Alps Alpine has been working on an indoor navigation system that would enable a very precise location acquisition, good enough to let autonomous drones fly accurately and safely for monitoring missions, among other things.
Indoor navigation is one of these technologies that have been promised but never delivered commercially to consumers, so anything that seems to go in the right direction is of great interest.
This system from Alps Alpine is based on a mesh/network of RF (radio-frequency) positioning module that broadcasts positioning information. You can think of it as an indoor GPS system because the fundamental idea is the same: using the location of several nodes, it is possible to compute a precise position.
The specifications say that the precision can be as high as ~30cm and that in theory, it was possible to trade precision for higher coverage. But for now, and in this specific context, 30cm is what the technology is aiming for.
You can imagine that flying a drone through a door entrance does require this kind of precision, but a store location application for Shopping Malls may do just fine with a 1-2 yards accuracy. It just depends on the situation.
The positioning module be embedded in the drone uses an ultra-low power technology, somewhat similar to Bluetooth at an RF level, but not at a protocol level. While it’s is excellent from a power consumption standpoint, the relatively short range means that the positioning network nodes density has to be relatively high.
We praise the use of low-power technology,but it’s very likely that the electricity consumed by the drone to fly will dwarf by several orders of magnitude the power used by this location system.
We’re not quite sure what the cost of deploying a mesh is since this is still under development, but in the past, mesh density has been a limiting factor for real-world deployments of various indoor technologies.
Yet, this is one of the most advanced systems we’ve seen, and we think that it is possible to use it for particular applications that may require laying a network on a specific path, and not a complete volume of a building for example.
Also, it might be possible to lower the precision if the drone was smart enough to guide itself visually for very close-proximity obstacles such as doors. Also, it seems inevitable to have some kind of proximity sensing (visual, radar or lidar) to handle things that are not in the indoor map database.
Would this make it to a mall or a warehouse soon, and what will the real use cases be? We’ll keep an eye on this technology.