Skip to content

Repository files navigation

Neuroplastic Drone Obstacle Avoidance

This folder is independent from approximate_controller_version and contains the neuroplastic two-neuron controller version of the PyBullet drone simulation.

What is implemented

  • Two front LiDAR rays, separated by 40 degrees.
  • Two recurrent non-spiking tanh neurons.
  • Self-excitatory synapses and mutual inhibition.
  • Online correlation-based synaptic plasticity with synaptic scaling.
  • A live PyBullet drone view with the blue travelled path and visible LiDAR rays.

The obstacle maps are reconstructed from the paper figure because exact map coordinates were not published.

Windows setup

Install Python 3.12 (64-bit) and select Add Python to PATH during its setup. Then double-click install_pybullet.bat once.

Run the neuroplastic simulation

Double-click run_neuroplastic_maze.bat for the maze environment.

Or run this from Command Prompt:

cd /d "C:\Users\work\OneDrive\Documents\drones\neuroplasticity_controller_version"
py -3.12 live_drone_sim.py --env maze --controller paper --duration 180 --show-neurons

--show-neurons prints the two neural outputs and changing synaptic weights in the Command Prompt while the PyBullet window runs.

About

Fully autonomous drone for adaptive obstacle avoidance, simulated across multiple platforms

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages