search

reiniscimurs / DRL-robot-navigation

Deep Reinforcement Learning for mobile robot navigation in ROS Gazebo simulator. Using Twin Delayed Deep Deterministic Policy Gradient (TD3) neural network, a robot learns to navigate to a random goal point in a simulated environment while avoiding obstacles.
MIT License
487 stars 97 forks source link

support for other planning methods #58

Closed SHITIANYU-hue closed 8 months ago

SHITIANYU-hue commented 1 year ago

Hello, it is an excellent repo for navigation. I am wondering in your paper( https://arxiv.org/abs/2103.07119) you mentioned several other planners, such as ROS local planner package, Dijkstra algorithm. May i know if it supports some classical path planning algorithm? Currently, I could only find TD3's implementation. thanks!

reiniscimurs commented 1 year ago

Hi,

This repo does not support that and is just about training the local planner through TD3 network. The repo referred to in the paper is https://github.com/reiniscimurs/GDAE However, that is also mostly to serve as a guideline for development and would not work "out of the box". The implementation of some sort of global planner is to create waypoints in known environment. Then these waypoints would be visited using local planner (TD3 policy in this case). It would not be that difficult to implement such a system from scratch.