This paper introduces a new methodology called OBLEA, which aims to optimise BLE anchor configurations in indoor settings. It takes into account various BLE variables to enhance flexibility and applicability to different environments. The method uses a data-driven approach, aiming to obtain the best configuration with as few anchors as possible.
The OBLEA method offers a flexible framework for indoor spaces where the occupants are fitted with wrist activity bracelets (beacons) and BLE anchors are set up. The anchors then collect and aggregate data, sending it to a central point (fog node) via MQTT.
A dataset was generated with the maximum number of anchors in the indoor environment, and different configurations were then trained and tested based on this dataset. The best balance between fewer anchors and high accuracy was chosen as the optimal configuration.
This methodology was tested and optimised in a real-world scenario, in a Spanish nursing home in Alcaudete, Jaén. The experiment involved seven inhabitants in four shared double rooms. As a result of this optimisation, the inhabitants could be located in real time with an accuracy of 99.82%, using a method called the K-Nearest-Neighbour algorithm and collating the signal strength (RSSIs) in 30-second time windows.
There’s new research from Scotland, UK on Calibration of a novel Bluetooth Low Energy (BLE) monitoring device in a sheep grazing environment. Knowing an animal’s location and proximity can offer insights into landscape use, animal performance, behaviour and social contacts. However, the technologies currently used to collect these data are costly and challenging to implement, particularly due to the low value of individual animals and typically large flock sizes.
A device was specifically designed for a study to assess the relationship between the Received Signal Strength Indicator (RSSI) of a BLE beacon and BLE reader and to develop a distance prediction model. This model was then applied in a static situation and on-sheep studies, using a multi-lateration approach to determine a beacon’s location within a field setting. A purpose-built Wearable Integrated Sensor Platform (WISP) was developed for the study, featuring a BLE reader and other sensors. It was designed to report the identity and RSSI of the 16 ‘closest’ beacons seen for each duty cycle.
The findings revealed that the height of the device had an impact, with fewer beacons reported at a shorter distance in WISPs at the lower height of 0.3 m. RSSI can vary greatly based on factors like transmission power, device orientation, enclosure and the operating environment.
Using the distance prediction and adjusted distance prediction, beacon locations could be estimated for most of the beacons. Not all beacons could be located due to issues such as being reported by too few WISPs or the resulting multi-lateration circles not intersecting.
The study suggests that BLE can potentially be used for sheep localisation in outdoor environments. The multi-lateration approach is dependent on receiving RSSI readings from multiple readers at a similar timepoint, it could offer more information about localisation and movement than simple proximity ranges or presence/absence. Locating a sheep to within about 30 m in a field environment represents a significant step forward.
Many people inquire about adjusting the transmission distance of a beacon. They often wish to either conserve battery or restrict the range at which a beacon is detectable.
While some third-party platforms and SDKs offer distance settings, it’s a misconception to think you can directly set the distance. What you’re actually adjusting is the transmission power, which in turn influences the transmission distance. But since this involves radio waves, which are prone to reflections and interference, it’s impossible to guarantee that a specific power will equate to a precise distance.
When using an app to detect beacons, you can employ the Received Signal Strength Indicator (RSSI) to focus on those within a desired range. However, it’s challenging to precisely correlate RSSI with the actual distance.
Some wonder if they can set the distance in terms of centimetres, similar to NFC. Typically, this isn’t feasible because even at their lowest power setting, most beacons transmit over a distance of about a metre.
Rather than asking if the transmitter’s distance can be minimised, it might be more practical to configure the receiver to disregard detections from further away. By using the RSSI value on the receiving app or another Bluetooth scanning device, you can filter out distant beacons. Specifically, you can dismiss detections with an RSSI below a certain threshold, allowing you to focus on detections within a centimetre range.
Most motion sensing applications usually place a sensor beacon on the things that will move. The accelerometer in the beacon reports movement. The research paper describes an alternative method of detecting movement of a person, an elderly person in this case, based on the change in blocking of the beacon signal over time. This has the advantage that the beacon doesn’t need to be worn. Also, it doesn’t have to be a accelerometer beacon as any beacon can be used.
The problem with using the strength of the beacon signal (RSSI), is that it varies over time even when there’s no change of blocking in the room. This is due to radio frequency (RF) noise and reflection. The authors of the paper looked into smoothing of the data to filter out such variance in the data:
The report concludes that when averaging over three or more RSSI values, it’s possible to minimise the RF variance and reliably detect the variance caused by human movement in the room.
Another, more reliable, way of detecting movement is to use a beacon with built-in PIR such as the iBS02PIR, M52-PIR, IX32 or MSP01.
Researchers from Idiap Research Institute and EPFL, Switzerland have been looking into the use of smartphone data, include Bluetooth metadata to try to infer social context, for example whether someone is alone or not.
The paper Understanding Social Context from Smartphone Sensing: Generalization Across Countries and Daily Life Moments (pdf) attempts to better understand human behaviour and mental well-being. The paper focuses on the use of passive smartphone sensors, including Bluetooth, to track the social context of individuals over time. In the past, this field of research has been limited by the fact that most studies have only been conducted in one or two countries and often focused on specific contexts such as eating or drinking.
This paper aims to overcome these limitations by using a new, extensive and multimodal smartphone sensing dataset that includes over 216,000 self-reports from more than 580 participants across five different countries – Mongolia, Italy, Denmark, the UK and Paraguay. The goal is to explore the feasibility of using sensor data to infer whether a person is alone or not and to examine how behavioural and country-level diversity influences this inference.
The sensor data comes from 34 different sensors, divided into continuous and interaction sensing modalities. Continuous sensing includes types of activity, step count, Bluetooth, WiFi, location, cellular, and proximity data, while interaction sensing involves app usage, touch events, screen on/off episodes, and notifications. In terms of Bluetooth, the study used both normal and low-energy Bluetooth capturing data on the number of connected devices and received signal strength indicators (RSSI).
The study’s key findings suggest that sensor features can be used to infer the social context. The research also found that models partially personalised to multi-country and country-specific data achieved similar accuracy levels, typically ranging from 80% to 90%. However, the models did not generalise well to unseen countries regardless of geographic similarity.
There’s new research from the University of Illinois titled Packet Reception Probability: Packets That You Can’t Decode Can Help Keep You Safe (pdf). Many existing systems estimate distance using the Receiver Signal Strength Indicator (RSSI) which is negatively impacted by sampling bias and multipath effects. As an alternative, the study uses Packet Reception Probability (PRP) that utilises packet loss to estimate distance.
Localisation is achieved through a Bayesian-PRP approach that also includes an explicit model of multipath. To facilitate straightforward deployment, there’s no need for any modifications to hardware, firmware, or driver-level on standard devices and only minimal training is required.
A variety of devices were used including Bluvision iBeeks, BluFi, a Texas Instrument Packet Sniffer, a laptop, and Android smartphones (Nexus5x). 60 iBeacons were deployed in a library and 38 in a retail store. The Texas Instrument Packet Sniffer, connected to a Windows laptop was used for packet reception from beacons. Android phones were equipped with a purpose-built Android app.
PRP was found to provide metre-level accuracy with just six devices in known locations and 12 training locations. Combining PRP with RSSI was found to be beneficial at short distances up to 2m. Beyond distances of 2m, fusing the two is less effective than using PRP alone because RSSI becomes de-correlated with distance.
There’s new research in the Bulletin of Electrical Engineering and Informatics on Bluetooth beacons based indoor positioning in a shopping malls using machine learning. Researchers from Algeria and Italy improved the accuracy of RSSI locating by using AI machine learning techniques. They used extra-trees classifier (ETC) and a k-neighbours classifier to achieve greater than 90% accuracy.
A smartphone app was used to receive beacon RSSI and send it to an indoor positioning system’s data collection module. RSSI data was also filtered by a data processing module to limit the error range. KNN, RFC, extra trees classifiers (ETC), SVM, gradient boosting classifiers (GBC) and decision trees (DT) algorithms were evaluated.
The ETC model gave the best accuracy. ETC is an algorithm that uses a group of decision trees to classify data. It is similar to a random forest classifier but uses a different method to construct the decision trees. ETC fits a number of randomised decision trees on sub-samples of the dataset and uses averaging to improve the predictive accuracy and control over-fitting. ETC is a good choice for applications where accuracy is important but the data is noisy and where computational efficiency is important.
RSSI stands for Received Signal Strength Indicator. It is a measure of the power level of a radio signal being received by a device, for example a smartphone, in dBm (decibel-milliwatts). The RSSI is accessible to receiving devices via APIs such as the standard iOS and Android Bluetooth libraries.
The RSSI value is typically used to get an indication of the distance between a device and a beacon. A higher RSSI value indicates a stronger signal and therefore a closer proximity to the beacon, while a lower RSSI value indicates a weaker signal and a farther proximity to the beacon. Note that RSSI is usually -ve so a larger negative more usually indicates the beacon is further away.
RSSI is not a perfect measure of distance, as it can be affected by factors such as the environment and the type of device that is receiving the signal. However, by comparing the RSSI value of a beacon’s signal with the known transmission power of the beacon, it is possible to estimate the distance between the device and the beacon.
RSSI is commonly used in wireless communications such as WiFi, Zigbee, Bluetooth and cellular networks to measure the signal strength of the received signal. It is also used to estimate the quality of the signal, and to determine if the signal is strong enough to maintain a reliable connection.
RSSI is not a standard or a regulated measure and varys depending on the technology and the manufacturer of the device.
The relationship between RSSI and distance is not linear, and can vary depending on the environment and the type of device that is receiving the signal. In general, as the distance between a device and a beacon increases, the RSSI value decreases. However, the rate at which the RSSI value decreases with distance can vary depending on factors such as the environment and the transmission power of the beacon.
In free space, the RSSI value decreases at a rate of approximately 6 dB per doubling of distance. This is known as the inverse square law, which states that the power of a signal decreases proportionally to the square of the distance from the source.
Inverse square law
However, in a real-world environment, the rate of decrease can be affected by factors such as walls, obstacles, and interference from other devices, which can cause the signal to weaken faster or slower than expected.
It’s also worth noting that the RSSI value can vary depending on the type of device that is receiving the signal, as well as the type of radio technology used. The sensitivity of the device’s radio receiver will also affect the received RSSI value, a more sensitive device will be able to detect weaker signals at farther distances than a less sensitive device.
While equations can be used to infer distance from RSSI, the above factors mean the most accurate way to determine distance is to compare with previously measured RSSI-distance values.
If accurate distance is essential, up to about 3m, consider using a beacon such as the iBS03R that uses a time of flight (ToF) sensor rather than using RSSI.
Bluetooth beacons advertising iBeacon can be used to perform indoor locating using trilateration. Trilateration is where three receivers are used to measure signal strength (RSSI) to calculate the position.
The implementation uses MQTT to send the data to a React app on a server where it’s displayed on a floorplan.
In practice, you might want to consider creating a more robust solution that uses Bluetooth gateways rather than ESP32 devices. There’s also the Bluetooth AoA Direction Finding standard that’s more accurate than using RSSI.
Machine learning was used to determine the line-of-sight distance in a multipath (reflective) environment. Due to the multipath effect, acquired signals indoors have complex mathematical models. A machine learning Artificial Neural Networks (ANN) is the most efficient way to process these signals.
The system achieved accuracy where 75% of the errors were less than 0.1 m with a median error of 0.037 m and a mean error of 0.092 m. This reduced ranging errors to under 10cm. The researchers were able to achieve high-precision indoor ranging without the need for a wide signal bandwidth nor synchronisation. The system was also simple and low cost to deploy due to low complexity of the equipment.
