A Machine-Learning-Based Access Point Selection Strategy for Automated Guided Vehicles in Smart Factories

Automated Guided Vehicles (AGVs) are becoming popular at many manufacturing facilities. To ensure mobility and flexibility, AGVs are often controlled by wireless communication, eliminating the constraints of physical cables. These AGVs require multiple Access Points (APs) to ensure uninterrupted coverage across the site. As AGVs move, they need to switch between these APs seamlessly. A primary challenge is that the communication downtime during this link-switching process must be minimal for effective AGV monitoring and control. Current AP selection strategies based on observed Received Signal Strength Indicator (RSSI) often fail in manufacturing environments due to RSSI's inherent instability. This paper introduces a new AP selection technique for AGVs navigating these sites. Our approach harnesses the distinct movement patterns of AGVs and uses machine learning techniques to learn location-, trajectory-, and orientation-specific RSSI from the APs. Real-world factory data from our unique dataset revealed that our method extends the potential communication duration per route by 1.34 times compared to the prevalent signal strength-based switching methods commonly implemented in current drivers provided by chipset vendors or open-source Wi-Fi drivers. These results indicate that the automatic evaluation and tuning of the wireless environment using the proposed method is beneficial in reducing the time and effort required to investigate the detailed propagation paths needed to adapt AGV to existing APs.

A Novel Optimized iBeacon Localization Algorithm Modeling

The conventional methods for indoor localization rely on technologies such as RADAR, ultrasonic, laser range localization, beacon technology, and others. Developers in the industry have started utilizing these localization techniques in iBeacon systems that use Bluetooth sensors to measure the object's location. The iBeacon-based system is appealing due to its low cost, ease of setup, signaling, and maintenance; however, with current technology, it is challenging to achieve high accuracy in indoor object localization or tracking. Furthermore, iBeacons' accuracy is unsatisfactory, and they are vulnerable to other radio signal interference and environmental noise. In order to address those challenges, our study focuses on the development of error modeling algorithms for signal calibration, uncertainty reduction, and interfered noise elimination. The new error modeling is developed on the Curve Fitted Kalman Filter (CFKF) algorithms. The reliability, accuracy, and feasibility of the CFKF algorithms are tested in the experiments. The results significantly show the improvement of the accuracy and precision with this novel approach for iBeacon localization.

Q-Learning-Based Pending Zone Adjustment for Proximity Classification

This paper presents a Q-learning-based pending zone adjustment for received signal strength indicator (RSSI)-based proximity classification (QPZA). QPZA aims to improve the accuracy of RSSI-based proximity classification by adaptively adjusting the size of the pending zone, taking into account changes in the surrounding environment. The pending zone refers to an area in which the previous result of proximity classification is maintained and is expressed as a near boundary and a far boundary. QPZA uses Q-learning to expand the size of the pending zone when the noise level increases and reduce it otherwise. Specifically, it calculates the noise level using the estimation error of a device deployed at a specific location. Then, QPZA adjusts the near boundary and far boundary separately by inputting the noise level into the near and far boundary adjusters, consisting of the Q-learning agent and reward calculator. The Q-learning agent determines the next boundary using the Q-table, and the reward calculator calculates the reward using the noise level. QPZA updates the Q-table of the Q-learning agent using the reward. To evaluate the performance of QPZA, we conducted an experimental implementation and compared the accuracy of QPZA with that of the existing approach. The results showed that QPZA achieves 11.69% higher accuracy compared to the existing approach, on average.

A Wi-Fi-Based Passive Indoor Positioning System via Entropy-Enhanced Deployment of Wi-Fi Sniffers

This study presents a Wi-Fi-based passive indoor positioning system (IPS) that does not require active collaboration from the user or additional interfaces on the device-under-test (DUT). To maximise the accuracy of the IPS, the optimal deployment of Wi-Fi Sniffers in the area of interest is crucial. A modified Genetic Algorithm (GA) with an entropy-enhanced objective function is proposed to optimize the deployment. These Wi-Fi Sniffers are used to scan and collect the DUT's Wi-Fi received signal strength indicators (RSSIs) as Wi-Fi fingerprints, which are then mapped to reference points (RPs) in the physical world. The positioning algorithm utilises a weighted k-nearest neighbourhood (WKNN) method. Automated data collection of RSSI on each RP is achieved using a surveying robot for the Wi-Fi 2.4 GHz and 5 GHz bands. The preliminary results show that using only 20 Wi-Fi Sniffers as features for model training, the offline positioning accuracy is 2.2 m in terms of root mean squared error (RMSE). A proof-of-concept real-time online passive IPS is implemented to show that it is possible to detect the online presence of DUTs and obtain their RSSIs as online fingerprints to estimate their position.

An Experimental Assessment of People's Location Efficiency Using Low-Energy Communications-Based Movement Tracking

(1) Background: public transport demand dynamics represents important information for fleet managers and is also a key factor in making public transport attractive to reduce the environmental footprint of urban traffic. This research presents some experimental results on the assessment of low-energy communication technologies, such as Wi-Fi and Bluetooth, as support for people density and/or movement tracking sensing technologies. (2) Methods: the research is based on field measurements to determine the percentage of discoverable devices carried by people, in relation to the total number of physical persons in interest, different scenarios of mobile devices usage and evaluation of influences on radio signals' propagation, RSSI / RX read values, and efficiency of indoor localization, or in similar GPS-denied environments. Different situations are investigated, especially public transport-related ones, such as subway stations, indoors of commuting hubs, railway stations and trains. (3) Results: diagrams and experiments are presented, and models of signal behavior are also proposed. (4) Conclusions: recommendations on the efficiency of these non-conventional traveler and passenger flow tracking solutions and models are presented at the end of the paper.

An Enhanced Indoor Positioning Algorithm Based on Fingerprint Using Fine-Grained CSI and RSSI Measurements of IEEE 802.11n WLAN

Received signal strength indication (RSSI) obtained by Medium Access Control (MAC) layer is widely used in range-based and fingerprint location systems due to its low cost and low complexity. However, RSS is affected by noise signals and multi-path, and its positioning performance is not stable. In recent years, many commercial WiFi devices support the acquisition of physical layer channel state information (CSI). CSI is an index that can characterize the signal characteristics with more fine granularity than RSS. Compared with RSS, CSI can avoid the effects of multi-path and noise by analyzing the characteristics of multi-channel sub-carriers. To improve the indoor location accuracy and algorithm efficiency, this paper proposes a hybrid fingerprint location technology based on RSS and CSI. In the off-line phase, to overcome the problems of low positioning accuracy and fingerprint drift caused by signal instability, a methodology based on the Kalman filter and a Gaussian function is proposed to preprocess the RSSI value and CSI amplitude value, and the improved CSI phase is incorporated after the linear transformation. The mutation and noisy data are then effectively eliminated, and the accurate and smoother outputs of the RSSI and CSI values can be achieved. Then, the accurate hybrid fingerprint database is established after dimensionality reduction of the obtained high-dimensional data values. The weighted k-nearest neighbor (WKNN) algorithm is applied to reduce the complexity of the algorithm during the online positioning stage, and the accurate indoor positioning algorithm is accomplished. Experimental results show that the proposed algorithm exhibits good performance on anti-noise ability, fusion positioning accuracy, and real-time filtering. Compared with CSI-MIMO, FIFS, and RSSI-based methods, the proposed fusion correction method has higher positioning accuracy and smaller positioning error.

Developing a Solution for Mobility and Distribution Analysis Based on Bluetooth and Artificial Intelligence

The purpose of this research was to develop a simple, cost-effective, but enough efficient solution for locating, tracking and distribution analysis of people and/or vehicle flowing, based on non-intrusive Bluetooth sensing and selective filtering algorithms employing artificial intelligence components. The solution provides a tool for analyzing density of targets in a specific area, useful when checking contact proximities of a target along a route. The principle consists of the detection of mobile devices that use active Bluetooth connections, such as personal notebooks, smartphones, smartwatches, Bluetooth headphones, etc. to locate and track their movement in the dedicated area. For this purpose, a specific configuration of three BT sensors is used and RSSI levels compared, based on a combination of differential location estimates. The solution may also be suited for indoor localization where GPS signals are usually weak or missing; for example, in public places such as subway stations or trains, hospitals, airport terminals and so on. The applicability of this solution is estimated to be vast, ranging from travel and transport information services, route guidance, passenger flows tracking, and path recovery for persons suspected to have SARS-COV2 or other contagious viruses, serving epidemiologic enquiries. The specific configuration of Bluetooth detectors may be installed either in a fixed location, or in a public transport vehicle. A set of filters and algorithms for triangulation-based location of detected targets and movement tracking, based on artificial intelligence is employed. When applied in the public transport field, this setup can be also developed to extract additional information on traffic, such as private traffic flowing, or passenger movement patterns along the vehicle route, improved location in absence of GPS signals, etc. Field tests have been carried out for determining different aspects concerning indoor location accuracy, reliability, selection of targets and filtering. Results and possible applications are also presented in the final section of the paper.

Environmental Impacts on Hardware-Based Link Quality Estimators in Wireless Sensor Networks

Hardware-based link quality estimators (LQEs) in wireless sensor networks generally use physical layer parameters to estimate packet reception ratio, which has advantages of high agility and low overhead. However, many existing studies didn’t consider the impacts of environmental changes on the applicability of these estimators. This paper compares the performance of typical hardware-based LQEs in different environments. Meanwhile, aiming at the problematic Signal-to-Noise Ratio (SNR) calculation used in existing studies, a more reasonable calculation method is proposed. The results show that it is not accurate to estimate the packet reception rate using the communication distance, and it may be useless when the environment changes. Meanwhile, the fluctuation range of the Received Signal Strength Indicator (RSSI) and SNR will be affected and that of Link Quality Indicator (LQI) is almost unchanged. The performance of RSSI based LQEs may degrade when the environment changes. Fortunately, this degradation is mainly caused by the change of background noise, which could be compensated conveniently. The best environmental adaptability is gained by LQI and SNR based LQEs, as they are almost unaffected when the environment changes. Moreover, LQI based LQEs are more accurate than SNR based ones in the transitional region. Nevertheless, compared with SNR, the fluctuation range of LQI is much larger, which needs a larger smoothing window to converge. In addition, the calculation of LQI is typically vendor-specific. Therefore, the tradeoff between accuracy, agility, and convenience should be considered in practice.

A Low-Cost On-Street Parking Management System Based on Bluetooth Beacons

In recent years, many city governments around the world have begun to use information and communication technology to increase the management efficiency of on-street parking. Among various experimental smart parking projects, deployment of wireless magnetic sensors and smart parking meters are quite common. However, using wireless magnetic sensors can only detect the occupancy of parking spaces without the knowledge of who are currently using these parking spaces; human labor is still needed to issue the parking bills. In contrast, smart parking meters based on image recognition can detect the occupancy of parking spaces along with the license plate numbers, but the cost of deploying smart parking meters is relatively high. In this research, we investigate the feasibility of building an on-street parking management system mainly based on low-cost Bluetooth beacons. Specifically, beacon transmitters are installed in the vehicles, and beacon receivers are deployed along the roadside parking spaces. By processing the received beacon signals using Kalman filter, our system can detect the occupancy of parking spaces as well as the identification of the vehicles. Although distance estimation using the received signal strength is not accurate, our experiments show that it suffices for correct detection of parking occupancy.

Performance, Accuracy and Generalization Capability of RFID Tags' Constellation for Indoor Localization

Indoor localization has recently witnessed an increase in interest due to its wide range of potential services. Further, the location information is very important in many applications, such as the Internet of Things, logistics, library management and so on. Hence, different technologies and techniques have been proposed in the literature for indoor localization systems. Most of these systems present the disadvantages of a poor performance, low accuracy and high cost. However, thanks to its low cost, high accuracy and non-line-of-sight detection, radio frequency identification (RFID)-based localization has increasingly become the most used technology for indoor localization. In this paper, we propose an innovative approach based on the multiple input single output (MISO) protocol to improve the accuracy of a low-cost RFID localization system. Whereas most traditional systems use a single tag for localization, the proposed architecture encourages the use of a group of RFID tags named as a constellation. According to experimental results and based on the signals’ diversity, the location accuracy is improved to get an estimated position error of 81 cm at the cumulative distribution function of 90%.

A Self Regulating and Crowdsourced Indoor Positioning System through Wi-Fi Fingerprinting for Multi Storey Building

Unobtrusive indoor location systems must rely on methods that avoid the deployment of large hardware infrastructures or require information owned by network administrators. Fingerprinting methods can work under these circumstances by comparing the real-time received RSSI values of a smartphone coming from existing Wi-Fi access points with a previous database of stored values with known locations. Under the fingerprinting approach, conventional methods suffer from large indoor scenarios since the number of fingerprints grows with the localization area. To that aim, fingerprinting-based localization systems require fast machine learning algorithms that reduce the computational complexity when comparing real-time and stored values. In this paper, popular machine learning (ML) algorithms have been implemented for the classification of real time RSSI values to predict the user location and propose an intelligent indoor positioning system (I-IPS). The proposed I-IPS has been integrated with multi-agent framework for betterment of context-aware service (CAS). The obtained results have been analyzed and validated through established statistical measurements and superior performance achieved.