Location-based service providers are continuously developing Radio Frequency (RF) based solutions to provide indoor navigation and positioning systems. Of these, BLE based systems are the most cost-effective and easily deployed, however the precision of RSSI (Recevied Signal Strength Indication) measurements are weakened by fading effects and the accuracy of the estimates are significantly affected by the distance between the receiver and transmitter. Traditional methods currently being used in the market such as trilateration and proximity are unable to meet the growing needs and expectations of LBS users. Studies have explored the possibility of enhancing accuracy and precision of indoor positioning in a secure way while decreasing costs using artificial neural networks. In fact, Boni Global has integrated an AI-based solution to its products, with which, real-time data received from the user is analyzed in a trained network model developed with fingerprinting-based artificial neural networks, eliminating fading effects to achieve high accuracy, precision and reliability.
Fingerprinting is a RSSI-based scene analysis that first collects features (fingerprints) of a surrounding, reference points at every location in the areas of interest, and then builds a fingerprint database. After that, the system estimates the location of an object by matching online measurements with the closest location fingerprints. During the offline training phase, RSSI signals are sampled at each reference point to build a database. In the online determination phase, real-time signals are received from a mobile device to calculate its most likely location. A neural network positioning model can be developed to train fingerprints instead of storing all received signal strength (RSS) data during the offline training phase.
This solution can significantly improve the performance of location-based products and applications to better serve the needs of LBS clients across various sectors.The AI based solution can achieve up to 2-3.5 metre accuracy with 98% success rate, while reducing costs by optimizing the use of BLE beacons with machine learning. Real-time positioning of people and assets will be faster and more accurate as the location information is communicated within a second.