Pampa Sadhukhan

Precise step length estimation (SLE) is a key necessity for not only navigation systems design but also gait health monitoring in neurological conditions. Among existing solutions, non-invasive inertial sensor-based approaches operating without dedicated infrastructure are more cost-effective. Many such methods, however, rely on bodily-affixed inertial sensors rather than freely held smartphone sensors. Traditional signal processing approaches, on the other hand, offer varying accuracy across diverse gait patterns due to user parameter calibration. This study, thus, proposes a regression-based SLE framework employing eight regression algorithms: linear regression (LR), k-nearest neighbours, support vector machine, decision tree, elastic network, random forest, histogram-based gradient boosting (HGB) regressor, and artificial neural network (ANN). Their extensive and rigorous evaluations across varied window sizes, using a dataset collected in normal and fast walking modes with two device positions (hand-held and trouser-pocket) during three evaluation scenarios, demonstrate the ANN’s exceptional generalization capacity over the other models and the previous method IRT-SD-SLE in unseen test evaluations with an achieved mean absolute error (MAE) not exceeding 6.3 cm. In seen test evaluations and leave-one-out cross-validation, the HGB regressor’s outstanding performance, achieving the lowest MAE below 1 cm across most contexts, is also reported. The extensive evaluations of training and testing times reveal the highest computational efficiency for LR, moderate efficiency for the HGB regressor, and the highest training cost for the ANN, indicating a clear trade-off between MAE and computational expense. Additionally, this study includes an insightful discussion on the performance results, including the trade-offs between accuracy and efficiency.

Precise estimation of step length constitutes a fundamental requirement in contemporary gait analysis, particularly for applications in healthcare monitoring, rehabilitation, and intelligent wearable systems. In contrast to conventional approaches, which often rely on wired, foot-mounted inertial sensors and computationally intensive deep learning architectures, this study presents a wireless, thigh-mounted Inertial Measurement Unit (IMU)-based framework for step length estimation, employing a supervised learning paradigm to enhance accuracy, portability, and practicality in natural walking environments. Using an MPU-6050 IMU interfaced with an ESP32 module, the proposed SLE framework establishes a comprehensive data acquisition pipeline to enable seamless wireless transmission and real-time gait recording. Moreover, noise attenuation via Butterworth filtering and statistical normalization was also applied to refine motion signals. Additionally, fourteen engineered gait features extracted from segmented step events are employed to train four supervised learning algorithms, namely artificial neural networks (ANN), sequential neural networks (SNN), k-Nearest Neighbour (k-NN), and support vector machine (SVM), by this SLE framework. Evaluation of the proposed SLE model under both normal and fast walking conditions using a leave-one-out cross-validation scheme demonstrates SNN’s superiority over the other considered supervised models, with an outstanding average accuracy of over 99.4% and a reasonably superior average accuracy of over 83.5% achieved in the wired and wireless environments, respectively, across diverse gaits under both walking modes. With a certain marginal performance degradation, the wireless configuration still underscores its robustness and exhibits its potential for real-time gait monitoring through resource-constrained devices.

Gait analysis with step length estimation is essential not only for assessing an individual's health status, mobility, and balance but also for designing a pedestrian navigation system in infrastructure-less environments. This study presents an improved methodology for gait analysis and step length estimation (SLE) by applying a sequential neural network (SNN), a kind of artificial neural network, to sensor data collected from a foot-mounted inertial measurement unit. Our proposed methodology employs a Butterworth filter and normalization technique to smooth the sensor signals. Key gait-related features are then extracted and fed to the SNN, predicting step length without involving any user-specific parameter. The performance of our proposed SNN-based SLE method is evaluated and compared with our previously developed peak-valley detection-based SLE in terms of accuracy. Ten participants including both male and female (aged 24-50) took part in a walking experiment on a 60-meter linear path, walking under two modes – normal and fast. The accuracy results show that our new SNN-based SLE method achieves superior performance with accuracy exceeding 99.4% for each participant and demonstrates greater resilience to variations in user dynamics compared to our previously developed traditional Peak-Valley detection-based SLE method, confirming the proposed SNN-based approach as a superior solution.

The widespread prevalence of Wi-Fi and Bluetooth or its low-power version, Bluetooth low energy (BLE), in modern society has significantly increased their use for setting up wireless connectivity in various Internet-of Things (IoT)-based applications, such as patient monitoring, indoor localization and environment monitoring. The growing demand for integrating efficient power management solutions into these IoT applications has motivated us to focus on the measurement and analysis of power consumption during wireless scan activities using Wi-Fi and BLE since these are crucial for the connection establishment. Although a few research exists on the measurement and analysis of power consumed by wireless technologies like Wi-Fi, BLE and Zigbee in the literature, none have paid attention to the accurate measurement and analysis of real-time power consumption during wireless scan operations at the application level. To address this gap, we have designed a novel mobile application for the accurate measurement of real-time power consumption during wireless scan operations in order to assess the energy efficiency of various IoT applications involving those wireless technologies. Extensive experimentation was carried out in this study to measure real-time power consumption during Wi-Fi and BLE scans as well as operations without any scans (i.e., no scan) at varying durations and distances of the measuring device from the beacon. Additionally, a comparative statistical analysis of the collected power measurements under three different cases (Wi-Fi scan, BLE scan and no scan) using mean, median and standard deviation was performed to demonstrate the real-time power consumption usage patterns of these operations. This is followed by an insightful discussion on the influence of duration and distance on such power consumption patterns. Furthermore, several avenues for future research based on the findings of this study are also presented here.

5G positioning is an essential part of many applications, such as driverless vehicles, drone tracking, emergency services, location-based services, etc. Such positioning can attain an accuracy on the order of 1 meter or even lower in both indoor and outdoor settings due to its several advantageous aspects, which include millimeter wave signal, multiple input multiple output antennas, device-to-device communication, etc. Hence, many researchers have paid attention to the development of 5G positioning systems over the past decade. This paper, therefore, presents a state-of-the-art review of the existing 5G positioning methods along with their taxonomy and comparative analysis in several metrics including their merits and demerits. Moreover, the architecture and advantageous aspects of 5G positioning are also presented in this paper. Similar to the satellite signals, 5G signals, however, can be impeded by various obstructions, further declining the accuracy of such positioning. Moreover, many security concerns and privacy issues are also associated with 5G positioning. Thus, the development of a precise and secure 5G positioning system faces a number of challenges. Therefore, various ways to improve the 5G positioning accuracy and several solutions to address the security concerns and privacy issues of such positioning are pointed out in this study.