Skip to content

cueltschey/5G_dynamic_compression

Repository files navigation

5G dynamic compression using DRL

Overview

This is a project designed to test if using AI for dynamic compression algorithm selection is possible. It uses either a ZMQ emulated channel or a UHD USRP B210 to send video data over a channel. Channel metrics are then fed to a deep reinforcement learning model, which selects the compression type.

action state:

  • bfp
  • zstd
  • lz7
  • no compression

reward state:

  • total packet throughput
  • compression ratio
  • overall latency

input state:

  • channel quality
  • antenna gain
  • frequency
  • distance between radios?

Building

install deps:

sudo apt install -y libopencv-dev libuhd-dev uhd-host libzmq-dev

build:

mkdir build
cd build
cmake ..
make -j$(nproc)

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages