Introduction
In environments where GPS signals are unreliable or unavailable, ensuring the safety and real-time location of personnel remains a critical challenge. This research project aims to develop an innovative alert system utilising Inertial Measurement Unit (IMU) devices to address this pressing need. The proposed system will focus on indoor environments, offering a robust solution for tracking personnel and detecting potential distress situations.
The research will build upon recent advancements in the field of inertial odometry, leveraging the robust, world-leading IMU platform called actibelt. This platform has established its reliability and accuracy over two decades of use in clinical trials, demonstrating such exceptional robustness that its devices have even been deployed on the International Space Station (ISS). We will adapt and extend the capabilities of this proven platform to address the specific challenges of human motion tracking in indoor environments where traditional positioning systems are ineffective.
Key Objectives
The core objectives of this research include:
1. Designing an accurate indoor positioning system using IMU sensors, implementing advanced hybrid odometry techniques that leverage both traditional filters and neural networks.
2. Developing a novel fusion algorithm that integrates Extended Kalman Filter (EKF), Unscented Kalman Filter (UKF) or other with Long Short-Term Memory (LSTM) networks or Transformer architectures. This hybrid approach aims to improve the accuracy and robustness of position estimation, particularly in challenging indoor environments with complex motion patterns.
3. Implementing reliable fall detection and lack of movement recognition capabilities using a combination of rule-based systems and machine learning classifiers. This dual approach will enhance the system's ability to identify potential emergencies while minimising false alarms.
Potential Impact
This project has wide-ranging applications, from enhancing firefighter safety in burning buildings to improving worker security in complex industrial environments. It also holds significant potential for protecting miners operating in underground tunnels where traditional positioning systems fail, and for monitoring deep sea divers working in extreme underwater conditions where GPS signals cannot penetrate. By combining cutting-edge sensor technology with sophisticated data analysis and machine learning techniques, we aim to create a versatile solution that significantly improves personnel safety and operational efficiency in GPS-denied settings across diverse and challenging environments. (Each tracking scenario calls for a different methodology; in this project we will focus on just one of those.)
Expected Outcomes
The scope of this overarching project is extensive, and the final outcomes of the student’s subproject will be significantly influenced by the student's familiarity with various technologies, their understanding of existing research in indoor positioning and safety systems, and the choices made in collaboration with our team. Given the project's breadth, several potential outcomes are possible. These could include:
- Development of a robust indoor positioning system that accurately tracks personnel movement in GPS-denied environments, adapting existing state-of-the-art inertial navigation techniques to the unique challenges of human motion and requirements of the actibelt device.
- Creation of an advanced fall detection and lack of movement recognition algorithm that minimises false alarms while ensuring high sensitivity to genuine distress situations.
- Implementation of a novel fusion approach that uniquely combines traditional filtering methods with neural networks to enhance the accuracy and reliability of position estimation in complex indoor environments.
The student might focus on one of these areas or some other, depending on their skills and interests, as well as the project's evolving priorities.
