A big problem here is - as already mentioned earlier - the accuracy of the GPS. It is good for navigation, telling where you are in a city or in a country. But it isn't good for controlling things, since roads are seldom wide enough to let you use the GPS to keep a vehicle on the road. And even if you have a very large grass area, much fun will be lost because of this accuracy issue.

When running a robot in a limited area, it is normally easier to draw cables on the floor or in the ground that will send a signal the robot can hear. Then you can frame the allowed area. Or you may use ultra-sound speakers/microphones to detect the location. Indoors, you could use optical detection, in case you have a patterned floor. As long as you have good enough contrast and large enough markers, you can use cameras outdoors too, scanning a 3D area.

So in the end - exactly how do you expect to use this GPS? You could calculate speed from the wheels, unless you have problems with wheel slip. You could calculate speed from photographing the ground. You are already planning to use a compass, in which case you don't need a GPS for direction. A GPS is way lousier at measuring height than it is at predicting your location.

The only time a GPS is really good enough for positioning a robot or vehicle is if you are navigating on the water not too close to the coast or if you have a DGPS that is computing a differential location based on a land-based radio reference. But a DGPS is quite expensive.