Fetal Movement Monitor

Fetal movement monitoring is crucial for assessing fetal well-being, as changes in activity can indicate potential perinatal complications. However, current methods, such as maternal perception and ultrasound, have limitations in accuracy and suitability for regular monitoring outside clinical settings. Therefore, there is a need for an objective, reliable, and accessible fetal movement monitoring system to reduce the risk of perinatal complications.

This project follows the proposal of clinical validation of the device, from manufacture to performance evaluation. The aim of the project is to achieve accurate recognition of fetal movements, ultimately improving the understanding of their relationship with birth outcomes.

Project Objectives:

• Ensure safe and ethical collection of high-quality sensor data and observed fetal movements from pregnant participants.
• Develop a signal processing and machine learning algorithm trained on collected data to recognize fetal movements.
• Evaluate the algorithm's performance on unseen data using appropriate performance metrics.

Multi-modal sensor device and sensor placement.

Key Findings:

• Successfully assembled and tested multiple prototypes of the device, integrating various sensor technologies and data annotation software to ensure seamless functionality.
• Conducted comprehensive data collection with pregnant volunteer participants, gaining valuable experience in collaborating with multidisciplinary teams of medical professionals and engineers.
• Developed a strong understanding of the importance of adaptability and thorough preparation when working with public participants, particularly in managing incomplete or imperfect data and dealing with real-world constraints.
• Leveraged advanced machine learning techniques, specifically a Transformer-based model, to accurately classify different types of fetal movements, contributing to the overall success of the project.