From what I have read, I would concur with your comments about in-door positioning. Some time back Motorola set the goal of 1 meter accuracy, anywhere and anytime; however, that was a very tall order. Apple and Google have made good progress using their WiFi databases, though I doubt they are anywhere near the 1-meter level - at least on a consistent basis. I have heard of one startup getting near to 20 meters indoors using RF signals. I think that there should be viable markets for that level of accuracy.
The idea of using MEMS in-doors suggests using a dead-reckoning system that would include a compass, and this is actually a three-axis magnetometer wherein the magnetic field strengths are tilt compensated to get a heading. This biggest problem for these systems is having to deal with any sort of magnetic interference. This type of solution could be severely degraded by having some other magnetic filed nearby, such as another mobile phone that's too close. The company Sensor Platforms claim to have a solution to this type of interference, but I haven't followed up with any white papers from the company.
Fundamentally, the quality of the MEMS sensors determine the quality of the in-door position estimates. There are current solutions (at or above $25,000) that can provide the required accuracy, but I believe no one has been able to make the lower-cost MEMS sensors work at that level. Someday, the quality will improve enough to make the indoor solutions robust, but it may take several years to get there. It will be a horse race between the players emerging now. We will need to determine which companies have the hardware and, critically, the software expertise to make it happen. |
