They have observed that the stability of the received Bluetooth signal strength RSSI depends on which Channel 37, 38 or 39 the signal is being received on. This is because the channels slightly overlap the WiFi channels and there can be other Bluetooth devices also using the same channels.
The method analyses the channels over time and chooses those it thinks has least interference and most stable RSSI. This reduces the positioning error by 0.2m, to 2.2m, at a distance of 3.6m.
Although the implementation is similar to SensorMesh™ and BeaconRTLS™ used together, their solution uses a proprietary mesh implementation and a proprietary data protocol. Consequently, their implementation suffers longer response time when used over longer physical distances. Their maximum inter-hop distance of 8 to 10 m also isn’t good due to non-optimal devices and non-optimal device positioning.
“The Priority Matrix shows that many IoT technologies are 5 years from mainstream adoption. However only one innovation profile will reach maturity in 2 years, indoor location for assets.
So why is ‘indoor location for assets’ more likely to achieve mainstream adoption sooner than other technologies? It’s because there are clear benefits for most companies and off-the-shelf software such as our BeaconRTLS™ is already available.
Our work with companies shows they are nevertheless cautious. Companies are taking time to understand the competing asset tracking technologies and are performing, sometimes lengthy, trials to determine how new systems will integrate with existing systems. They are considering the implications of SAAS vs on-premise solutions, the availability of second-sourced beacon hardware and the compromises of accuracy vs system complexity and cost.
We sometimes get asked for location beacons or which beacons are best for determining location. All beacons can be used for locating. While there are physical aspects such as battery size/life and waterproofing that make some beacons more suitable for some scenarios, locating capability is determined more by the software used rather than the beacons themselves.
If you have been attracted to Bluetooth by recent announcements on Bluetooth direction finding, be aware that no ready-made hardware or software solutions exist yet. It will take a while, perhaps years, before silicon vendors support Bluetooth 5.1 direction finding, silicon vendors create SDKs and hardware manufacturers create hardware.
There are many industries where the inability to find assets leads to the requirement to have many more of those assets. This is especially so in areas, such as hospitals, where not finding things can cost lives.
It also tends to be the case that such urgently required items are also expensive as they are critical pieces of equipment. When equipment is very expensive, lack of redundancy can end up causing key staff spending their time finding things rather than doing their main job.
Even when not finding things isn’t mission critical, a lot of time, human effort and hence cost can be wasted if assets aren’t available. Examples include vehicles in fleet management, tools in construction and equipment in manufacturing.
Beacons and locating systems allow you to reduce asset redundancy, save costs and make working processes more efficient.
In our previous post on iBeacon Microlocation Accuracy we explained how distance can be inferred from the received signal strength indicator (RSSI). We also explained how techniques such as trilateration, calibration and angle of arrival (AoA) can be used to improve location accuracy.
Their study was based on a large-scale exhibition where they placed scanning devices:
They implemented a LSTM neural network and experimented with the number of layers:
They obtained best results with the simplest machine learning model with only 1 LSTM:
As is often the case with machine learning, more complex models over-learn on the training data such that they don’t work with new, subsequent data. Simple models are more generic and work not just with the training data but with new scenarios.
The researchers managed to achieve an accuracy of 2.44m at 75 percentile – whatever that means – we guess in 75% of the cases. 2.44m is ok and compares well to accuracies of about 1.5m within a shorter range confined space and 5m at the longer distances achieved using conventional methods. As with all machine learning, further parameter tuning usually improves the accuracy further but can take along time and effort. It’s our experience that using other types of RNN in conjunction with LSTM can also improve accuracy.
A mentioned in our post on location accuracy, two methods of improving accuracy are calibration and trilateration. There’s a recent research paper on iBeacon indoor localization using trusted-ranges model, that explores an alternative ‘trusted-ranges’ method. The method is still based on the RSSI measurements between the beacon and detector. It builds up a trusted-range model to describe how the RSSI varies over time and distance.
The model supplies reliable ranges of received signal strength values from nearest neighbours classifying received signal strength values into various levels of range. It performs better than calibration, especially at shorter ranges, while having a low complexity and hence computationally fast speed.