Lab: Hardware Accelerator

KI-IoT

Nowadays, there is a high demand for developing a computing platform to meet the requirements of internet of things (IoT) applications such as energy efficiency, low power consumption, etc. Meanwhile, artificial intelligence (AI) methods have become the main part of several IoT systems like autonomous driving, mobile communication, etc. Therefore,  new requirements for AI-based IoT systems should be taken into consideration such as reliability, robustness, and security. AI-based IoT systems are expected to enhance operating efficiency, improve risk management, increase IoT Scalability, and introduce new security servers.

KI-IoT is a research project that investigates and develop a robust AI-based IoT system.

Project Goals

KI-IoT aims to develop a RISC-V-based platform including embedded non-volatile memories for the design of system-on-a-chip solutions. In order to ensure the reliability and trustworthiness of the proposed platform, the platform will follow the concept of "Safety and Security by Design" and provide optimized power management and AI accelerator components with new memory technologies. For more information, you can visit BMBF website.

Our Role:

The role of the Institute of Computer Engineering (ITI) is to first develop a system model of overall SoC. Such a model is expected to provide an overview of SoC reliability. ITI is also responsible for developing the RISC-V-based platform and establishing hardware interfaces of the system components such as the interface between the main processor and  RRAM-based Accelerators and the interface between the main processor and classical DSPs.

Our Partners:

In KI-IoT, we are woking together with the following partners:

  • IHP - Leibniz-Institut für innovative Mikroelektronik, Frankfurt (Oder).
  • Infineon Technologies AG,
  • NXP Semiconductor,