@Karle - yes there are multi-constellation receivers that use
GPS, Glonass, Galileo and BDS. The problem, if for example GPS fails, is "what is the failure mode?" and would the receiver know not to use GPS. That is to say if the GPS satellites are still transmitting, just with bad info, then the combined solution will be bad.
@SeanPatrick - I think the existing MEMS gyroscopes are more than adequate for your suggestion of an artificial horizon. The problem with the
current small, low cost gyros is their long term drift. In an inertial nav unit you combine accelerometer data inferring the direction of gravity and gyro data to account for the near term motion to determine the direction of "up", and this can be done very will with
current cheap small devices. The reason you need a better gyroscope for a full inertial
navigation solution is because you cannot use the direction of gravity to determine the direction of "North" and so need a gyro that you can count on for a long time period without drift. You can use a magnetometer for this but of the three
sensor types, gyros, accelerometers, and magnetometers, mags are the most problematic. One reason is there is much more variation in the magnetic field from place to place than in the gravitational field so the produce an accurate solution you need a very accurate magnetic map/model, or a very good gyroscope.