A car that drives itself
Our self-driving technology
Our vehicles are equipped with the best sensors available, giving us a 360-degree picture of the world hundreds of meters from the car. We extensively leverage machine learning to develop state-of-the-art perception models, and have built a rich data pipeline that allows us to continuously improve based on real-world driving. Our multiple sensor modalities give us great perception, while improving the safety and operational redundancy of our autonomous service.
Predicting the future
To make the right driving decisions, our vehicles don’t just need to know where other actors are right now, but where they’re going next.
Our prediction engine uses a combination of advanced probability models, high-definition maps, and time-based behavior models to predict exactly what’s going to happen around our vehicles. This is the foundation of a safe, comfortable ride - allowing us to know when to stop, go, yield, or pursue a more advanced driving maneuver.
Focus on safety
Safety is the top priority at Voyage. Our vehicles are equipped with hardware redundancies across all critical compute, power, and control surfaces. Algorithms are designed to produce orthogonal views of the world, and we work with third-party experts to review our systems in accordance with industry best practices. The Voyage team has a private closed-course test facility, and amazing simulation tools to help validate every software change before it makes it way onto open roads.
Enabling our engineers
Real-world miles are critical in building a self-driving car. Our in-house tooling supports a near real-time feedback loop from the field to engineering teams. From in-field note taking to in-browser visualization and playback tools, we’ve optimized our data pipeline for engineering efficiency. Voyage team members can easily search for, review, and request resimulation for any event observed during our autonomous driving sessions.