Category: Location

Solving Indoor-Outdoor Transitions

New research titled Seamless Indoor–Outdoor Localization Through Transition Detection by Jaehyun Yoo presents a system for maintaining accurate and continuous positioning of a person or object as they move between indoor and outdoor environments. Standard GPS technologies struggle in indoor areas due to signal loss, while indoor systems often fail outdoors. Transition zones, such as building entrances, pose particular challenges. This paper addresses those issues by introducing a transition detection algorithm that combines data from WiFi and BLE signals, GPS metrics, and inertial sensors.

The core of the proposed system is a handover strategy that classifies an environment as indoor, transition, or outdoor using a probabilistic model based on signal strength (RSSI), position estimation, and satellite signal quality. The classification process relies on machine learning, specifically a neural network model trained on unlabelled data using an unsupervised approach. The transition zones are especially crucial for switching accurately between positioning engines.

The system comprises three separate engines. The indoor engine uses WiFi and BLE fingerprinting fused with inertial sensor data through a particle filter. The transition engine works exclusively with AI-augmented inertial data to maintain tracking where both WiFi and GNSS are unreliable. The outdoor engine integrates GNSS data with inertial measurements using an extended Kalman Filter, adapting dynamically to account for signal reliability and expected movement.

Experimental tests were conducted using three Samsung smartphones in three office buildings of various sizes and layouts. The new method demonstrated improved positional accuracy and better state classification compared to Google’s Fused Location Provider (FLP) and standard GPS. Results showed the proposed system consistently outperformed the alternatives, particularly in recognising transition zones and reducing localisation errors. However, the method was sensitive to direction errors and GNSS multi-path interference, which could affect the overall accuracy, particularly in densely built environments.

The study concludes that the approach offers significant improvements in seamless localisation using consumer smartphones without requiring labelled data or external GNSS hardware. Future work aims to replace manually tuned parameters with data-driven methods and to further address GNSS multi-path errors by leveraging raw satellite data.

Location in Underground Mines

New research titled Reliable Proximity Sensing for Underground Mining by Johnny Lam explores the challenge of accurately tracking personnel and vehicles in underground mines using digital positioning systems. While vehicle tags use onboard sensors like accelerometers and gyroscopes for precise positioning, personal tags rely on less accurate signal strength readings from fixed reference points. This discrepancy can lead to safety issues, especially when personnel are inside vehicles but their tags report separate positions.

To resolve this, the study investigates a method for personal tags to inherit the vehicle’s position by first determining if they are actually inside the same vehicle. The approach uses Bluetooth Low Energy (BLE) to detect nearby devices, then samples and compares accelerometer data from both the personal and vehicle tags. Using dynamic time warping (DTW) and variance analysis of the motion data, the system classifies whether a personal tag is inside the vehicle.

The system was implemented as an Android application and tested under various conditions. BLE proved effective for detecting proximity, while combining DTW and variance analysis yielded a classification accuracy of about 90%. Power consumption increased by roughly 22% with full functionality.

Obtaining Distance from RSSI

RSSI is the signal strength at the Bluetooth receiver. The signal type, for example, iBeacon, Eddystone or sensor beacon is irrelevant. The value of the RSSI can be used to infer distance.

The accuracy of the distance measurement depends on many factors such as the type of sending device used, the output power, the capability of the receiving device, obstacles and importantly the distance of the beacon from the receiving device.

The output power isn’t known to the receiver so it’s sometimes added to the advertising data in the form of the ‘measured power’ which is the power at 1m from the sender.

The closer the beacon is to the receiver, the more accurate the derived distance. As our article mentions, projects that get more detailed location derived from RSSI, usually via trilateration and weighted averages, usually achieve accuracies of about 5m within the full range of the beacon or 1.5m within a shorter range confined space.

Some beacons, mainly those with output RF amplifiers, transmit more than 50m and in these cases, while the beacon can be detected, using RSSI to infer distance isn’t usually reliable due to noise (variation) in the RSSI value.

There’s some Android Java code on GitHub if you want to experiment with extracting distance from RSSI. There’s an equation for iOS on GitHub.

Need more help? Consider a Feasibility Study.

Beacons that flash/vibrate at a given distance.

Continuous Transmission in Asset Tracking

Continuous transmission is a fundamental distinction between Bluetooth-based systems and traditional RFID or barcode scanning systems in asset tracking solutions. Unlike RFID or barcode methods, which rely on individual scans that can easily be missed, especially when performed manually, Bluetooth beacons emit signals at regular intervals. This regularity means that updates on an asset’s presence and location are ongoing, rather than sporadic. It also enables more accurate reporting, as the absence of a Bluetooth signal can reliably indicate that an asset has genuinely gone missing.

In reality, beacons do not transmit continuously in the literal sense. Instead, they emit signals periodically, with intervals ranging from a few hundred milliseconds to several seconds or even minutes. The frequency of these transmissions affects both power consumption and the immediacy of data in the reporting system. More frequent transmissions provide more up-to-date information but consume more power, which can be a critical consideration in battery-operated devices.

To balance power usage and data accuracy, some simpler systems opt for a transmission interval of around ten seconds. Additionally, certain advanced beacons are equipped with motion sensors that trigger extra updates when movement is detected, thereby providing more dynamic and context-aware tracking.

Mobile Device Power Saving Proximity Beacons

New research (pdf) introduces a method to reduce battery consumption in mobile devices, particularly small, wearable computing devices like smartwatches, rings, or glasses, by using proximity beacons instead of traditional satellite-based location tracking.

Typically, these devices use GNSS (Global Navigation Satellite Systems) such as GPS, which consume significant power due to the effort required to detect and process weak satellite signals. This is especially problematic for small devices with limited battery capacity. The proposed method embeds low-power proximity beacons, such as Bluetooth trackers (e.g., AirTags), into wearable components like watch bands. These beacons emit Bluetooth signals detected by nearby devices, which can then use their own GNSS services to triangulate the location and report it via existing networks like Apple’s Find My or Android’s equivalent.

The system allows dynamic association of the mobile device with new beacons. If a user switches out a watch band, for instance, the device can recognise which beacons are moving with it and prompt the user to reassign the new beacons for location tracking. This flexibility helps maintain tracking without user intervention in most cases.

In scenarios where beacons or networks are unavailable, such as remote locations, the device may still fall back on traditional GNSS. However, in populated areas with ample nearby devices, the reliance on GNSS can be minimised, significantly conserving battery life.

The document also outlines user control features, allowing manual pairing or unpairing of beacons, enabling/disabling location tracking based on time, place, or battery level, and even letting a device query a location network for its own location. This approach not only reduces power consumption but also extends tracking functionality even when the device’s battery is depleted, as the beacons can continue to transmit signals.

The technique provides an efficient and user-friendly alternative to power-intensive GNSS location services, enhancing practicality and longevity for small wearable devices while retaining robust location tracking capabilities.

Auto-Adjusting Location Algorithm

A new study uses an indoor localisation system that integrates Bluetooth Low Energy (BLE) with an Internet of Things (IoT) framework to improve accuracy in tracking individuals, particularly those with cognitive impairments such as Alzheimer’s and dementia. The system employs an auto-adjusting algorithm that dynamically optimises received signal strength indicator (RSSI) coefficients based on real-time environmental factors, leading to improved location estimation precision.

Existing systems relying on RSSI often suffer from inaccuracies due to environmental interferences, signal fluctuations, and the use of static coefficient assignments. To address these challenges, this study develops an auto-adjusting algorithm that dynamically selects coefficients based on RSSI classifications.


The system consists of a central unit, a Raspberry Pi, and BLE peripheral nodes that communicate wirelessly. It collects real-time RSSI data and applies a path loss model to estimate distances. A web interface was developed to facilitate real-time tracking and data visualisation. The system was tested in a healthcare environment with five rooms, comparing the performance of fixed coefficient models against the proposed dynamic approach.

The experimental results showed that using fixed coefficients in distance estimation led to an initial error of 28.03%. By implementing the auto-adjusting algorithm, the error was reduced to 8%, while the maximum localisation error was decreased to 2.01 meters. Additionally, the system demonstrated high energy efficiency, with BLE peripherals operating for approximately 499 hours on a standard 230 mAh battery, reinforcing its suitability for IoT applications.

One of the main advantages of the auto-adjusting algorithm is its ability to dynamically adapt parameters. The system adjusts the path loss exponent (n) and reference signal strength (A) based on real-time RSSI classifications, improving accuracy significantly. This approach minimises localisation inaccuracies by continuously recalibrating signal strength values. The system is also energy-efficient, making it ideal for continuous tracking in various environments. Additionally, it is scalable and can be integrated with other indoor positioning systems such as Ultra-Wideband (UWB) and Wi-Fi.

The system achieves higher accuracy, maintaining a maximum error of only 2.01 meters compared to fixed coefficient models. Additionally, the BLE-based approach ensures long battery life and cost-effectiveness, making it suitable for healthcare and security applications. Compared to previous studies, the proposed algorithm proved more reliable for positioning in real-world environments.

Using Support Vector Regression (SVR) with Beacons

A new study (pdf) explores optimising Bluetooth Low Energy (BLE) beacon-based indoor positioning systems using support vector regression (SVR). It addresses the challenge of accurately identifying building occupants’ locations in real time, a critical requirement for applications such as emergency evacuations and asset tracking. Traditional methods, including trilateration and RSSI-based techniques, can face limitations like signal interference and non-line-of-sight issues.

The research adopts a fingerprinting method that uses pre-trained SVR models to improve positioning accuracy. BLE beacons, which are cost-effective and energy-efficient, were deployed across a controlled environment, and extensive RSSI data was collected and pre-processed. The model’s hyperparameters were fine-tuned to achieve optimal performance. Experimental results demonstrated a significant improvement in accuracy, with the lowest root mean squared error (RMSE) recorded as 0.9168 feet.

The findings underscore the potential of machine learning, particularly SVR, in enhancing the reliability of indoor positioning systems. This study provides a benchmark for future research, highlighting its practical applications in emergency scenarios and the advantages of BLE technology in such implementations.

Improving Bluetooth Location Accuracy

New research focuses on enhancing indoor localisation using Bluetooth Low Energy (BLE) technology by addressing challenges in signal instability and noise. The authors propose a system combining the Kalman filter for signal smoothing and deep learning models, specifically Autoencoders and Convolutional Autoencoders, for feature extraction from Received Signal Strength Indicator (RSSI) data. The method uses a fingerprinting approach, collecting data in two phases, offline for creating a reference database and online for matching new measurements to predict locations.

The study demonstrates that integrating the Kalman filter with the Convolutional Autoencoder model yields an average localisation error of 0.98 metres, significantly improving accuracy. Experimental comparisons with existing methods highlight the proposed system’s effectiveness in balancing cost, energy efficiency, and precision. The findings suggest this approach as a robust solution for indoor localisation in environments requiring high accuracy and low energy consumption.

UWB vs Bluetooth Beacons

Ultra-Wideband (UWB) technology has recently emerged as a contender to Bluetooth beacons, with some companies traditionally focused on Bluetooth now marketing UWB as the next generation solution. But does UWB live up to the promise?

UWB undeniably offers a key advantage: more accurate location tracking. With its ability to determine positions down to tens of centimetres, it surpasses Bluetooth in precision. However, this comes with significant trade-offs that should be carefully considered before adopting the technology.

One of the critical drawbacks of UWB is the lack of standardisation. Unlike Bluetooth, which operates on a well-defined and widely supported Bluetooth LE standard, UWB devices are proprietary. This means users are locked into a single vendor’s ecosystem, unable to mix and match devices from different suppliers. If the chosen vendor’s devices become obsolete, the entire solution becomes redundant, forcing costly upgrades or a complete overhaul.

The lack of standardisation also impacts the broader ecosystem. Bluetooth devices benefit from a vibrant market with multi-vendor compatibility, driving competition and keeping costs low. In contrast, UWB solutions rely on custom protocols, devices, and specialist skills, leading to higher costs and limited interoperability. While Bluetooth beacons have a range of up to 50 metres, and even 200 metres or more for certain devices, UWB typically operates within a range of 30 to 40 metres. Some advanced Bluetooth devices can even reach up to 1 kilometre, providing greater flexibility in many applications.

Power consumption is another area where Bluetooth outshines UWB. Bluetooth beacons are designed to operate efficiently, often lasting months or even years on a single battery. UWB devices, on the other hand, are more power-hungry, typically lasting only weeks in positioning applications. This makes them less practical for long-term deployments, especially in IoT scenarios where low maintenance is a priority.

Scalability is a growing concern with UWB. The technology generates and needs to process more data than Bluetooth, which can lead to bottlenecks and reduced performance as the network expands. This poses challenges for large-scale deployments, where simplicity and efficiency are critical.

Moreover, UWB’s compatibility is limited when compared to Bluetooth’s universal presence. UWB devices are primarily detected by iOS devices, with limited support on Android. This constrains their usability in a diverse market. Bluetooth, in contrast, is supported by virtually every modern smartphone and a large number of third party gateways, making it a more versatile choice.

Bluetooth beacons also offer greater functionality beyond location tracking. They can perform various sensing tasks, such as monitoring temperature, humidity, air pressure, light levels, and even detecting smoke, water leaks, or proximity. UWB, being narrowly focused on location tracking, lacks this flexibility, limiting its utility in IoT applications.

Ultimately, the decision between UWB and Bluetooth depends on your specific needs. If you require extremely precise location tracking within a limited range and can accommodate the higher costs and proprietary nature of UWB, it may be worth considering. However, for most use cases, Bluetooth remains the more efficient, flexible and cost-effective option. Its standardisation, broad compatibility, and multi-functional capabilities make it a reliable choice for tracking and IoT applications alike.

Framework for Evaluating Indoor Tracking Systems

There’s new research outlining the use of the MobiXIM framework for developing, evaluating, and refining indoor tracking systems (ITS), addressing challenges related to the lack of standardisation in the field. Indoor tracking, necessary where GPS is ineffective, relies on methods such as infrastructure-based (e.g., Bluetooth beacons using Received Signal Strength Indication), infrastructure-less (inertial and magnetic sensors) and collaborative systems (peer-to-peer communication between devices). These approaches encounter issues like accuracy, reproducibility and data collection costs.

MobiXIM integrates tools to streamline the ITS creation process, incorporating a mobile app for data collection and a web-based orchestrator platform. It employs Bluetooth Low Energy (BLE) iBeacons, both physical and virtual, to enhance location estimates. Physical iBeacons are commercial devices broadcasting signals detectable by smartphones, while virtual iBeacons simulate these signals for testing scenarios without physical deployment. The signals allow devices to calculate their proximity to a beacon, correcting their location estimates based on signal strength.

The framework’s plugin-based architecture promotes modularity, enabling researchers to mix and match existing algorithms. The methodology includes filtering noise from sensor data, positioning via algorithms like Pedestrian Dead Reckoning, and correcting errors through collaborative adjustments among devices and beacon signals. The corrected data is evaluated using metrics such as positioning accuracy and trajectory similarity.

Experiments in a university building demonstrated how collaboration between devices and interaction with beacons significantly improved accuracy. The replay feature of MobiXIM allows researchers to simulate and adjust experimental setups, testing variables like beacon density and device collaboration.

iBeacons play a critical role by providing a reliable reference point for error correction and enhancing the overall accuracy of indoor positioning systems, particularly when combined with collaborative algorithms.