Today we’re excited to announce the release of Falcon 5.1, a focused update that brings several user experience and workflow enhancements based on our work with customers and EDU users. 

How can you use Falcon to train a vision model like YOLOv8? In this blog we present a high-level walkthrough from simulation scenario creation to model training and testing. Let's dive in!

Today marks the arrival of our educational license — designed for anyone looking to learn to build better AI models with digital twin simulation. Meet EDU.

Interview with Duality AI co-founders Apurva Shah and Michael Taylor on how cutting edge technologies of the entertainment industry are solving the steepest challenges in off-road autonomy.

With an upgrade to Unreal Engine 5.3, the Latest Falcon release brings even higher simulation accuracy with enhanced real-time performance — all while lowering the bar for digital twin creation!

Falcon 4.2 brings a host of exciting updates — chief of which is FalconEditor: an intuitive visual tool designed to streamline and accelerate the adoption of digital twin simulation.

Putting task coding to the test: the same natural language prompted mission is carried out in a digital site twin in Falcon, and then in the corresponding real-world location. Let's see how it went!

An effective approach to make robotic systems less brittle and more adaptable, while keeping their behavior predictable — And how digital twin simulation in Falcon is bringing it to the real world.

Speculating how generative AI synthetic data may compare with real world data and physically based simulation techniques. 

Our first release of 2024 brings exciting new updates to FalconSim, FalconCloud, and FalconEditor. In this blog we dive into the brand new features rolling out today.

ViperGPT shows great potential as a visual system for Embodied AI. We use FalconCloud to learn how VipertGPT sees the world. 

Modularity and reusability are key features driving digital twin simulation — so how do we create real interoperability that ensures a meaningful and long lasting digital twin infrastructure? 

KEF Robotics provides state-of-the-art autonomy software to fly aircraft without a human pilot. With Falcon, KEF successfully took on unique challenges faced by tethered drones. 

Today we’re announcing an expansion of the Falcon product suite with the debut of two new products: FalconCloud and FalconEditor.

Creating high-fidelity Digital Site Twins can be a daunting task. Duality's AI-powered pipeline enables our customers to build a digital twin of any desired location from already available data!

Can advanced internet AI models be harnessed to make robots smarter and safer — making them our natural collaborators? We explore Embodied AI, and how digital twin simulation can be its bridge.

Powerful integrations are vital to Falcon. We show one way of co-simulating a rocket launch with MATLAB Simulink and Falcon for generating higher accuracy results.

Duality's visual designer Mia Rincón takes us through her journey of creating the new Duality logo.

This project earned Duality a spot on Fast Co's World Changing Idea List. We walk through how we built it by integrating Falcon with AWS SimSpace Weaver.

The Open Geospatial Consortium wants geospatial data to be Findable, Accessible, Interoperable, and Reusable. We agree, and propose our standard for truly interoperable Digital Twins.

Field reporting from SIGGRAPH 2022 — and the importance of the open metaverse for geospatial data projects.

The enterprise metaverse is a Fortnite but for digital twins. Duality co-founder, Apurva Shah presents a perspective for developers on the challenges, and opportunities, ahead.

Part 3 of how we quantify if our synthetic data is relevant and valuable using the Three I’s frameworks: Indistinguishability, Information Richness, and Intentionality.

Part 2 of how we quantify if our synthetic data is relevant and valuable u.

Training a model on poor quality data can cause it to pick up “bad habits” that don’t transfer well to real-world scenarios. So how do we quantify if our synthetic data is relevant and valuable?

USD is game-changing for composing simualtion scenarios, but parsing it is complex and slow. Meanwhile loading pak files in Unreal is a breeze. Can we have the best of both worlds?

Efficiency drives much of Duality’s purpose — and it led us to develop a novel approach to simulating a probe sensor for guiding an autonomous warehouse robot fleet (with many more applications).

Duality’s founder, Apurva Shah first shared the framework for metaverse design while teaching his “Designing the Metaverse” class at California College of the Arts (CCA). This year, we brought it i...

How to design high-fidelity, realistic environments for digital twins and simulation with maximum workflow efficiency.

“Game engines cannot be physically accurate simulators” is a maxim in need of retirement. We prove it false by demonstrating determinism in Unreal Engine.

Falcon, built on top of Unreal Engine, enables testing and training of autonomous systems in environments so realistic that machine learning networks can’t tell the difference between the synthesiz...

In this end-to-end operations demo, two autonomous vehicles (quadcopter and a compact track loader), are trained to perform a search and retrieve mission of airdropped supplies.

Through the lens of interaction design, Duality CEO, Apurva Shah, proposes a unified, practice oriented definition of the metaverse.

The Metaverse is an elusive idea. Duality’s CEO Apurva Shah provides a brief overview of what we mean by this term in diverse and useful contexts.

Unreal Engine is not only meant for large, expensive GPUs...

Achieving a 90% reduction in the development timeline of a drone-based detection system via virtual environments and high-quality synthetic data.

We use digital twins of consumer off-the-shelf Quadcopter and digital camera to determine the land coverage of a fire at a specific GPS location.

Improving pipeline integration for robotics teams and reducing ramp up time with Unreal Engine and Duality’s USD importer.