DASL Wiki

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
====== 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 : 

{{youtube>UBDt9J7M5PU?large}}\\ 

===== 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 [[https://sites.google.com/site/drcsimulatorforhubo/home|here]].

To understand how this tutorial works and what it is doing follow this tutorial: [[drexel_darwin_ball_tracking_gazebo|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.