Fleet Learning Technology
As new Wayve vehicles take to the road in new cities, our Fleet Learning Technology continually improves the AI software that drives them.
Fleet Learning Technology
What makes AV2.0 unique is that our intelligent autonomy software constantly improves from every mile driven across all fleet deployments. Our Fleet Learning Technology enables this network effect.
This platform brings together all of the capabilities needed to train, evaluate, and deploy machine learning systems that can operate safely at scale in the real world. Today, we’re building technology to collect data from fleets of vehicles, train models on vast amounts of driving data, create simulated worlds where it can learn how to drive in different environments and situations, and ensure that these models are safe before they’re deployed on real roads.
Large scale driving data
Data is a key ingredient for our autonomous driving software to learn how to drive safely and reliably in different environments. It is also crucial in evaluating the performance of our AI Driver.
We’ve built a robust system for collecting and analysing data from all on-road experiences and simulated environments to enable continuous learning. We continually expand our driving datasets to cover various environments, such as urban and suburban roads, during the day and night and in many weather conditions.
Groundbreaking MLOps
To ensure our models are developed responsibly, we’re developing groundbreaking MLops workflows—creating new tools, processes and pipelines that allow us to build, train and deploy billion-parameter AI models on vehicles more quickly and confidently.
We’re building new methods to manage our foundation models faster and more confidently across the entire machine learning lifecycle so that we can demonstrate that our AI Driver is safe and trustworthy.
Infinity Simulator
We are building a world-class simulator to train and evaluate our autonomous driver to function in any environment. Our simulator enables us to test our AI Driver on an infinite number of miles and an endless array of driving scenarios in diverse generated worlds. Virtual testing allows us to test situations that would otherwise be difficult or unsafe to try on real roads. We can also experiment in simulation with new vehicle platforms and different geographies before we build prototypes for real-world applications.
Learn more about our technology
Introducing GAIA-1: A Cutting-Edge Generative AI Model for Autonomy
GAIA-1 is a new generative AI research model that creates realistic driving videos and offers fine-grained control over ego-vehicle behaviour and driving scene features by leveraging video, text and action inputs. Ideal for research, simulation, and training purposes, it provides endless possibilities for innovation in autonomy.
Discover more
Evaluating driving performance in diverse simulated worlds
Wayve's latest research on how we are simulating the diversity of the real-world dynamic behaviour on our streets using multi-agent reinforcement learning
Discover more
Wayve and DPD launch a Fleet Data Collection Pilot Program in the UK
The companies are working together to explore how innovations in computer vision and machine learning can be applied to existing fleet and delivery operations to increase the safety of smart urban delivery solutions
Discover more