Intrinsic Flowstate aims to change robotic software development

Listen to this article

Intrinsic unveiled a new robotic software development solution today with the goal of making it easier to develop applications for robots. The Intrinsic Flowstate software adds a layer of abstraction to traditional robotic software tools with the goal of reducing the development skillset required while making it easier than ever before to program a robot in simulation.

Wendy Tan White, CEO of Intrinsic said this in her product keynote, “We’re making the ability to program intelligent robotic solutions as simple as standing up a website or mobile application. To be clear, we don’t believe a single software product or service will have this impact. This is an ecosystem effort. Only through the collective contributions of a large development community can we achieve this kind of change.”

Intrinsic acquired ROS maker Open Source Robotics Corp in December 2022, and Brian Gerkey, co-founder and now-former CEO of Open Robotics, joined Intrinsic as part of the OSRC team and continues to serve on the board of OSRF.

Gerkey said, “We’re continuing to increase our pace of development on these tools, as you’ll see most imminently in the next ROS distribution, named Iron Irwini, which will be released just next week. As a first step, we’re focusing on those of you who are designing solutions with industrial robots. We’re starting with solution designers because we know first-hand how hard it is to build, test, and deploy robot applications. Intrinsic aims to deliver a platform that fills the gaps in current robot software and overcomes the hurdles that prevent broader participation in robotics today.”

Intrinsic Flowstate combines simulation and graphical programming

Flowstate is starting with support for industrial robots. This class of robotics operates in a semi-structured setting, which simplifies many of the system variables. The product roadmap is to build a set of “skills” (in software) that can be extended to other classes of robots in the future. Ultimately, the roadmap for Flowstate is that it will function for mobile robotics and mobile manipulators at some point in the future.

Industrial robots are expensive and difficult to use, making them a niche tool. They require domain-specific knowledge and programming skill sets, making it difficult to modify how they work over time. Often, industrial robot workcells are programmed, certified and then left unchanged until a production line is rebuilt.

Torsten Kroeger, CTO said, “We want solutions to be portable across robots – starting with industrial robots. We want those solutions to seamlessly integrate with the other equipment in production environments. That’s why we’re building our platform with interoperability in mind: making it extensible and compatible with a wide range of robots, sensors, and automation equipment. We are betting on open interfaces and established industry standards.”

Intrinsic Flowstate software combines process editing with work cell state management and simulation. | Credit: Intrinsic

The Flowstate beta release is coming

Intrinsic Flowstate includes a number of features in the beta release. The primary capabilities enable the editing of the work cell process, the management of the work cell “scene”, and the capability to simulate the function of the program.

In detail, the beta release includes:

• An intuitive, web-based developer experience to facilitate robotic solution design from concept to deployment.
• Ability to author a sequence and layout your work cell in the same digital environment.
• Ability to simulate and validate solutions – all without touching a single piece of hardware.
• Skills provide a way for experts to encode their domain knowledge into reusable modules. A skill could be anything: pose estimation, manipulation, path planning, if it’s something a robot solution needs to do, it can be packaged as a skill.
• A fully configured development environment providing clear APIs to contribute new skills to the platform.
• A hardware-abstracted approach to development, primarily through the real-time control framework and Hardware Abstraction Layer.
• In the scene editor, you will be able to upload static geometries or a complete scene of objects. Add components, such as a robot, from the resource catalog, and onboard new components if you don’t see what you need in the catalog.

Jake McCarter, head of product, said: “Anyone should be able to train a machine learning skill and put it to work. We’re now lucky enough to have many of the core ROS and Gazebo developers as colleagues here at Intrinsic. Together we’re going to offer you a seamless and powerful developer experience for creating robotics solutions.”

The skills catalog is a large library of “reusable skills”, including skills provided by Intrinsic, and skills that have been contributed by developers or other users on the platform. Skills can be easily parameterized from the graphical process editor, with:

• User-provided values
• References to resources or products in the world
• Or robot poses that have been taught and saved in frontend

Stefan Hinterstoisser, perception engineering lead explains, “At Intrinsic, we’ve been developing advanced technologies that allow anyone to add perception capabilities to their solution – without needing specific domain knowledge or a background in computer vision. These capabilities include advanced AI and machine learning models that can be trained to recognize an object and determine its position and orientation in space. This can happen via one or multiple cameras mounted on the robot or in its environment. No complex training pipelines. No days were spent annotating training data. All of our pose estimators can be trained using synthetic data, with just a CAD model of the object you want to detect.”

Intrinsic announces partnership with Comau

Intrinsic announced a partnership with robot manufacturer and global systems integrator Comau. The companies have worked together over several years to validate this approach through real-world use cases. Using Flowstate and the underlying platform, Comau has recently created a modular solution to enable the assembly of a plug-in hybrid electric vehicle (PHEV) supermodule. This is just one example of the possibilities envisioned for solution builders and the types of solutions that could soon become possible to build and affordable to buy.

Flowstate is welcoming applications for the beta program, which will run for roughly eight weeks beginning in July 2023. Apply today at intrinsic.ai/beta