Full-stack autonomous driving moved out of simulation and into two test vehicles
A next-generation urban mobility platform had an autonomous driving stack that performed in simulation and had never been integrated into a production-representative vehicle.
Perception, planning and control had been developed by separate teams against separate assumptions. Sensor calibration, compute budgets and real-time timing constraints only surface on real hardware — and the programme needed a demonstrable vehicle, not a report, to justify continued investment.
We took ownership of integration end to end: sensor set definition and calibration, middleware and timing architecture, then staged bring-up from bench to closed course. Once the vehicle was driving, we wrote the technical proposal that translated the capability into a fundable shuttle service line.
- —Sensor suite specification, mounting and calibration
- —Perception, planning and control integration on target compute
- —Middleware, logging and real-time timing architecture
- —Staged validation from bench to closed-course driving
- —Technical proposal for an autonomous shuttle business line
Two production-representative vehicles ran the full stack on-vehicle, giving the programme a demonstrable capability rather than simulation results. The accompanying proposal established a new shuttle business line as a fundable route to market.
Have a problem that looks like this one?
Start with a 45-minute discovery call. No pitch, no deck — we align on the problem and tell you honestly whether we are the right team for it.