Isaac SDK 2020.1 Release Announcement

We’re happy to announce the latest version of NVIDIA Isaac SDK 2020.1 for robot developers everywhere. Download the Isaac SDK and get more information on this release @ https://developer.nvidia.com/isaac-sdk; Highlights of this release include:

New Isaac GEMs for Manipulation as well as Navigation

  • 3D perception with Refined Object Pose Estimation DNN using depth image.
  • Multi-LIDAR support for improved localization & obstacle avoidance.
  • Motion Planning for Manipulation using LQR Planner.
  • Docker container to enable Object Detection DNN training using Isaac Sim and NVIDIA Transfer Learning Toolkit.

New capabilities for Isaac Engine

  • Python API: Access an Isaac GEM or one of the reference applications directly from another development environment.

New Isaac Reference applications and platform

  • Laikago robot and much more

Isaac Sim 2020.1 (Unity 3D)

  • With Unity3D Plugin and support for HDRP pipeline.
  • Virtual Factory of the Future to Train, Test, Visualize and Debug Robotics AI.
  • Multi-robot simulation for validation and testing of the navigation stack.
  • Reinforcement Learning Framework and RL Docking Policy Example.

Isaac Sim 2020.1 (Omniverse Early Access)

  • This will be available soon.
  • Based on NVIDIA’s Omniverse Platform.
  • Support for Isaac SDK and Manipulation use cases.
  • To get more information and to register for Early Access to this simulator, please visit (https://developer.nvidia.com/isaac-sim).

Thank You!

The Isaac SDK team

1 Like

Great New of this week. :)
So glad to hear about Isaac SDK 2020.1.

Excited about this release! A couple of questions:
a) Does Omniverse have any enhancements to the navigation stack? Or is it primarily for manipulation? Should I use omniverse for navigation purposes or should I stick to the Unity3D build?

b) The documentation for 2019.3 mentioned that the navigation stack will have support for camera based navigation in a future release (as opposed to LIDAR based). Is the camera based navigation stack there as yet?

Hi arjunsengupta98

a) Does Omniverse have any enhancements to the navigation stack? Or is it primarily for manipulation? Should I use omniverse for navigation purposes or should I stick to the Unity3D build?
We are so glad that you are excited about the Isaac Sim (omniverse). The Isaac Sim 2020.1 (both omniverse and unity) support both navigation and manipulation with Isaac SDK. There are no enhancements on the navigation stack in the omniverse Isaac Sim. Is your choice to work with which Sim (omniverse or Unity) . if you have any questions about the Isaac Sim (omniverse) you can ask it here to get faster response.

b) The documentation for 2019.3 mentioned that the navigation stack will have support for camera based navigation in a future release (as opposed to LIDAR based). Is the camera based navigation stack there as yet?
No we still dont have camera based navigation.

1 Like

Hello,
What is the compatibility matrix for ISAAC with Jetpack versions. Everywhere I have seen, that “Jetpack 4.3” is explicitly mentioned. Does it mean “Jetpack 4.3 and Above”, can I use Jetpack 4.4. Or is this explicitly tied to the jetpack 4.3 only as mentioned?
Please advise.
Regards.
Kunal Vora.

Hi Kunal, you got it right, it is just JP4.3. We are working to get a release that supports JP4.4 soon. Stay tuned.

1 Like

Shrinv,
Thank you for that clarification. Helped save a lot of time.
While at it, can we also go about clarifying the CUDA GPU compatibility.
1)Jetson NANO : is this capable for development of application and deployeing as well. (aarch64 vs x86 question)?
2)I see you folks have some reference to Cross-compile packages, in the ISAAC folder is see 3 ‘packages’ folder.
packages, packages_jetpack43, packages_x86_64. Does it mean that i can develop and build the application on x86 laptop with no GPU, and deploy it on Nano(aarch64) like the L4T and VisionWorks samples?. Or does it mean that even the x86 development laptop also requires the a CUDA supported GPU for development and build.?
3) What all CUDA Supported GPUs work for isaac for resources(8gb, 12gb, 24 gb cards) and compatibilty point for view. Can I pick any CUDA GPU from that page, and RTX, any Quadro to Develop+Build+Deploy the application?

Please advise.
Thank you for the time.
Regards.
Kunal Vora

Hi Guys! I am installing dependencies for isaac 2020 dependencies on my jetson. As discussed in guide, I am installing dependencies by using install dependencies_jetson.sh in 2020isaac/engine/build/scripts/install dependencies_jetson.sh but when it installs ipy widgets with command python3 -m pip install --user ipywidgets it stuck by giving errors on one of the python library named argon2-cffi cffi like this:

  Failed building wheel for cffi
  Running setup.py clean for cffi
Failed to build argon2-cffi cffi
Installing collected packages: pygments, setuptools, decorator, wcwidth, prompt-toolkit, backcall, parso, jedi, pickleshare, ptyprocess, pexpect, ipython-genutils, six, traitlets, ipython, jupyter-core, tornado, python-dateutil, pyzmq, jupyter-client, ipykernel, pycparser, cffi, argon2-cffi, Send2Trash, entrypoints, webencodings, pyparsing, packaging, bleach, MarkupSafe, jinja2, pandocfilters, testpath, mistune, defusedxml, zipp, importlib-metadata, attrs, pyrsistent, jsonschema, nbformat, nbconvert, terminado, prometheus-client, notebook, widgetsnbextension, ipywidgets
  Running setup.py install for cffi ... error
    Complete output from command /usr/bin/python3 -u -c "import setuptools, tokenize;__file__='/tmp/pip-build-zquz_n6d/cffi/setup.py';f=getattr(tokenize, 'open', open)(__file__);code=f.read().replace('\r\n', '\n');f.close();exec(compile(code, __file__, 'exec'))" install --record /tmp/pip-e6ti_omb-record/install-record.txt --single-version-externally-managed --compile --user --prefix=:
    Package libffi was not found in the pkg-config search path.
    Perhaps you should add the directory containing `libffi.pc'
    to the PKG_CONFIG_PATH environment variable
    No package 'libffi' found
    Package libffi was not found in the pkg-config search path.
    Perhaps you should add the directory containing `libffi.pc'
    to the PKG_CONFIG_PATH environment variable
    No package 'libffi' found
    Package libffi was not found in the pkg-config search path.
    Perhaps you should add the directory containing `libffi.pc'
    to the PKG_CONFIG_PATH environment variable
    No package 'libffi' found
    Package libffi was not found in the pkg-config search path.
    Perhaps you should add the directory containing `libffi.pc'
    to the PKG_CONFIG_PATH environment variable
    No package 'libffi' found
    Package libffi was not found in the pkg-config search path.
    Perhaps you should add the directory containing `libffi.pc'
    to the PKG_CONFIG_PATH environment variable
    No package 'libffi' found
    running install
    running build
    running build_py
    creating build
    creating build/lib.linux-aarch64-3.6
    creating build/lib.linux-aarch64-3.6/cffi
    copying cffi/vengine_cpy.py -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/recompiler.py -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/commontypes.py -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/setuptools_ext.py -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/model.py -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/cffi_opcode.py -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/lock.py -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/ffiplatform.py -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/api.py -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/pkgconfig.py -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/verifier.py -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/error.py -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/__init__.py -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/backend_ctypes.py -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/vengine_gen.py -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/cparser.py -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/_cffi_include.h -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/parse_c_type.h -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/_embedding.h -> build/lib.linux-aarch64-3.6/cffi
    copying cffi/_cffi_errors.h -> build/lib.linux-aarch64-3.6/cffi
    running build_ext
    building '_cffi_backend' extension
    creating build/temp.linux-aarch64-3.6
    creating build/temp.linux-aarch64-3.6/c
    aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -DUSE__THREAD -DHAVE_SYNC_SYNCHRONIZE -I/usr/include/ffi -I/usr/include/libffi -I/usr/include/python3.6m -c c/_cffi_backend.c -o build/temp.linux-aarch64-3.6/c/_cffi_backend.o
    c/_cffi_backend.c:15:10: fatal error: ffi.h: No such file or directory
     #include <ffi.h>
              ^~~~~~~
    compilation terminated.
    error: command 'aarch64-linux-gnu-gcc' failed with exit status 1
    
    ----------------------------------------
Command "/usr/bin/python3 -u -c "import setuptools, tokenize;__file__='/tmp/pip-build-zquz_n6d/cffi/setup.py';f=getattr(tokenize, 'open', open)(__file__);code=f.read().replace('\r\n', '\n');f.close();exec(compile(code, __file__, 'exec'))" install --record /tmp/pip-e6ti_omb-record/install-record.txt --single-version-externally-managed --compile --user --prefix=" failed with error code 1 in /tmp/pip-build-zquz_n6d/cffi/

i tried python3 -m pip install --user ipywidgets and python3 -m pip install ipywidgets but still this error is coming, Can anyone please guide… Also is it necessory thing to install for running subsequent tasks on jetson ??

thanks in advance.

1 Like

Thanks for replying! I am running this command on the Jetson TX2 terminal, which is acting as companion or target Platform for Isaac SDK. I don’t know why it’s not building the wheel ipy widgets.

Having the same problem when trying to install jupyter lab on my xavier-nx JetPack 4.4 (L4T R32.4.3)

Found the missing package needed for the argon2-cffi error!

sudo apt install -y libffi-dev

1 Like

Thanks for this. I had the same problem with the Jetson Nano with JetPack 4.4 from a fresh 18.04.5 install.

Hello,
I did face same issue on Nano. Now my question is that is this command ‘engine/build/scripts/install_dependencies_jetson.sh -u <jetson_username> -h <jetson_ip>’ expected to run natively on the Jetson Bot or is it expected to run from the host?

Also austin70i2l: that thing he suggested did get me away from argon2-cffi build failure, but the end of the dependency install looks like following:

environment: line 13: jupyter: command not found
Connection to 192.168.1.80 closed.
Rebooting the Jetson device-

Is this a successful install? or is this aborted . I dont see the argon2-cffi issue afte rruning what austin suggested. But how to verify successful install?
Any one please advise.
Regards,
Kunal Vora.

This command runs from the host and connects to the Jetson over SSH. Obviously you need to supply <jetson_username> and <jetson_ip> accordingly.

I also noted this, but deploying to the Jetson seems to work OK (so far as I have tested it). I checked that ‘jupyter’ is recognised command on the Nano and it is in my case.

The Audio GEMS which I need have been removed from 2020.1NX for some reason, can you find out what is going on? Should I just downgrade to 2020.1 or will they be available somewhere else?