Location, Location, Location

We are exhibiting RTLS technology at the Healthcare World Asia conference today, so just thought to share a little bit about my experiences playing around with different variants of the technology.

We purchased an active RTLS system a couple of years back, this was based on proprietary 2.4GHz protocol using time-of-flight trilateration technology. The intent was to track the positions of forklifts as they moved about a warehouse. Installation was pretty painful, we mounted 4 readers/emitters 25 metres apart high up in the corners of a small warehouse. The claimed accuracy for the system was 0.3m. We expected the accuracy to drop due to the dense warehouse racking environment, but what we experienced surprised us - we could not even get anywhere close to even 5m of accuracy.  This means we couldn't even place the forklift truck in the correct aisle. We were in for another shocker: the 4 pieces of "AAA" batteries in a tag lasted only 3 weeks! We were later told by the vendor to reduce the tag's beaconing frequency to once every half hour - which would make the locating system about as useful as a flashlight on a sunny day.

We later explored Wi-Fi based RTLS solutions, but cost alone was absolutely prohibitive: we were quoted six figures for the "locating engine" software alone. We also quickly found out that if we wanted to do RTLS, we couldn't use "normal" routers and access points; we had to use industrial grade high sensitivity APs that cost a couple of thousand dollars - same as the price of a typical active reader. From chatting with industry experts, we found out that WiFi RTLS deployments are simply not working well in real life. Users complained about accuracy problems, with the system often pointing them to not only the wrong room, but the wrong floor!

After more research, we found an alternative to trilateration based technology - Infra-Red RTLS. We tested this and were happy to discover that it works really well, thus we are partnering with RF Code as the first company to bring this technology into the region. The theory is simple, RF waves passes through walls but infra-red waves don't. A lost-cost infra-red emitter floods a room with infra-red signals, with no spillover into adjacent rooms. Thus, a tag positively identifies it is inside a certain room, and reports its room ID to a single shared reader than can be more than 50m away. The advantages:

- better battery life than WiFi (up to 5 REAL years). WiFi is designed as a data transmission protocol with speed as its primary objective, and not power consumption

- lower hardware cost - much fewer readers needed, as well as lower network installation costs

- lower backend cost - trilateration is computationally resource intensive. Imagine a scaled-up deployment of 50,000 tags with a 10-second resolution - this means 5,000 trilateration algorithm computations every second, resulting in high-end server hardware requirements

- lower software cost - middleware costs a fraction of the 6-figure "locating engine" we were quoted

- more accurate positioning - sure, with IR you won't be seeing fancy visualisations of 1,000 tag moving about on a map in real time. But in practical terms, all a user is probably interested in is the answer to a simple question: Where is my asset now? And my bet is that  infra-red RTLS can answer this question better! 

Demo in a box: location accuracy up to 10 cm!