Multi-modal Navigation Interaction Recommendation with a Driver Demand-Based Knowledge Graph

The personalized human-vehicle interaction system provides the driver with a way of interacting that suits them, allowing them to interact with the car system without interfering with normal driving activities, which is essential for driving safety. In general, drivers’ preferences or needs vary depending on the person, the scenario, and the task of the trip. This article takes navigation, a high-frequency human-vehicle interaction, as an example, and constructs a Driver demand Knowledge Graph (DKG) by using multi-modal and hierarchical thinking that focuses on the driver’s needs and consideration of multiple fine-grained factors. Ultimately, a multi-modal navigation interaction recommendation method through knowledge graphs is proposed, which represents nodes from multiple aspects and extracts a multi-relational graph structure for DKG.