Localisation had to be trusted in situations impossible to drive safely and repeatedly, with no reliable map to check against. We generated those scenarios, fused the sensors and stress-tested the software.
Result. a repeatable validation platform, built to OpenDRIVE and ADASIS standards.
The programme needed high-quality image, video and 3D training data at volume, for years. We ran it as a long-term data operation with layered quality checks.
Result. consistent, production-grade data, project after project.
Dense cities and bad weather break most localisation. Our researchers built self-learning maps and weather-adaptive methods using lower-cost sensors.
Result. new methods for low visibility and heavy occlusion, published internationally.
The field lacked a rich dataset of real, unstructured traffic. We collected and labelled more than 100GB of driving behaviour and released it openly.
Result. METEOR, used by researchers around the world.
We drove a national highway stretch with AXIRA to show condition data, asset inventory and a digital twin from a single drive.
Result. the full pipeline proven end to end on a live highway.
Cameras in every store, but no view of how customers shopped. We read the existing footage for footfall, demographics and behaviour in real time.
Result. clear daily insight for store teams, with no new hardware.
Hyper-realistic avatars that listen, speak and respond naturally, built as an Intel software solution for service, education and events.
Result. personal, round-the-clock digital engagement.
Operators in the US, North America and Africa needed platforms that could keep up. We built cloud-native platforms, a recommendation engine and analytics.
Result. reliable systems, sustained over multi-year engagements.
Trusted by teams at
