Uber’s has shared its blueprint in hopes that it will become the industry standard for developing safe autonomous systems. It starts with a premise, that the company’s “self-driving vehicles are acceptably safe to operate on public roads,” and derives a comprehensive series of steps that must be developed and verified to meet it.

Rather than separately address the hardware, software, testing, simulation, or human training as separate components in an automated system—something standardization bodies like the International Organization for Standardization (ISO) are already doing—it places them within a single, larger safety framework. Uber’s safety guidelines end up looking like a cross between a checklist and a decision tree, with each item flowing from that premise that its vehicles are safe to drive on public roads.
More>>