Workflow

Setting up an IsaacROSDev Container

https://nvidia-isaac-ros.github.io/concepts/docker_devenv/index.html#development-environment
https://nvidia-isaac-ros.github.io/getting_started/dev_env_setup.html
https://nvidia-isaac-ros.github.io/robots/nova_carter/getting_started.html#nova-carter-dev-setup

Reference Example

https://nvidia-isaac-ros.github.io/reference_workflows/isaac_perceptor/run_perceptor_in_sim.html

Let's Get Started

We begin by cloning isaac_ros_common and nova_carter repos to the src folder of our local workspace. My local workspace is ~/workspace/humble_ws

git clone https://github.com/NVIDIA-ISAAC-ROS/isaac_ros_common.git
git clone https://github.com/NVIDIA-ISAAC-ROS/nova_carter.git  

Start the IsaacROSDev Container (from the workspace...)

./src/isaac_ros_common/scripts/run_dev.sh  -d  ~/workspace/humble_ws/

Then run ls in the terminal to confirm that isaac_ros-dev is set to your host host workspace.

Alternatively, you can run this one from inside workspace/src/isaac_ros_common/scripts. The -d command explicity sets the workspace to workspaces/isaac_ros-dev inside the IsaacROSDev container.

./run_dev.sh -d  ~/workspace/humble_ws/

Installations onto the Container....

First, lets get updated...

sudo apt-get update  
rosdep update

We'll need curl too to for later when we download assets into the isaac_ros_assets folder...

sudo apt-get install -y curl tar

Install LTTng-UST...

We'll need this for SMACC later...

sudo apt-get install -y lttng-tools  
sudo apt-get install -y lttng-modules-dkms  
sudo apt-get install -y liblttng-ust-dev  

Install Jetson Stats

sudo apt-get install -y ros-humble-isaac-ros-jetson-stats

Use rosdep to install Nova Carter bringup dependencies...

rosdep install -i -r --from-paths ${ISAAC_ROS_WS}/src/nova_carter/nova_carter_bringup/ --rosdistro humble -y

Install Nvblox From Debian...

sudo apt-get install -y ros-humble-isaac-ros-nvblox && \
rosdep install isaac_ros_nvblox

Download the nvblox assets

Set variables for isaac_ros_assets workspace folder...

NGC_ORG="nvidia"
NGC_TEAM="isaac"
NGC_RESOURCE="isaac_ros_assets"
NGC_VERSION="isaac_ros_nvblox"
NGC_FILENAME="quickstart.tar.gz"

REQ_URL="https://api.ngc.nvidia.com/v2/resources/$NGC_ORG/$NGC_TEAM/$NGC_RESOURCE/versions/$NGC_VERSION/files/$NGC_FILENAME"

Create isaac_ros_assets workspace folder...

mkdir -p ${ISAAC_ROS_WS}/isaac_ros_assets/${NGC_VERSION} && \
   curl -LO --request GET "${REQ_URL}" && \
   tar -xf ${NGC_FILENAME} -C ${ISAAC_ROS_WS}/isaac_ros_assets/${NGC_VERSION} && \
   rm ${NGC_FILENAME}

Download the isacc_ros_ess and isaac_ros_peoplesemsegnet models into the isaac_ros_assets folder (takes a while)

Source setup.bash since the packages are already installed...

source /opt/ros/humble/setup.bash

Set env variable so you don't have to manually accept every EULA...

export ISAAC_ROS_ACCEPT_EULA=1

Run the shell scripts

ros2 run isaac_ros_ess_models_install install_ess_models.sh
ros2 run isaac_ros_peoplesemseg_models_install install_peoplesemsegnet_vanilla.sh
ros2 run isaac_ros_peoplesemseg_models_install install_peoplesemsegnet_shuffleseg.sh

Install isaac_ros_object_detection and other perception packages from Debian... (optional)

We'll start with pointcloud_to_laserscan...

sudo apt-get install -y ros-humble-pointcloud-to-laserscan

Install isaac_ros_detectnet

Then we'll get into isaac_ros_object_detection, starting with detectnet

sudo apt-get install -y ros-humble-isaac-ros-detectnet 

Download the isaac_ros_detectnet assets

Set variables for isaac_ros_assets workspace folder...

NGC_ORG="nvidia"
NGC_TEAM="isaac"
NGC_RESOURCE="isaac_ros_assets"
NGC_VERSION="isaac_ros_detectnet"
NGC_FILENAME="quickstart.tar.gz"

REQ_URL="https://api.ngc.nvidia.com/v2/resources/$NGC_ORG/$NGC_TEAM/$NGC_RESOURCE/versions/$NGC_VERSION/files/$NGC_FILENAME"

Create isaac_ros_assets workspace folder...

mkdir -p ${ISAAC_ROS_WS}/isaac_ros_assets/${NGC_VERSION} && \
   curl -LO --request GET "${REQ_URL}" && \
   tar -xf ${NGC_FILENAME} -C ${ISAAC_ROS_WS}/isaac_ros_assets/${NGC_VERSION} && \
   rm ${NGC_FILENAME}

Install isaac_ros_rtdetr

sudo apt-get install -y ros-humble-isaac-ros-rtdetr

Download the isaac_ros_rtdetr assets

Set variables for isaac_ros_assets workspace folder...

NGC_ORG="nvidia"
NGC_TEAM="isaac"
NGC_RESOURCE="isaac_ros_assets"
NGC_VERSION="isaac_ros_rtdetr"
NGC_FILENAME="quickstart.tar.gz"

REQ_URL="https://api.ngc.nvidia.com/v2/resources/$NGC_ORG/$NGC_TEAM/$NGC_RESOURCE/versions/$NGC_VERSION/files/$NGC_FILENAME"

Create isaac_ros_assets workspace folder...

mkdir -p ${ISAAC_ROS_WS}/isaac_ros_assets/${NGC_VERSION} && \
   curl -LO --request GET "${REQ_URL}" && \
   tar -xf ${NGC_FILENAME} -C ${ISAAC_ROS_WS}/isaac_ros_assets/${NGC_VERSION} && \
   rm ${NGC_FILENAME}

Now download the Nvidia SyntheitcaDETR model...

mkdir -p ${ISAAC_ROS_WS}/isaac_ros_assets/models/synthetica_detr && \
cd ${ISAAC_ROS_WS}/isaac_ros_assets/models/synthetica_detr && \
  wget 'https://api.ngc.nvidia.com/v2/models/nvidia/isaac/synthetica_detr/versions/1.0.0/files/sdetr_grasp.etlt'

Then we'll convert the encrypted model (.etlt) to a TensorRT engine plan and drop it in the isaac_ros_assets/models/synthetica_detr folder...

/opt/nvidia/tao/tao-converter -k sdetr -t fp16 -e ${ISAAC_ROS_WS}/isaac_ros_assets/models/synthetica_detr/sdetr_grasp.plan -p images,1x3x640x640,2x3x640x640,4x3x640x640 -p orig_target_sizes,1x2,2x2,4x2 ${ISAAC_ROS_WS}/isaac_ros_assets/models/synthetica_detr/sdetr_grasp.etlt

Then get back to the workspace...

cd /workspaces/isaac_ros-dev/

and install this package so you can test your vision pipeline later...

sudo apt-get install -y ros-humble-isaac-ros-examples

Install isaac_ros_image_pipeline

sudo apt-get install -y ros-humble-isaac-ros-depth-image-proc
sudo apt-get install -y ros-humble-isaac-ros-gxf-extensions
sudo apt-get install -y ros-humble-isaac-ros-image-pipeline

These should already be installed but if you just want to make sure..

sudo apt-get install -y ros-humble-isaac-ros-image-proc
sudo apt-get install -y ros-humble-isaac-ros-stereo-image-proc

Assemble the Workspace

Clone the Application Repos...

git clone https://github.com/robosoft-ai/SMACC2.git  
git clone https://github.com/robosoft-ai/sm_isaac_perceptor_1.git  
git clone https://github.com/robosoft-ai/rrt_exploration.git  

Workspace Operations

Create a file called .isaac_ros_common-config with the following context:

cd src/isaac_ros_common/scripts

echo -e "CONFIG_IMAGE_KEY=ros2_humble.nova_carter\nCONFIG_DOCKER_SEARCH_DIRS=(../../nova_carter/docker ../docker)" > .isaac_ros_common-config

Create a file called .isaac_ros_dev-dockerargs with the following context:

 echo -e "-v /etc/nova/:/etc/nova/\n-v /opt/nvidia/nova/:/opt/nvidia/nova/" > .isaac_ros_dev-dockerargs

Clone the perceptor dependency repositories using the vcstool file in the nova_carter repository:

cd /workspaces/isaac_ros-dev
vcs import --recursive src < src/nova_carter/nova_carter.repos

A ton of beta ish dependencies are added in this step to make the original perceptor demo work.

Build Workspace

source /opt/ros/humble/setup.bash

rosdep update
rosdep install --ignore-src --from-paths src -y -r

colcon build --symlink-install

Launch Application

Source the workspace...

source install/setup.bash

ros2 launch sm_isaac_perceptor_1 navigation.launch.py \
mode:=simulation run_rviz:=True

Reference Application Launch Command...

ros2 launch nova_carter_bringup navigation.launch.py \
mode:=simulation run_rviz:=True