Materials for electronics

+ Learn more

Memristive devices are relatively new (first implemented in 2008 in a solid state device) and are based on oxides rather than on semiconductors. Although some memristor devices have already found their path towards product commercialization, the lack of a solid theory and basic understanding of the underlying physics is impeding mass production in many cases. There is not even agreement on the material system which presents the best properties for the different applications. Research on oxide materials is required for the development and design of new devices with better properties and enhanced reliability. Thus, neuromimeTICs is involved in the research of new materials, both complex and binary oxides, to enable innovative memristive devices.

Our research is devoted to the development of silicon-based technologies for the fabrication of microelectronic devices based on the resistive switching phenomenon, and to the study of the related phenomena through the electrical characterization of high-k dielectrics-based MIS and MIM structures, as memristors for memory and neuromorphic applications.

Electron Devices.png

Electron Devices

+ Learn more

Electron devices is the enabling technology for the design and implementation of the neuromorphic circuits which should lead to the fabrication of hardware-based machine learning systems to face actual societal challenges. neuromimeTICs is involved in advanced research on memristor devices based on binary and complex oxides. The Institute of Microelectronics of Barcelona of the National Microelectronics Centre (IMB-CNM) and the Institute of the Materials Science of Barcelona (ICMAB), provide fabrication and integration facilities, and physical /electrical characterization infrastructure. This allows neuromimeTICs to develop physical and compact models for these devices based on real data. Compact SPICE-like models to be incorporated into neuromorphic circuit simulators and the design of innovative devices are the final products of our research on electron devices.


Neuromorphic circuits

+ Learn more

Memristors are resistances with non-volatile memory effects. These devices show behavior similar to that of biological synapses showing both short term and long term plasticity effects. On the other hand, relatively simple CMOS circuits can implement integrated and fire electronic neurons. Combining these two elements, electronic neurons and synapses, it is possible to design and fabricate neuromorphic circuits that implement deep learning artificial neural networks. These hybrid memristor/CMOS circuits are the basis for deep learning industrial control systems (ICs) which should be able to implement spatially distributed, energy efficient, secure and trustful artificial intelligence applications in many fields requiring real-time local processing of data and immediate action (e-Health, intelligent mobility, etc.). neuromimeTICs research in this field is based on SPICE simulation of neuromorphic circuits, investigation of circuit architecture, characterization of learning mechanisms and algorithms for supervised and unsupervised systems.


AI for health

+ Learn more

The goal in this research line is twofold. On one side, to assist professionals with predicitive data analysis, which takes into account data from existing disparate monitoring devices and data from the Electronic Health Record (EHR). On the other hand, to unveil non-evident patterns in the data with the aim of gaining insight into biological processes and their causal effect


 AI for Automotive

+ Learn more

Design of Deep Learning (DL) solutions based on the real-time prediction of traffic congestion and car collisions. This research line also considers the enhancement of DL pre-processing techniques by using knowledge of Internet of Things/Internet of Vehicles systems (IoT/IoV) collecting Mobility data and sensors measurements.