Artificial intelligence in GPS navigation systems

GPS navigation systems use stored map information for determining optimal route selection based on a shortest path algorithm. This technique is quite successful in getting you to where you want to go in a reasonable time and is fault tolerant in the sense that it can automatically reroute in case of error. One disadvantage of this approach is that it does not have any memory. It does not automatically remember the actual time it took you to get there nor does it learn from that experience and use the actual measurements to improve future route selection. A simple method for modifying a GPS navigational system to incorporate a simple learning paradigm using velocity profiles is described. In addition to learning, these velocity profiles can also be used to extract features from the environment which can then be used to further improve the accuracy of optimal route selection. It is assumed to be completely autonomous which means that it requires no user input or intervention. All of the required information is derived from recording GPS location, date and time.