• capsule endoscopy
• maximum power point tracking
• active rectifier
• wireless power transfer
• implantable medical devices
• analog memory
• accelerators
• neuromorphic computing
• analog design

In recent years, despite the increasing popularity of digital circuits, analog Integrated Circuit (IC) design has remained a fundamental tool in the electronic field, thanks to some unique features such as the ability to handle continuous signals, and to the growing development of mixed-signal systems requiring expertise in both domains.
In particular, analog circuits have revealed themselves particularly suited for the field of neuromorphic computing, which is ideal for efficient, low-power processing of data at the edge of the Internet of Things (IoT) network. Of particular importance are IoT devices for healthcare applications such as Implantable Medical Devices (IMDs), which require real-time monitoring and control with minimal power consumption. Neuromorphic computing for IMDs could enable local analysis and processing of the acquired data, provide personalized treatment, all with the highest efficiency and minimal power consumption.
Additionally, the world of IMDs in in sharp development and offers several new challenges for analog IC design, since it aims to integrate a growing number of electronic sensors and systems in miniaturized devices, which need to work in harsh conditions with limited power budget. An important field of research is therefore represented by Wireless Power Transfer (WPT) techniques to open the possibility for IMDs to operate in autonomy and safety, avoiding the use of wires or batteries.
This is the motivation at the basis of this thesis, that proposes an analog approach to neuromorphic IMDs. Our primary focus has been on the development of analog ICs tailored to the deep neural networks and IMDs domains. To thoroughly explore unique analog design issues for each application, we have separated these two areas into distinct parts: in the first part, we present the design of an analog vector-matrix multiplier, a neuromorphic engine based on high-density non volatile memories realized in a single-poly CMOS technology; the second part of the thesis will deal with the design and fabrication of a WPT system for an implantable capsule, with particular focus on the receiver IC embedding maximum delivered power tracking and minimization of power losses in the body.