Redundancy of inputs might be the secret sauce that will lead to the expected safety improvements of autonomous transportation. In the above interview, Sean McCormack, Director, FOG/OEM Sales for KVH Industries, talks about KVH’s inertial sensor as a way for an autonomous vehicle to understand where it is at all times, even, in those cases where a tunnel might block a GPS signal.
He explains that their fiber optic-based, gyroscope provides the precision necessary to stay in lanes and detect drift. He doesn’t go into detail regarding how it works, but as pointed out by Jim Hayes in this interview, it is possible to infer the physical environment by its impact on fiber and the resulting change in optical characteristics. McCormack points out that, although it isn’t necessarily ready from a cost standpoint for single occupancy vehicles, there are other higher-value applications, such as shuttle and military, that can more easily absorb the costs. .
Stay tuned for an interview with Nvidia’s Danny Shapiro where he adds more detail about the concept of sensor-fusion and how the neural networks are teaching the cars to be better drivers.