Wayve, a developer of Embodied AI technologies, has launched GAIA-2, the latest iteration of its video-generative world model for assisted and automated driving. Building on its first iteration, GAIA-1, the new model offers enhanced diversity, realism, and control in synthetic video data generation. Through it, Wayve is aiming to accelerate the development and validation of its end-to-end AI software for driving.
The company highlighted that GAIA-2 represents a deviation for general-purpose text of video generative AI models by being purpose-built for assisted and automated driving, citing its ability to maintain consistency across several camera viewpoints, while generating diverse geographies and driving conditions.
In carrying out this purpose, GAIA-2 leverages new advancements that enhance its ability to support the training and validation of advanced driving technologies. Here, its Enhanced Fine-Grained Control over driving dynamics enables greater scene generation with control over the ego-vehicle behavior, the behavior of other road agents, and environmental factors such as road configurations (lane structure, intersections, crossings), weather, and time of day.
At the same time, the updated model has been trained on a large-scale, curated, dataset from the UK, U.S., and Germany; diverse vehicle platforms such as cars and vans; as well as various sensor configurations and frame rates. This training helps the model generate realistic, adaptable and case-rich synthetic data aligned with the surround camera setup typically featured on modern, software-defined vehicle architectures.
Throughout its operations, GAIA-2 more broadly ensures spatial and temporal coherence across multiple camera viewpoints, providing a surround-view perspective of the vehicle’s environment. This, Wayve underscored, is crucial for training and testing driver assistance and automated driving AI as it replicates the real-world multi-camera setups used in these systems today.
Beyond generating video, Wayve further highlighted the validation and verification of AI models across everyday and safety-critical driving scenarios. The company said that GAIA-2 enables this at scale by augmenting real-world data with highly controlled, repeatable, and diverse synthetic scenarios.
Using GAIA-2, Wayve can simulate multiple common safety-critical scenarios, generating rare and high-risk events that can be impractical or unsafe to capture through real-world data collection alone. This also facilitates the stress testing of Wayve’s AI driving models, ensuring they are prepared for both common and safety-critical situations in a controlled environment.
