Say Hello to FalconCloud — Digital Twin Simulation for the Entire Smart System Product Lifecycle
November 9, 2023
· Written by
Today, we’re announcing an expansion of the Falcon product suite with the debut of two new products: FalconCloud and FalconEditor (see our press release). This expansion of the Falcon digital twin platform serves to make digital twin simulation workflows even more seamless, efficient, and accessible, enabling collaboration on a much larger scale.
Duality was founded with a vision inspired by what we know best: building virtual worlds and developing robots. Falcon is the simulation platform that takes that knowledge and translates it into real customer results. FalconSim has been employed by organizations like DARPA to take on some of the thorniest autonomy hurdles in the RACER challenge. Companies like Honeywell and P&G have spoken about the accuracy of results they were able to achieve with FalconSim. We even made it on to Fast Company’s World Changing Idea list thanks to the work carried out with our partners at Amazon Web Services simulating an emergency response scenario in a pulsing metropolis.
Over the course of these diverse projects our vision has persisted: empower anyone to solve the most difficult problems in robotics by making advanced digital twin simulation as accessible as possible. While we always made sure that our users could run Falcon without need for custom hardware, the smart system market has evolved, and simulation needs are extending into collaborative workflows well beyond research and development.
Engineers and data scientists need accurate synthetic data without the hurdle of becoming simulation experts. Logistics coordinators need to test and tune new robotic systems in simulation prior to costly deployment. Marketing professionals need simulation solutions that let them exhibit smart system performance in scenarios relevant to their customers. Use cases like these call for a simulation platform that is flexible, collaborative, and accessible.
This is why we are thrilled to share today’s exciting news: the arrival of digital twin simulation in the cloud, available to anyone via a user-friendly web app.
FalconCould debuts as part of the Falcon 3.3 release. In this blog, we’ll tell you all about FalconCloud, and cover all the exciting updates that we’ve made to the Falcon product suite.
Let’s get into the details!
Simulation has long faced challenges that hinder teams interested in adopting it for the full product lifecycle. These hurdles range from IT burdens, to long ramp up times, and complicated scenario building. Until today, Duality’s best-in-class digital twin simulator, FalconSim, was only available for installation on local hardware with a GPU requirement.
FalconCloud is a major step towards changing this paradigm.
With FalconCloud, Duality is making FalconSim fully accessible from a web browser, enabling users to run high-fidelity simulations at massive scale, without any special hardware or software installations. By removing these infrastructure barriers, FalconCloud is enabling collaborative simulation projects for users across entire organizations and with external stakeholders.
Key Features and Benefits
Digital Twin Catalog All digital twins, environments, systems, and items, live in FalconCloud’s digital twin catalog. Simulation scenarios are composed from lego-like, modular digital twins, and as users’ digital twin assets increase, a management solution quickly becomes imperative. A centralized digital twin catalog that enables searching, filtering, and sharing by (or between) teams, makes it much easier to find and reuse relevant twins to author any number of novel simulation scenarios.
Scenario Catalog Scenarios, just like digital twins, are infinitely reusable and shareable; thus, managing scenarios in the catalog works exactly the same way. Users can manage, share, and run any simulation scenarios effortlessly in the cloud.
To simulate any available scenario, users simply launch it directly from the catalog. This in turn creates a session — a unique simulation instance in which the scenario plays out, sensor data is generated, and outcomes are observed.
This allows users to run simulation scenarios in the cloud without complex hardware or specialized knowledge, enabling entire teams to share scenarios, letting anyone leverage simulation for their unique workflows.
Session Manager Users can seamlessly download session data from all cloud-based sessions and immediately use it for AI/ML training, sharing with collaborators, or for any other purpose.
Enhanced Documentation System Effective documentation is critical for enabling productive simulation work. The new system is more user-friendly, with significant improvements like advanced search capabilities to help users find relevant answers and related topics intuitively.
Tutorials FalconCloud features a growing library of tutorials to introduce anyone to digital twin simulation workflows, and all the ways they can be leveraged to generate valuable synthetic data.
Pricing Tiers that Reflect Use Types FalconCloud is designed to enable collaboration, and this is reflected in our new pricing structure. Team members will engage with Falcon on different levels. A developer may need to author brand new scenarios, while an ML engineer may only need to run the scenarios to generate data. An external stakeholder may only need to see a specific simulation run or a scenario set up. FalconCloud tiers (including the free Share tier) are now structured to better meet our customers’ needs. See more details here.
We are also introducing an early release of FalconEditor. A defining feature of simulation in Falcon is the modular creation of digital twins and simulation scenarios. The addition of FalconEditor enables more efficient scenario creation supported by powerful visualization tools.
Key Features and Benefits
Unreal Editor Foundation The best-in-class level editor is extended with added features specifically designed for digital twin and scenario authoring.
Visual Scenario Authoring FalconEditor enables users to visualize and manage the position of assets spawned from the USD-based scenario. This forms the foundation of an integrated development environment (IDE) where users can ensure that their scenario is set up as intended, and that everything from environment to sensors is functioning as expected.
Integrated FalconSim FalconEditor allows users to evaluate scenario execution in FalconSim, including observing sensor output, all without leaving FalconEditor. By verifying the scenario directly in FalconEditor, users can rapidly iterate while authoring scenarios, and experience a more frictionless simulation process.
Scenario and Pak file Creation FalconEditor allows users to create both the simulation asset and scenario files for use in FalconSim. FalconEditor directly saves scenario edits to your scenario file and it enables users to modify Unreal levels and directly create Pak files with those changes.
Early Release Requirements Available as a local install for Windows. Offered to current Pro-tier customers for an additional fee.
FalconSim - Version: 3.3.1025
FalconSim is the simulation platform at the heart of the Falcon product suite. Along with making it accessible via FalconCloud, this release introduces significant FalconSim improvements.
Key Features and Benefits
Capture Sensor The capture sensor represents a major improvement to Falcon’s sensing architecture, resulting in improvements in both usability and sensing speed for our users. The Capture Sensor offers all of Duality’s camera sensors in a single sensor, simplifying sensor configuration. It can be queried for individual sensing modalities, such as an RGB color image, but also provides additional scene information, thanks to Duality’s custom build of the Unreal Engine, enabling sensing data available nowhere else. This new architecture also provides a noticeable increase in sensor rendering speed.
Importantly, the Capture Sensors enables new camera modes, with zero loss of any pre-existing sensor options and functionality.
User-Created Sensors This new system also allows advanced users to write their own highly performant sensor models. Users can now write sensor models that operate on the GPU, resulting in extreme performance, but with the ease of working inside Falcon’s Python API. This enables users to build new sensing modalities or implement custom noise models beyond Falcon’s current sensor library.
Ready to try Falcon for yourself? Start exploring digital twin simulation today with a free account at https://falcon.duality.ai/ .