DriveCap Navigator is an in-vehicle navigation autonomous system that captures relevant navigation features and provides driver-specific navigation advice. We use machine learning techniques to build models of driver-specific preferences so that the navigator system can provide tailored information to her. Navigation is a special case for common denominator driving tasks due to its inherent map-level planning nature. As such, this system is a special case of DriveCap that solely focuses on transforming traditional safety agnostic and somewhat uniform in-vehicle navigation approaches into a something that directly supports safer driving at a personalized level.
For example, an older driver might have considerable trouble with left hand turns, glare when cresting a particular hill at night, and congestion. Driver specific rules will be learned by the system based on observed behavior and scenario perception. The route planning system will then utilize this information to provide driver-specific routing (e.g., avoiding left hand turns). Likewise, vehicle limitations will be inherently captured as rules too (e.g., conversion vans have trouble with ground clearance).