THESIS - Embedded implementation and optimization of radar-based sensor fusion algorithms for patient monitoring
Izegem, Belgium
About Televic Healthcare
At Televic Healthcare, we are at the forefront of revolutionizing healthcare through cutting-edge communication and information technology. We specialize in designing, developing and manufacturing innovative solutions that enhance patient care, streamline processes and empower healthcare professionals. We work in a channel model through certified partners internationally and our ambition is to become the reference in our niche market through innovation.
Topic Description
The integration of advanced radar-based sensor fusion techniques in healthcare environments enables real-time and unobtrusive monitoring of patients’ activities and movements. This thesis focuses on implementing and optimizing digital signal processing and sensor fusion algorithms on an embedded platform with constrained computational and memory resources. The goal is to assess the feasibility of running these algorithms efficiently in a real-world healthcare setting while ensuring optimal performance in terms of accuracy, power consumption, and data efficiency.
During this thesis, you will:
- Study and compare sensor fusion algorithms: analyze the selected algorithms for movement classification and localization using radar data, comparing them with state-of-the-art alternatives. Evaluate their computational complexity, memory footprint, and real-time applicability for embedded systems.
- Implement and optimize for embedded hardware: Deploy the algorithms onto an embedded platform, ensuring efficient execution within system constraints. Optimize performance through techniques such as memory management, pipelining, and parallel processing.
- Evaluate system performance: Measure execution time, power consumption, and memory usage of the embedded implementation. Assess trade-offs between accuracy and efficiency.
- Analyze communication requirements: study the amount of data the embedded system needs to transmit to a backend system to achieve reliable performance. Analyze the relationship between data resolution, update frequency, and accuracy of movement classification and localization. Evaluate the impact of data reduction techniques on accuracy and computational efficiency.
- Validation and benchmarking: conduct controlled tests to evaluate system reliability and compare results with existing solutions. Provide recommendations for further optimizations based on findings.
Nature of the work
- Level: Bachelor, Master
- AI, Networking, Wireless, Embedded Software, Hardware
- Location: University and Televic Healthcare (Izegem)
- Number of students: 1 or 2