Education

In this project, we will develop an ADL monitoring system for smart wheelchairs using wearable sensors (i.e., camera, inertial measurement units). The core idea is to classify ADL by leveraging different sensing and feature computation modalities to extract relevant and complementary context bits. For example, the presence of and the interaction with distinctive objects within the camera scene provides strong cue about the ongoing user activity [1, 2, 3]. From another perspective, inertial sensors record distinct orientation and acceleration signatures of ADL-related temporal patterns [4]. Moreover, previous literature provides strong evidence that the combination of video cameras and inertial sensors improves the recognition performance compared to one or the other [5, 6]. Based on the general multi-modal activity recognition paradigm, we are seeking a novel solution tailored to the smart wheelchair. The outcome of the study will be key to analyse the expected cardiovascular function at different moments of the day in free living.

1. Bensland, S., Paul, A., Grossmann, L., Eriks-Hogland, I., Riener, R., & Paez-Granados, D. (2023). Healthcare Monitoring for SCI individuals: Learning Activities of Daily Living through a SlowFast Network. IEEE International Conference on System Integration.

2. M. Wang, C. Luo, B. Ni, J. Yuan, J. Wang and S. Yan, "First-Person Daily Activity Recognition With Manipulated Object Proposals and Non-Linear Feature Fusion," in IEEE Transactions on Circuits and Systems for Video Technology, vol. 28, no. 10, pp. 2946-2955, Oct. 2018

3. G. Schiboni, F. Wasner and O. Amft, "A Privacy-Preserving Wearable Camera Setup for Dietary Event Spotting in Free-Living," 2018 IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops), 2018, pp. 872-877

4. Lopez-Nava IH, Muñoz-Meléndez A. Human action recognition based on low- and high-level data from wearable inertial sensors. International Journal of Distributed Sensor Networks. 2019.

5. C. Chen, R. Jafari and N. Kehtarnavaz, "UTD-MHAD: A multimodal dataset for human action recognition utilizing a depth camera and a wearable inertial sensor," 2015 IEEE International Conference on Image Processing (ICIP), 2015, pp. 168-172

6. S. K. Yadav, K. Tiwari, H. M. Pandey, S. A. Akbar, “A review of multimodal human activity recognition with special emphasis on classification, applications, challenges and future directions”, Knowledge-Based Systems, 2021, Volume 223

• Review the previously trained network outcomes for activity classification and re-evaluate at higher time resolution.
• Describe the state-of-the-art of sensor-based activity monitoring systems with a focus on multi-modal activity recognition techniques, i.e., combining object detection, pose estimation, and wearable inertial sensor features.
• Define a multi-sensor configuration according to a definition of (sub)optimality with the goal to achieve a robust and generalisable recognition system.
• Define a taxonomy of activity classes and patterns to model wheelchair user behaviour in free living.
• Design and implement a multi-stage machine learning framework.
• When required, implement hyperparameter optimisation methods.
• Implement a validation framework (e.g., nested k-fold cross-validation strategy) to provide unbiased evaluation of the system’s predictive performance.
• Write a comprehensive report on the project outcomes.

• Gain unique access and first-hand experience in one of the leading institutions on long-term health management - At the Swiss Paraplegic Center at Nottwil.
• Get introduced to state-of-the-art machine learning techniques and contribute to their application in health management
• Learn about intelligent health systems, modelling for human conditions and apply regression and classification models to available data

• Enrolled student at ETH Zurich or EPFL (or another European University):
• ETHZ: D-MAVT, D-INFK / EPFL: IMT, CS (or equivalent)
• Structured and reliable working style
• Strong programming skills in Python/Matlab
• Deep learning experience on video data
• Ability to work independently on a challenging topic
  • Strong knowledge of bash, python and data structures
  • Knowledge of virtual environments (conda / docker)

Host: Dr. Diego Paez (SCAI Lab, ETHZ, SPF)

Please send your CV and the latest transcript of records from my studies to Mehdi Ejtehadi (mehdi.ejtehadi@hest.ethz.ch)

Central to the project is the integration of technological innovation with clinical validation. Beginning with a comprehensive review of existing literature and clinical proof of concept, the study establishes a foundation for understanding the current state-of-the-art in OA management. Through iterative prototyping and biomechanical testing, the knee orthosis is refined to optimize functionality and comfort while ensuring efficacy in real-world scenarios. Leveraging advancements in materials science and biomechanics, the orthosis is customized to address the unique biomechanical demands of OA patients.

The envisioned solution emphasises a cost-effective product that aligns with the broader goal of improving patient outcomes while promoting sustainability in healthcare delivery.

By bridging the gap between research and clinical practice, this thesis contributes to the advancement of orthotic technology and underscores the potential for transformative change in OA treatment.

Key Requirement: 1. Strong background in CAD design 2. Biomechanics understanding 3. Strong motivation for product development 4. Ability to communicate with experts in different fields (preferably in German)

Collaborating with a multidisciplinary team of medical engineers and health science professionals, the thesis aims to design a Swiss-customized knee orthosis tailored to individual patient needs. By focusing on both the stand and swing phases of gait, the orthosis is engineered to relieve pain and improve mobility, thereby potentially reducing the need for surgical interventions.

Host: Dr. Diego Paez (SCAI-Lab, ETHZ | SPZ)

Please send your CV and latest transcript to: Dr. Diego Paez-Granados (diego.paez@hest.ethz.ch)

We are looking for a student assistant to support a project that standardizes signal processing methods in robotics, including data management, data engineering, and software stack design.

Your profile

• Student of a Swiss university or university of applied sciences (ETH Zurich, EPFL)

• European National (or valid work permit in CH)

• Availability during the semester: ~15 hours/week - (to be discussed)

• Further availability during semester break: ~3+ days/week, to be discussed

• Mobile robots and sensors experience

• Knowledge of bash, python, and data management principles

• Strong knowledge of ROS

• Simulation experience in IsaacSIM (preferred)

• Experience in databases (preferred) - SQL
• ML and DL experience (preferred)
• Experience with Robotics Operating System (ROS) - (desirable)

We offer - Learn and investigate Machine Learning methods for healthcare. - Develop new data-driven models of the body and diseases. - Build personalized devices and sensor systems. - Develop algorithms to aid the decision-making process in therapeutic and clinical settings. - Experience industry-level data processing and management systems. - Salary: As defined by ETH regulations for student research assistants. - Flexible working hours, possibility for part-time home office. - Work with a team of interdisciplinary health professionals at the Nottwil Campus (Luzern).

Create algorithms, tools, and visualization of healthcare biosignal data, applied in ROS for robotic care. Implementing within different projects on motion analysis, rehabilitation & embodied sensing from wearables.

Analysing, data processing (40%) visualisation and graphical mapping (20%), validation, testing, and benchmarking ML algorithms (30%) and writing documentation (10%).

Starting Date: February 2025 (or shortly thereafter)

Applications: Submit your CV, transcripts and one page of fitting experience to the position. oriella.gnarra@hest.ethz.ch

Questions regarding the position should be directed to Mehdi Ejtehadi (mehdi.ejtehadi@hest.ethz.ch (no applications).

We provide the student with a large Fitbit dataset sampled at a 1-minute resolution. Initially, the student must develop suitable algorithms to extract powerful features from the multi-modal data. The student may hand-craft the features and/or leverage state-of-the-art techniques of deep representation learning. These features are then used as input to a model that detects napping episodes. The student is encouraged to start with simple threshold-based algorithms and advance to AI-based approaches. The results from the novel algorithm are then compared against methods from clinical routine. The comparison must include agreement analyses (e.g., Bland-Altman) and performance metrics (e.g., sensitivity and specificity). Eventually, the detected naps are tested as potential biomarkers in the central disorder of hypersomnolence using clustering techniques and statistical tests.

The student will develop a nap detection algorithm based on artificial intelligence in this project. The student will then assess the predictive power of the novel algorithm against established methods. Finally, the naps are tested as a potential biomarker in the central disorder of hypersomnolence.

oriella.gnarra@hest.ethz.ch

The main goal of this project is to analyze data acquired from a network of wearable and nearable devices for the automatic classification of ADL (Activities of Daily Living). The second aim is to provide a score representing the subjects' ability to perform those activities and correlate with the ICF (International Classification of Functioning, Disability, and Health) scores provided by the clinics. The third endpoint is to find correlations between the day and night sensor data.

1. Data Exploration and Analysis: You will explore a multimodal dataset collected with patients with neurological disorders and healthy individuals. Mixed-type variables include demographic features, health conditions observed during consecutive days, and multivariate time series extracted from wearable and nearables sensors.
2. Methodology Development and Deployment: You will help define what data to use, design the experiments, and develop and validate both the modeling foundation and implementation pipeline for the automatic classification of ADL, correlation with ICF scoring and correlation among day and night sensors data.
3. Presentation and Documentation: You will prepare a written report highlighting the background and the state of the art together with the implemented novelty of your work. Methodology, results, and discussion will be reported, and the code will be documented.

oriella.gnarra@hest.ethz.ch

This project offers a unique opportunity for students with exceptional data analysis skills to contribute to impactful research in healthcare innovation. Your work will focus on leveraging the StrongAge dataset to address the following research questions: 1. How is REM sleep associated with cognition and deep sleep with physical activity? 2. What are the correlations between sleep patterns and metrics such as cognition (MoCa scale), depression (validated questionnaires), frailty, in-apartment activity, and door exits? 3. How do longitudinal sleep patterns evolve in the cohort over one year? 4. What is the relationship between sleep and environmental conditions such as seasonal changes?

Literature Review: • Conduct an in-depth review of the state-of-the-art literature on the StrongAge Cohort Dataset. • Identify gaps and build a novel foundation for this research project. Data Exploration: • Assess the quality and completeness of time-series data (sensor signals). • Examine the quality and completeness of clinical data. Data Analysis: • Apply advanced statistical and machine learning techniques to answer the research questions. • Uncover meaningful patterns and relationships to identify potential digital biomarkers. Presentation & Documentation: • Compile findings in a written report detailing the background, methodology, results, and discussion. • Ensure the codebase is well-documented and reproducible. • Outstanding results may lead to the opportunity to prepare a high-quality manuscript for publication.

oriella.gnarra@hest.ethz.ch

This project proposes the development of a constrained reinforcement learning (RL) framework for social navigation, emphasizing explicit safety and collision avoidance constraints rather than relying solely on the reward function. By utilizing constrained RL, the approach seeks to eliminate the need for intricate reward tuning that often leads to overly conservative or risky behaviors. This work involves exploring various constraints based on collision detection, velocity limits, and geometric considerations like velocity obstacles.

Tasks

• Perform a literature review on Constrained Reinforcement Learning.

• Design a Constrained Reinforcement Learning Framework for Social Navigation.

• Train and Evaluate RL Agents for Social Navigation.

• Conduct Validation Experiments.

Requirements

• Proficient programming skills in Python.

• Solid understanding of RL algorithms, preferably constrained RL algorithms.

• Experience with neural networks, including frameworks like PyTorch.

• Familiarity with robotics, navigation algorithms, and path planning.

A safe Navigation policy for social environments.

ralyassi@ethz.ch