Self-Driving Benchmarks

The age of connected vehicles (human and self-driven) raises several major concerns regarding safety and responsibility. What legislation or guarantees will protect the consumer?

First measurements have been collected by the California DMV regarding disengagement reports. The frequency of human driver intervention as well as the time elapsed are required information to be reported. A clear benefit is that improvements in self-driving capabilities as well as the important safety issue of human driver handoff is measured.

However, is this enough? What other metrics for security, privacy, and availability are required?

The cloud infrastructure (e.g., AWS, Azure) largely regulates itself in regards to availability and uptime. Cloud services compete with each other based on reputation and track record of service features, availability, and resistance to attacks.

Is it reasonable to expect car manufacturers to follow suit and compete on safety, security, and privacy issues? How will this be enforced without transparency?

Largely, privacy has been seen as a consumer driven initiative where each individual mandates and asks for privacy. However, in the age of connected vehicles it will be the companies that are pushing for privacy as they are completely unwilling to share both the collected information and algorithms they have developed.

Is it reasonable to expect that such companies will be transparent and forthcoming regarding security attacks, zero-days, and privacy breaches? At what time scale is acceptable?

These are all issues that we will examine in future posts.