/DL models;

2. Selection of the most important sensors and signals;

3. Train and test of DL algorithms, using team’s datasets;

4. Evaluation, selection, and tuning of the best model;

5. Exploration of different XAI algorithms;

6. Integration of the AI model in an app, with an interface to present intuitively the ergonomic risk of the body segments and overall.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:André W. Catarino, Cristina P. Santos and PhD Student Gilberto Martins

Descrição:

With rapid technological advances, the development and utilization of wearable technologies have experienced exponential growth. This market has already proven its worth,with an estimated global value of USD 32.63 billion in 2019 and expected annual growth of 15.9% until 2027. Wearable technology offers a wide range of applications, such as tracking user movement using accelerometers and gyroscopes. Thus, these technologies can become valuable tools for the prevention of work-related musculoskeletal disorders, through ergonomic assessment of the user, and for monitoring performance during physical activity or even during the rehabilitation process. The main goal of this proposal is the development of a wearable garment, prepared for sensorization, exploring the use of E-Textiles for the integration of surface electrodes and conductive tracks. The garment should be developed focusing on user comfort and in order to maximize the contact of the sensors with the skin surface, by integrating compression zones.

Objectives:

1. Literature review on wearable devices and E-Textiles used in biomedical applications, focusing on material properties such as breathability, hypoallergenic performance, heat and sweat dissipation.

2. Analysis of wearable devices currently used for ergonomic assessment, physical performance assessment and rehabilitation, listing materials used, production techniques and techniques for sensor integration.

3. Gathering of mechanical and physiological requirements, identifying the regions of interest where the sensors will be located and where the areas of highest compression should be positioned.

4. Evaluation of textile materials for integration into the device in terms of flexibility, permeability and skin compatibility (sample testing).

5. Proposal of a first layout of the garment, highlighting areas of greater compression and considering aspects such as comfort during long periods of use and ease of putting don/doffing the garment.

6. Development of the first prototype, through 2D layout and 3D modeling, integrating areas for sensor placement and textile surface electrodes.

7. Development of the test protocol. Evaluation of parameters such as user comfort, sweat dissipation, thermal permeability and breathability of the garment.

8. Validation of the prototype with usability tests on real users while performing physical activities and work-related tasks.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos and PhD Student Gilberto Martins

Descrição:

The World Health Organization predicts that by 2050, the proportion of people aged 60 and up will increase from 12% to 22%. This emphasizes the importance of promoting healthy aging strategies, such as ensuring optimal occupational health and encouraging positive behaviors like regular exercise to help the elderly maintain autonomy. One important aspect of occupational health is addressing Work-Related Musculoskeletal Disorders (WRMSDs), which account for 53% of all occupational diseases and affect 22% of the European population, resulting in an economic strain of €240 billion. Current ergonomic risk assessment relies on time-consuming observational methods centered on the ergonomist's expertise.   Lately, researchers have proposed systems that combine the use of Inertial Measurement Units (IMUs) for postural monitoring and the acquisition of surface electromyography (sEMG), thus combining postural assessment with the evaluation of muscle activity and fatigue. However, electromagnetic interferences can compromise the quality of the electromyographic signal and, consequently, the muscle monitoring of the user.

This proposal focuses on exploring the use of Functional Near Infrared Spectroscopy (FNIRs) to assess muscle activity and fatigue through variations in muscle blood oxygenation. This technology has the potential to work as a complement to the acquisition of the electromyographic signal or even as an alternative in environments where external interferences can compromise the quality of the EMG signal. Objectives:

1. Literature review on the application of fNIR in the detection and assessment of muscle oxygenation, hemodynamics, activity and fatigue.

2. Identification of fNIR systems and their key features (optode configuration, wavelengths and signal processing methodologies).

3. Requirements gathering for the application, including acquisition location (muscles to be monitored), expected changes in the signal during physical activity, light penetration distance.

4. Development of the hardware architecture (light sources, photodetectors, signal amplification and filtering circuits).

5. Development of software for system control (control of sampling frequencies, selection of acquisition channels, signal visualization) and signal processing to extract characteristics such as oxyhemoglobin and deoxyhemoglobin concentrations and their correlation with muscle activation and fatigue.

6. Development of the prototype taking into account aspects such as ergonomics, comfort during use and the contact between the optodes and the skin surface.

7. Validation of the prototype with regards to muscle activity detection in static and dynamic conditions, fatigue and comparison with conventional systems such as EMG.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos and PhD student Gilberto Martins

Descrição:

The World Health Organization predicts that by 2050, the proportion of people aged 60 and up will increase from 12% to 22%. This emphasizes the importance of promoting healthy aging strategies, such as ensuring optimal occupational health and encouraging positive behaviors like regular exercise to help the elderly maintain autonomy. One important aspect of occupational health is addressing Work-Related Musculoskeletal Disorders (WRMSDs), which account for 53% of all occupational diseases and affect 22% of the European population, resulting in an economic strain of €240 billion. Current ergonomic risk assessment relies on time-consuming observational methods centered on the ergonomist's expertise. Lately, researchers have proposed systems that combine the use of Inertial Measurement Units (IMUs) for postural monitoring and the acquisition of surface electromyography (sEMG), the reference for muscle activity and fatigue monitoring. The main objective of this proposal is to develop a sEMG acquisition circuit, with the potential for miniaturization, which allows the acquisition of the electromyographic signal using textile surface electrodes embedded in the textile fabric. The circuit developed should make it possible to adjust the gain of the acquired signal, since the aim is to acquire the electromyographic signal from different muscle groups, whose intensities can vary significantly.

Objectives:

1. Literature review on sEMG acquisition circuits (amplification, filtering, adjustable gain).

2. Familiarization with the Delsys Trigno Avanti gold-standard EMG acquisition system, to be used during system validation.

3. Definition of the circuit specifications, namely frequency range, and recommended gain for the signal acquired for each of the muscle groups evaluated.

4. Requirements gathering regarding the system's functionality and list of necessary components for system development.

5. Designing user-friendly interfaces for gain adjustment to accommodate varying muscle intensities.

6. Test protocol definition and sensory data collection.

7. System validation while performing work-related tasks.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos and PhD student Gilberto Martins

Descrição:

will rise from 12% to 22%. This emphasizes the importance of promoting healthy aging strategies, such as ensuring optimal occupational health and encouraging positive behaviors like regular exercise, in order to help the elderly maintain their autonomy. One critical aspect of occupational health is addressing Work-Related Musculoskeletal Disorders (WRMSDs), which account for 53% of all occupational diseases and affect 22% of the European population, costing €240 billion. Current ergonomic risk assessment is based on time-consuming observational methods that focus on the ergonomist's expertise. Promoting physical activity is critical for healthy aging. However, seniors' engagement in exercise is low. Electrical muscle stimulation (EMS) is increasingly popular in the fitness community. It encourages user engagement and may be a viable alternative to voluntary exercise in seniors. The main objective of this proposal will be to explore the applicability of an EMS system using surface electrodes embedded in textiles, through the development of an actuation circuit and the study of the use of textile surface electrodes typically used for sEMG signal acquisition.

Objectives:

1. Literature review regarding the use of surface electrodes in the context of EMS and possible actuation systems for wearable application

2. Definition of the circuit specifications, namely output current and voltage, frequency, pulse width and duration.

3. Requirements gathering regarding the system's functionality and list of necessary components for system development.

4. Designing user-friendly interfaces that allow the control of intensity, frequency and duration.

5. Test protocol definition.

6. System validation with healthy subjects in the context of physical activity.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Cristina P. Santos and PhD student Gilberto Martins

Descrição:

The World Health Organization predicts that by 2050, the proportion of people aged 60 and up will increase from 12% to 22%. This underscores the importance of promoting strategies for healthy aging, such as ensuring optimal occupational health and encouraging positive behaviors like regular exercise to help the elderly maintain autonomy. A crucial aspect of occupational health is addressing Work-Related Musculoskeletal Disorders (WRMSDs), which account for 53% of all occupational diseases and affect 22% of the European population, resulting in an economic burden of €240 billion. Current ergonomic risk assessments rely on time-consuming observational methods centered on the ergonomist's expertise. Recently, researchers have proposed systems that combine Inertial Measurement Units (IMUs) for postural monitoring with the acquisition of surface electromyography (sEMG), the gold standard for muscle activity and fatigue monitoring. The primary objective of this proposal is to develop a smartphone app designed to control a wearable device and its various components, including sEMG acquisition modules, IMUs, and Electrical Muscle Stimulation (EMS) capabilities. To ensure seamless control of the device, the app will feature an intuitive, user-friendly interface. This interface will enable users to easily display real-time data, calibrate the modules, and manage settings for optimal performance.

Objectives:

1. Literature review on existing smartphone apps used in biomedical contexts, focusing on user interface design, data display, and control functionalities.

2. Study current market-leading apps that control wearable devices, such as those for fitness trackers or medical devices, to understand best practices and common features.

3. Outlining of the necessary specifications for the app, including data visualization requirements, control functionalities, and compatibility with various smartphone operating systems.

4. Requirements gathering regarding the app’s capabilities, such as real-time data display, module calibration, EMS control, and user feedback mechanisms.

5. Design of a user-friendly interfaces that facilitate easy control of the device’s sEMG, IMU, and EMS functionalities, ensuring that users can adjust settings effortlessly.

6. Definition of evaluation protocols to test the app’s usability, including data interpretation, and control responsiveness.

7. App validation through user testing

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A

Tipo: Extracurricular Internship or MSc. Thesis

Local: Universidade do Minho, Braga

Supervisores:Professora Cristina Santos (cristina.santos@dei.uminho.pt)

Descrição:

Motor learning in human-robot interaction is strongly influenced by the structure and relevance of task variability. Recent evidence suggests that task-relevant variability—that is, variability that affects task success—promotes neural adaptation by inducing motor errors that the central nervous system (CNS) can exploit to update internal models [Todorov, 2004; Wu et al., 2014]. Conversely, task-irrelevant variability is often ignored [Krakauer & Mazzoni, 2011].

In limb rehabilitation and assistive robotics, vibrotactile feedback has been explored to improve user engagement and performance. However, non-directional or ambiguous feedback may impose an implicit neural resource allocation problem, where the CNS must allocate attention or processing to decode additional sensory channels without clear task benefit [Sigrist et al., 2013; Mahani et al., 2021]. To mitigate this, the proposed work explores the use of feedback mechanisms that guide movement toward biomechanical and ergonomic optima.

Objectives:

- To design adaptive feedback systems—both vibrotactile and mechanical—that guide posture based on effort and ergonomics.

- To evaluate the influence of task-relevant variability and feedback directionality on postural learning.

- To assess whether feedback combining assistive and penalizing actions (push-pull strategy) accelerates learning.

Requerimentos:

Prazo da Candidatura: 31/08/2025

Modelo de Trabalho: N/A

Compensações: N/A