Navigation System Parking Availability Prediction

The patent creates a virtual copy of real-time parking availability data by training machine learning models on historical parking data. These models generate predicted parking availability information that replicates the essential characteristics of real-time data without requiring continuous communication infrastructure. The virtual copy is sufficiently accurate for navigation decisions while eliminating the need for expensive real-time sensing and communication systems at parking facilities.

The system performs preliminary action by collecting and analyzing historical parking availability data in advance to train machine learning models. These pre-trained models are then deployed to generate parking availability predictions without requiring real-time data collection. The preliminary training phase captures parking patterns and behaviors, enabling the system to predict future availability based on current time, date, and contextual factors without continuous communication infrastructure.

The patent replaces the mechanical communication system (sensors, transmitters, receivers, and communication networks) with an information processing system based on machine learning. Instead of physically transmitting real-time data from parking facilities to navigation systems, the system uses trained models to compute parking availability predictions locally based on historical patterns and current contextual inputs. This substitution eliminates communication reliability concerns while maintaining prediction accuracy.

The navigation system performs self-service by generating its own parking availability information through machine learning models rather than relying on external communication from parking facilities. The system uses its own historical data and contextual information (time, date, weather, events) to independently predict parking availability, making it autonomous and independent of external communication infrastructure.