Asynchronous ADC design for neuromorphic systems

At IMEC (Thesis)

We propose to work towards the development of analog-to-digital converted tailored for neuromorphic processing systems. This novel type of analog-to-digital converter will be compatible with standard Address Event Representation scheme that is currently used in state-of-the art neuromorphic event-based systems. The ADC will have to perform data compression in the analog domain and near the sensor.

 

What you will do

Battery life time is essential for many wireless communication systems in IoT and biomedical applications. To prolong life time, non-uniform sampling ADC approach was introduced to reduce the number of conversions while still being able to conserve the required information. The ADC operation is asynchronous without a sampling clock, and the output of the ADC is processed further with a neuromorphic network. Neuromorphic computing systems exploits digital pulses for both computation and communication, it is therefore crucial to develop novel and low-power ADCs that can directly interface with these new class of massively parallel processing units.

 

In this project, we propose to do develop an ultra-low-power analog-to-digital converter system that can be directly interfaced to neuromorphic spiking neural processing circuits. This novel ADC will enable data compression by directly computing on the large amount of data available in the analog domain, resulting in the transmission of only the low-bandwidth outcome of the preprocessing. The output of the ADC will be a pulse of digital events (spikes) that will be sent to remote neuromorphic computing or actuating modules.

 

The project duration is of 12 months. This project can result, depending on maturity and performance, of the design in a tape-out of the designed circuits; part of the work will be spent in carrying out testing and characterization of the system. We target to publish the outcome of the project in high impact conference or journal.

 

We seek very motivated candidates with electrical engineering and/or biomedical background, with strong knowledge of the RF and analog design workflow and acquainted with the Cadence tool chain.

 

The tasks in this project include:

  • Literature survey on low-power asynchronous ADCs and the specific needs for neuromorphic devices.
  • Propose a low-power and miniaturized ADC with event-based output.
  • Schematic and layout design of selected architecture.
  • Implementation and measurement of the test-chip.
  • Thesis writing and documentation in imec-Holst Centre.

What we do for you
As the world-best research institute, we are the center of the excellence in nano-electronics design for Internet of Things and Healthcare applications. In this internship project, you will be working on the cutting-edge research project, under the supervision of the world-renown researchers from diverse background.

 

Who you are

  • Electrical Engineering Background with good analytical skills.
  • Knowledge of analog design circuits using CMOS transistors.
  • Structured way of reporting, both orally and written.
  • Motivated student eager to work independently and expand knowledge in the field.
  • Good written and verbal English skills.

 

Interested

Click on apply  to submit your application. You will then be redirected to e-recruiting.


Please be advised that non-EU/EEA country students that are studying outside of the Netherlands, need to have a work-permit to be able to do an internship in the Netherlands.

Back