Skip to content

Researching and subsequently implementing data compression algorithms on Internet of things(IoT) devices

Notifications You must be signed in to change notification settings

ASCII-Mentorships/Data-Compression-in-IoT-devices

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 

Repository files navigation

Data-Compression-in-IoT-devices

Researching and subsequently implementing data compression algorithms on Internet of things(IoT) devices

Project Idea

Why do we need Data Compression on IoT devices?

  • To enable Federated Learning on IoT devices.
  • Federated Learning removes the need to send data from IoT devices to cloud, for applying machine learning algorithms. The machine learning model is downloaded in the device itself (example smartphone) and the data generated updates the model, which is then encrypted and sent to cloud.
  • Federated Learning allows for smarter models, lower latency, and less power consumption, all while ensuring privacy.
  • However to accomplish the training of the model on the IoT device itself, the enormous data generated by the device needs to be effectively compressed and stored in the device. This poses a problem as the IoT devices have limited memory and processing capabilities.
  • Thus the aim of our project is to find effective data compression algorithms for resource constrained IoT devices to enable federated learning.

Project tentative timeline

  1. Summer 2022 - Read and analyze research papers regarding the project (In progress).
  2. September 2022 - Setup the IoT devices in the DASH Lab and start preliminary implementation of algorithms.
  3. October and November 2022 - Find the algorithms that help achieve optimum compression ratio and speed with given constraints.
  4. December 2022 - Finalize algorithms and tools. Carry out real-time simulations.
  5. Subsequent work for 2023 - Publish research paper if the results are promising.

Prerequisites

  1. Good grasp on fundamentals of Data Structures and Algorithms and/or Machine learning.
    • We will use concepts like sliding window (LZ77 compression algorithms) and trees (Huffman coding)
    • We will use certain machine learning techniques like data deduplication for achieving our goal.
    • Either DSA or ML knowledge is sufficient.
  2. Proficiency in either C or C++ or Java or Python, knowledge of more than one language is better.
  3. Understanding of basic computer components such as what is a processor, memory etc.
  4. Willingness to learn and a passion for problem solving.
    • We will be working on some cool stuff such as Raspberry PIs and Arduino :)

Project leaders

  1. Hitarth Kothari
  2. Amogh Sinha

Project supervisor

Dr Arnab K Paul

About

Researching and subsequently implementing data compression algorithms on Internet of things(IoT) devices

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published