Optimized CNNs to Indoor Localization through BLE Sensors Using Improved PSO

Abstract

Indoor navigation has attracted commercial developers and researchers in the last few decades. The development of localization tools, methods and frameworks enables current communication services and applications to be optimized by incorporating location data. For clinical applications such as workflow analysis, Bluetooth Low Energy (BLE) beacons have been employed to map the positions of individuals in indoor environments. To map locations, certain existing methods use the received signal strength indicator (RSSI). Devices need to be configured to allow for dynamic interference patterns when using the RSSI sensors to monitor indoor positions. In this paper, our objective is to explore an alternative method for monitoring a moving user’s indoor position using BLE sensors in complex indoor building environments. We developed a Convolutional Neural Network (CNN) based positioning model based on the 2D image composed of the received number of signals indicator from both x and y-axes. In this way, like a pixel, we interact with each 10 × 10 matrix holding the spatial information of coordinates and suggest the possible shift of a sensor, adding a sensor and removing a sensor. To develop CNN we adopted a neuro-evolution approach to optimize and create several layers in the network dynamically, through enhanced Particle Swarm Optimization (PSO). For the optimization of CNN, the global best solution obtained by PSO is directly given to the weights of each layer of CNN. In addition, we employed dynamic inertia weights in the PSO, instead of a constant inertia weight, to maintain the CNN layers’ length corresponding to the RSSI signals from BLE sensors. Experiments were conducted in a building environment where thirteen beacon devices had been installed in different locations to record coordinates. For evaluation comparison, we further adopted machine learning and deep learning algorithms for predicting a user’s location in an indoor environment. The experimental results indicate that the proposed optimized CNN-based method shows high accuracy (97.92% with 2.8% error) for tracking a moving user’s locations in a complex building without complex calibration as compared to other recent methods.

An RSS Transform—Based WKNN for Indoor Positioning

“…It is time-and labor-consuming. To address this issue, researchers have done some significant work: (1) the large UJIIndoorLoc dataset covers multiple buildings [8]; (2) the IPIN2016 Tutorial dataset focuses on small scene positioning [9]; and (3) Zenodo dataset contains both long-and short-term changes [10].…”

“…However, the extensive deployment of wireless infrastructure and the proliferation of mobile devices have facilitated positioning in indoor scenes, and positioning based on received signal strength (RSS) has been an attractive solution [ 1 ]. Common wireless signals such as Bluetooth [ 2 ], Wi-Fi [ 3 ], ultra-wideband (UWB) [ 4 ], and radio frequency identification (RFID) [ 5 ] are often used for positioning. Positioning is also dependent on the existing position calculation algorithms, such as direct positioning, geometrical calculations, and fingerprint localization.…”

LightGBM Indoor Positioning Method Based on Merged Wi-Fi and Image Fingerprints

“…However, owing to the complex and variable natures of indoor environments, the large-scale application of indoor positioning solutions has yet to be achieved. Researchers have used a variety of indoor signals for positioning, including wireless local area network (WLAN) facilities widely distributed in indoor environments, cellular networks [ 2 ], Bluetooth [ 3 ], radio-frequency identification [ 4 ] and other radio frequency signals, microelectromechanical system gyroscopes [ 5 ], ultra-wideband (UWB) [ 6 , 7 ], laser ranging [ 8 ], and visual information [ 9 ]. Wi-Fi fingerprint positioning does not require the distances and angles to be known in advance; however, it is seriously affected by indoor multipath effects.…”

A Sequential Handwriting Recognition Model Based on a Dynamically Configurable CRNN

“…Based on the expert's perspective, it is more convenient to search for architecture modifications automatically, such as simple scalar hyperparameters, e.g., the regularization coefficient and learning rate. In recent years, many existing studies have applied neuroevolutionary mechanisms for automatically detecting the appropriate CNN topology and hyperparameters based on the target dataset [28][29][30][31]. Additionally, the existing automation-based CNN models have been used for character recognition.…”