This is an old revision of the document!
Table of Contents
Optimal Preview Control Simulation on DRC-Sim(Gazebo)
The simulation on Gazebo is very similar to the cycloid simulation done in the previous tutorials. Only this time the trajectory points to for the robot to follow is different.
The trajectory points generated is imported into the plugin that.
In this plugin the robot is programmed to take 6 pairs of steps and can be modified to take and pre-defined number of steps.
Here is a video link of the Darwin-OP walking :
Walk through tutorial to implement the above on the DRC Simulator(Gazebo)
This tutorial assumes that you have Gazebo - version 1.7.1 installed and also Github Installed and have basic knowledge of Gazebo. To understand Gazebo you can follow the tutorials I have authored here.
To understand how this tutorial works and what it is doing follow this tutorial : Simulation on Gazebo
Step 1
Clone from the Git repository :
git clone https://github.com/Bharadwaj-Ramesh/Darwin_OP.git
Step 2
Copy the contents to the gazebo Models directory
cp -r Darwin_OP ~/.gazebo/models/
Step 3
Go into the directory where everything is copied into
cd ~/.gazebo/models/Darwin_OP/Darwin_Optimal_Preview_Control_Walking
Step 3
Create a build directory and compile the plugins in the build directory.
mkdir build cd build cmake .. make
Step 4
Add the plugins to Gazebo plugin path.
export GAZEBO_PLUGIN_PATH=`pwd`:$GAZEBO_PLUGIN_PATH
Step 5
Move the world file to the build directory.
cd .. cp Darwin_OPC_walking build cd build
Step 6
Open the world file that we just copied using Gazebo in pause state.
gazebo -u Darwin_OPC_walking
Run the world and you will see that the Darwin-OP walks to the ball.
The current version has friction issues and hence the Darwin is not programmed to turn towards the ball if it is not at the center of the Darwin-OP's frame. But it walks closest(.1m) to the ball and stops.