Mixed-signal Frontend Design with Convolutional Neural Network Computational Capacity

Leuven - Master projects/internships
More than two weeks ago

Enhancing the performance of data driven deep learning network by exploring the analogue and mixed signal sensor frontend with convolutional neural network computational capacity.

Recent effort on biomedical signal processing for health monitoring has shown promising results by using machine learning (i.e. deep learning) based algorithms. While the training phase can be performed offline on optimized hardware, implementing such neural network models on resource constrained wrist-worn platforms incurs research challenges in terms of chip area and power consumption. This introduces a new opportunity to design an energy-efficient sensor frontend within a mixed-signal network to perform part of the neural network related computation, which offloads digital processing. The mixed-signal network provide high power/area efficiency together with good reconfigurabilities. The research will involve understanding a) PPG-applications; b) deep learning algorithms, c) hardware architecture (conventional, CIM, time-domain, mixed-signal etc.) implementing the ML algorithm, d) low power analog and digital circuit design to perform basic neural related computation and implementing neural network related computational capabilities.

Type of project: Combination of internship and thesis

Duration: 6-9 month

Required degree: Master of Engineering Technology, Master of Engineering Science

Required background: Biomedical engineering, Electrotechnics/Electrical Engineering

Supervising scientist(s): For further information or for application, please contact: Didac Gomez Salinas (Didac.GomezSalinas@imec.be) and Shuang Song (Shuang.Song@imec.be) and Dwaipayan Biswas (Dwaipayan.Biswas@imec.be)

Imec allowance will be provided

Share this on


This website uses cookies for analytics purposes only without any commercial intent. Find out more here. Our privacy statement can be found here. Some content (videos, iframes, forms,...) on this website will only appear when you have accepted the cookies.

Accept cookies