I couldn't tell you about any patents. But my company (former, really) did a lot of work on precise automotive location tracking systems, and found that GPS is not accurate enough or consistent enough. We and most other automotive researchers in advanced safety systems used a hybrid of GPS and inertial guidance, usually with a map database with enhanced detail.
Accuracy: commercial or easily available GPS is really only accurate to the 5 or 10 meter neighborhood, depending on the circular error probable. Adding fixed ground GPS transmitters helps a lot, but now we are depending upon new infrastructure. Plus, when you are out training in the mountains, will the signals be available? Generally stationary transmitters are placed to cover urban areas and shorelines.
Consistency: GPS results get better or worse depending on the number of satellites that can be recieved. It tunnels or city business districts ("urban canyons") satellites can be outright blocked. If you need to determine the vertical dimension you need more minimum satellites than if you only need latitude and longitude.
I know that some of these methods are patented.
Map databases have issues, one being cost of licensing the database, say from NavTek. Plus they can have errors, and might not be totally accurate in a less-commonly used dimension, like altitude. Bikes might add some new requirements, like tracking the difference between riding the inner edge of the curve versus the outer edge. Precision tracking systems such as I worked on are stretching to tell you for sure what lane you are in, or distinguishing lane from shoulder.
I could see work being done to resolve some of these issues. There is someplace to go, more room for development. But the magnet/sensor unit is just a part of the inertial navigation sensors that, for high accuracy and consistency, would need to supplement GPS.