Transfer and Online Reinforcement Learning in STT-MRAM Based Embedded Systems for Autonomous Drones

  • Authors:
    Insik Yoon (Georgia Tech), Malik Aqeel Anwar (Georgia Tech), Titash Rakshit (Samsung), Arijit Raychowdhury (Georgia Tech)
    Publication ID:
    Publication Type:
    Received Date:
    Last Edit Date:
    2776.044 (University of Notre Dame)
    2777.006 (Purdue University West Lafayette)


In this paper we present an algorithm-hardware codesign for camera-based autonomous flight in small drones. We show that the large write-latency and write-energy for nonvolatile memory (NVM) based embedded systems makes them unsuitable for real-time reinforcement learning (RL). We address this by performing transfer learning (TL) on meta environments and RL on the last few layers of a deep convolutional network. While the NVM stores the meta-model from TL, an on-die SRAM stores the weights of the last few layers. Thus all the real-time updates via RL are carried out on the SRAM arrays. This provides us with a practical platform with comparable performance as end-to-end RL and 83.4% lower energy per image frame.

4819 Emperor Blvd, Suite 300 Durham, NC 27703 Voice: (919) 941-9400 Fax: (919) 941-9450

Important Information for the SRC website. This site uses cookies to store information on your computer. By continuing to use our site, you consent to our cookies. If you are not happy with the use of these cookies, please review our Cookie Policy to learn how they can be disabled. By disabling cookies, some features of the site will not work.