ROS ANFIS RL

Setup

Clone the repository into the Python 3 catkin workspace

Simulation

1. In a non python 3 environment. Source the anfis repository
2. Run ./simulation.sh script to start Gazebo environment

Program

1. Source the anfis repository in Python 3 environment
   1. source the Python 3 virtual environment as stated below in section Python 3 Setup using souce /path/to/py3venv/bin/active
   2. Source the ROS catkin workspace source ~/python3_ws/devel/setup.bash
2. Ros run or run in Python 3 virtual environment main.py

Activate the environment:

run: source ../../py3venv/bin/activate

do another source: source ~/python3_ws/devel/setup.bash

Tensorboard log visualizations

Local machine

ssh -L 16007:127.0.0.1:16007 nvidia@192.168.0.12

Server

tensorboard --logdir=runs --samples_per_plugin images=999 --port=16007

tensorboard --logdir=runs --samples_per_plugin images=999

Otherwise, run the script: tensorboard.sh in there select the port to set up the server to.

Python 3 environment setup for ROS Melodic

The tutorial is based on https://youtu.be/oxK4ykVh1EE?t=1232

The instructions steps are also demonstrated here. There are some differences demonstrated below. https://github.com/AUVSL/Jackal-maintenance-and-upgrade/blob/main/python3_setup_instuctions.sh

Line 28 be sure to change the file path to the location of your python virtual environment. Rebuilding the Python environment catkin clean; catkin build --cmake-args -DPYTHON_EXECUTABLE:FILEPATH=/home/auvsl/python3_ws/py3env/bin/python

Ros launch setup

git clone https://github.com/ros/ros_comm.git

pip install defusedxml netifaces

option 1 to try and get roslaunch and rosrun to run in Python3 Melodic

pip install -U catkin_tools

option 2 to try and get roslaunch and rosrun to run in Python3 Melodic

git clone https://github.com/catkin/catkin_tools.git

cd catkin_tools

pip3 install -r requirements.txt --upgrade

python3 setup.py install --record install_manifest.txt

python3 setup.py develop

git clone https://github.com/ros/roscpp_core