I really don't want this job. It relates to intellectual property I developed a few years ago, but I simply could not contain myself. I must point out that your approach is wrong. Dijkstra does not apply well in a wireless environment. It would work well if all nodes exist in a wired mesh but in a wireless universe "links" are very transient in nature and under certain conditions we are lucky if they persist even for the duration of a packet transmission. Any information you have about the network link-state is obsolete by the time you realize it. By the way, I was able to achieve location within 6-inches in an office environment over a 10m2 area using 6 reference nodes (e.g. Anchors). Also, RSSI is a very poor indication of link connectivity. Good Luck with your project! It sounds like fun!