Motion Sensor Data Research Articles

A Novel Platform and Workflow for Underwater Photogrammetry Surveys

Abstract. Underwater photogrammetry has gained popularity for creating three-dimensional (3D) maps and ortho-images of marine environments as compared to traditional echosounder surveys, they can be more cost-effective at creating high-resolution 3D models, and orthoimages are usually more informative than acoustic backscatter maps. This paper builds on previous work by the authors that developed an underwater image-capturing platform with several additional sensors, including GNSS, IMU, pressure depth sensor and single beam echosounder. This study aims to analyse the impact of calibration and sensor data integration into the photogrammetric processing workflow. The tests were performed using two underwater sites. The low-cost device's pressure sensor and tide station data outperform GNSS PPK-derived heights. Furthermore, it was observed that incorporating IMU motion sensor data did not improve the processing results. Additionally, utilising the echosounder point cloud proves valuable for enhancing the overall quality of the survey. Despite its lower density, it serves a dual purpose by validating the photogrammetry dataset and, more importantly, can be employed for correcting DSM height. This study further underlined the importance of reliable camera calibration for accurate 3D reconstruction.

Ensem-DeepHAR: Identification of human activity in smart environments using ensemble of deep learning methods and motion sensor data

Recognizing human activity plays a crucial role in many applications such as medical care services in smart healthcare environments. Inertial or motion sensors can measure physiognomies such as acceleration and angular velocity of body movement while performing the activities and we can use them to learn the models capable of activity recognition. Over the past decades, many state-of-the-art activity recognition systems have been developed but there is still room to improve. In this paper, we have proposed a novel approach to identify human activity from motion sensor data by employing an enormous analysis of sensor data. Based on data analysis, we yielded quality data by preprocessing using a preprocessing chain for human activity recognition (PC-HAR) which also includes the Synthetic Minority Over-sampling Technique to balance the data of the dataset. As a recognition model, we proposed an ensemble of three different deep learning algorithms, namely, modified DeepConvLSTM, modified InceptionTime, and modified ResNet which is named ‘Ensem-DeepHAR’. The outcome of the proposed model is carried out by stacking predictions from each of the mentioned models and then a Random Forest as a meta-model uses those predictions to recognize the final activity. We evaluated our method on both person-dependent and person-independent cases and achieved 99.31 %, 99.08 %, and 97.52 % accuracies for the former case and 97.95 %, 98.11 %, and 99.51 % accuracies for the latter case using three common benchmark datasets: WISDM_ar_v1.1, PAMAP2, and UCI-HAR respectively. The various performance metrics and measures of experimental results establish the supremacy of the proposed model over the state-of-the-arts.

A physics-informed Bayesian data assimilation approach for real-time drilling tool lateral motion prediction

In this study, a Bayesian data assimilation method that fuses physics with motion sensor data is demonstrated to infer the dynamic states at points of interest on the bottomhole assembly (BHA) with proper uncertainty quantification. A 4.75 inch-LWD (Logging-while-drilling) tool has been used as a use case, where the dynamic states at the formation evaluation sensor can be predicted in real time with the measurements at the motion sensor as the required inputs. This was achieved with a developed transfer function that utilizes unscented Kalman filtering technique. The robustness of the transfer function was evaluated with synthetic data obtained from finite element analysis (FEA) simulations for various BHA configurations and drilling conditions. It was found that the prediction by the transfer function agrees favorably well with the true states of motion at the formation evaluation sensor. Specifically, using the developed transfer function can help reduce the relative errors for the motion trajectories at the formation evaluation sensor by a factor of 3, and can significantly enhance measurement quality risk classification. The developed transfer function method was further assessed with experimental roll test data, which is considered as close to drilling conditions. The prediction by the transfer function was found consistently close to the ground truth in the presence of backward whirl. The developed modeling method can potentially have broader impacts by enabling fit-for-basin virtual V&V (Verification and Validation) to accelerate LWD tool development, or enabling future drilling optimization.

Implementing Gait Kinematic Trajectory Forecasting Models on an Embedded System.

Smart algorithms for gait kinematic motion prediction in wearable assistive devices including prostheses, bionics, and exoskeletons can ensure safer and more effective device functionality. Although embedded systems can support the use of smart algorithms, there are important limitations associated with computational load. This poses a tangible barrier for models with increased complexity that demand substantial computational resources for superior performance. Forecasting through Recurrent Topology (FReT) represents a computationally lightweight time-series data forecasting algorithm with the ability to update and adapt to the input data structure that can predict complex dynamics. Here, we deployed FReT on an embedded system and evaluated its accuracy, computational time, and precision to forecast gait kinematics from lower-limb motion sensor data from fifteen subjects. FReT was compared to pretrained hyperparameter-optimized NNET and deep-NNET (D-NNET) model architectures, both with static model weight parameters and iteratively updated model weight parameters to enable adaptability to evolving data structures. We found that FReT was not only more accurate than all the network models, reducing the normalized root-mean-square error by almost half on average, but that it also provided the best balance between accuracy, computational time, and precision when considering the combination of these performance variables. The proposed FReT framework on an embedded system, with its improved performance, represents an important step towards the development of new sensor-aided technologies for assistive ambulatory devices.

MotionID: Towards practical behavioral biometrics-based implicit user authentication on smartphones

Traditional one-time authentication mechanisms cannot authenticate smartphone users’ identities throughout the session — the concept of using behavioral-based biometrics captured by the built-in motion sensors and touch data is a candidate to solve this issue. Many studies proposed solutions for behavioral-based continuous authentication; however, they are still far from practicality and generality for real-world usage. To date, no commercially deployed implicit user authentication scheme exists because most of those solutions were designed to improve detection accuracy without addressing real-world deployment requirements. To bridge this gap, we tackle the limitations of existing schemes and reach towards developing a more practical implicit authentication scheme, dubbed MotionID, based on a one-class detector using behavioral data from motion sensors when users touch their smartphones. Compared with previous studies, our work addresses the following challenges: ① Global mobile average to dynamically adjust the sampling rate for sensors on any device and mitigate the impact of using sensors’ fixed sampling rate; ② Over-all-apps to authenticate a user across all the mobile applications, not only on-specific application; ③ Single-device-evaluation to measure the performance with multiple users’ (i.e., genuine users and imposters) data collected from the same device; ④ Rapid authentication to quickly identify users’ identities using a few samples collected within short durations of touching (1–5 s) the device; ⑤ Unconditional settings to collect sensor data from real-world smartphone usage rather than a laboratory study. To show the feasibility of MotionID for those challenges, we evaluated the performance of MotionID with ten users’ motion sensor data on five different smartphones under various settings. Our results show the impracticality of using a fixed sampling rate across devices that most previous studies have adopted. MotionID is able to authenticate users with an F1-score up to 98.5% for some devices under practical requirements and an F1-score up to roughly 90% when considering the drift concept and rapid authentication settings. Finally, we investigate time efficiency, power consumption, and memory usage considerations to examine the practicality of MotionID.

Revealing the Mysteries of Population Mobility Amid the COVID-19 Pandemic in Canada: Comparative Analysis With Internet of Things-Based Thermostat Data and Google Mobility Insights.

The COVID-19 pandemic necessitated public health policies to limit human mobility and curb infection spread. Human mobility, which is often underestimated, plays a pivotal role in health outcomes, impacting both infectious and chronic diseases. Collecting precise mobility data is vital for understanding human behavior and informing public health strategies. Google's GPS-based location tracking, which is compiled in Google Mobility Reports, became the gold standard for monitoring outdoor mobility during the pandemic. However, indoor mobility remains underexplored. This study investigates in-home mobility data from ecobee's smart thermostats in Canada (February 2020 to February 2021) and compares it directly with Google's residential mobility data. By assessing the suitability of smart thermostat data, we aim to shed light on indoor mobility patterns, contributing valuable insights to public health research and strategies. Motion sensor data were acquired from the ecobee "Donate Your Data" initiative via Google's BigQuery cloud platform. Concurrently, residential mobility data were sourced from the Google Mobility Report. This study centered on 4 Canadian provinces-Ontario, Quebec, Alberta, and British Columbia-during the period from February 15, 2020, to February 14, 2021. Data processing, analysis, and visualization were conducted on the Microsoft Azure platform using Python (Python Software Foundation) and R programming languages (R Foundation for Statistical Computing). Our investigation involved assessing changes in mobility relative to the baseline in both data sets, with the strength of this relationship assessed using Pearson and Spearman correlation coefficients. We scrutinized daily, weekly, and monthly variations in mobility patterns across the data sets and performed anomaly detection for further insights. The results revealed noteworthy week-to-week and month-to-month shifts in population mobility within the chosen provinces, aligning with pandemic-driven policy adjustments. Notably, the ecobee data exhibited a robust correlation with Google's data set. Examination of Google's daily patterns detected more pronounced mobility fluctuations during weekdays, a trend not mirrored in the ecobee data. Anomaly detection successfully identified substantial mobility deviations coinciding with policy modifications and cultural events. This study's findings illustrate the substantial influence of the Canadian stay-at-home and work-from-home policies on population mobility. This impact was discernible through both Google's out-of-house residential mobility data and ecobee's in-house smart thermostat data. As such, we deduce that smart thermostats represent a valid tool for facilitating intelligent monitoring of population mobility in response to policy-driven shifts.

Higher awakening threshold of preterm infants in prone position may be a risk factor for SIDS.

The supine sleeping position in the prevention of sudden infant death syndrome in preterm infants is poorly understood. We aimed to investigate the effect of sleep posture on cardiorespiratory parameters and movement patterns in preterm infants close to discharge. This observational study included neonates born in 2022 at the University Hospital Schleswig-Holstein, Lübeck, Germany. Motion sensor data, heart rate, respiratory rate and oxygen saturation were recorded for infants with postconceptional age 35-37 weeks during sleep in the prone and supine positions. We recorded data from 50 infants, born at 31 (24-35) weeks of gestation (mean(range)), aged 5.2 ± 3.7 weeks (mean ± SD), of whom 48% were female. Five typical movement patterns were identified. In the prone position, the percentage of calm, regular breathing was higher and active movement was less frequent when compared to the supine position. The percentage of calm irregular breathing, number of apnoeas, bradycardias, desaturations and vital sign changes were not influenced by position. The prone position seems to be associated with a higher arousal threshold. The supine position appears advantageous for escape from life-threatening situations such as sudden infant death syndrome.

Combining Deep Learning with Signal-image Encoding for Multi-Modal Mental Wellbeing Classification

The quantification of emotional states is an important step to understanding wellbeing. Time series data from multiple modalities such as physiological and motion sensor data have proven to be integral for measuring and quantifying emotions. Monitoring emotional trajectories over long periods of time inherits some critical limitations in relation to the size of the training data. This shortcoming may hinder the development of reliable and accurate machine learning models. To address this problem, this article proposes a framework to tackle the limitation in performing emotional state recognition: (1) encoding time series data into coloured images; (2) leveraging pre-trained object recognition models to apply a Transfer Learning (TL) approach using the images from step 1; (3) utilising a 1D Convolutional Neural Network (CNN) to perform emotion classification from physiological data; (4) concatenating the pre-trained TL model with the 1D CNN. We demonstrate that model performance when inferring real-world wellbeing rated on a 5-point Likert scale can be enhanced using our framework, resulting in up to 98.5% accuracy, outperforming a conventional CNN by 4.5%. Subject-independent models using the same approach resulted in an average of 72.3% accuracy (SD 0.038). The proposed methodology helps improve performance and overcome problems with small training datasets.

Machine learning-based novel continuous authentication system using soft keyboard typing behavior and motion sensor data

Smartphones utilize various authentication methods, including passwords, fingerprints, and face recognition. While this information is quite practical and easy to remember, it introduces several security issues. The primary concerns involve theft, password forgetfulness, or unauthorized password copying. Implementing behavioral biometrics for user authentication adds an extra layer of security. The main contribution of this study is the utilization of soft keyboard typing behavior, a behavioral biometric, for continuous user recognition. To achieve this, the phone's grip style and typing characteristics of users are scrutinized using data collected from motion sensors and the touchscreen panel. Another challenge in mobile device authentication pertains to recognition accuracy and processing time. To expedite and optimize data classification, a hybrid classification structure is suggested. This structure incorporates correlation-based feature selection and a straightforward logistic regression method, offering rapid and highly accurate classification outcomes—a further contribution of this study. Experimental results demonstrate that user identification can be accomplished in as little as 0.03 ms, with a classification accuracy of up to 93%. Continuous authentication systems offer greater security compared to one-time authentication systems. Nevertheless, these systems might not always yield the most precise results. Overcoming this challenge necessitates the development of an efficient software architecture. In line with this, an additional contribution of this study is an explanation of how to construct a continuous authentication system using the developed architecture.

Afternoons can be a critical time of the day that reflects associations between life‐space mobility and loneliness among living alone older adults: longitudinal analyses with motion sensor data

AbstractBackgroundDaily activity patterns of life‐space mobility can serve as indicators for loneliness and subsequent cognitive change. Assessment of life‐space mobility often uses self‐reported data, and the measurement intervals are often less frequent than needed to capture change over time. This study examined whether the life‐space mobility changed when loneliness was reported weekly.MethodsParticipants were 139 older adults who lived alone (age = 78.1±8.6, 103 (74%) were female, 32 (23%) were African Americans, 19 (14%) had MCI diagnosis) and were enrolled in ORCATECH (ORegon Center for Aging & TECHnology) longitudinal aging studies and Minority Aging Research Study. Passive infrared motion sensors were placed in common rooms of each participant’s home (bathroom; bedroom; kitchen; living room). Time spent in each room across the day (morning; afternoon; evening; night) and out‐of‐home were used as indicators of life‐space mobility and were derived from home sensor data. Participants reported whether they felt lonely on a weekly basis using an online survey sent via email. Generalized estimating equations were used to correlate the time spent in each room and out‐of‐home and the likelihood of feeling lonely. Age, gender, education, race, cognitive status (MCI/ non‐MCI), and average daily time out‐of‐home were controlled in the model.ResultsWe analyzed 4994 weeks of data; 232 weeks had an incidence of loneliness (5%). Spending an additional hour in the bedroom in the afternoon (noon‐6 pm) was associated with a 21% increase in the likelihood of experiencing loneliness (p = .03). Spending an additional hour out‐of‐home in the afternoon was associated with a 13% decrease in the likelihood of experiencing loneliness (p = .02). Time spent in the kitchen, bathroom, and living room in a day was unrelated to weekly loneliness.ConclusionTime spent outside the home may suggest an expansion of life‐space, while spending time in one’s bedroom may indicate contraction of life space. Our findings suggest carefully monitoring a person’s location in the afternoon might help detect loneliness in this population. Frequent and objective measurements of life‐space mobility can produce renewed insights into the experience of loneliness among older adults who live alone, and thus, inform timely interventions for loneliness and prevent cognitive decline.

Fine-Grained Intoxicated Gait Classification Using a Bilinear CNN.

Consuming excessive amounts of alcohol causes impaired mobility and judgment and driving accidents, resulting in more than 800 injuries and fatalities each day. Passive methods to detect intoxicated drivers beyond the safe driving limit can facilitate Just-In-Time alerts and reduce Driving Under the Influence (DUI) incidents. Popularly-owned smartphones are not only equipped with motion sensors (accelerometer and gyroscope) that can be employed for passively collecting gait (walk) data but also have the processing power to run computationally expensive machine learning models. In this paper, we advance the state-of-the-art by proposing a novel method that utilizes a Bi-linear Convolution Neural Network (BiCNN) for analyzing smartphone accelerometer and gyroscope data to determine whether a smartphone user is over the legal driving limit (0.08) from their gait. After segmenting the gait data into steps, we converted the smartphone motion sensor data to a Gramian Angular Field (GAF) image and then leveraged the BiCNN architecture for intoxication classification. Distinguishing GAF-encoded images of the gait of intoxicated vs. sober users is challenging as the differences between the classes (intoxicated vs. sober) are subtle, also known as a fine-grained image classification problem. The BiCNN neural network has previously produced state-of-the-art results on fine-grained image classification of natural images. To the best of our knowledge, our work is the first to innovatively utilize the BiCNN to classify GAF encoded images of smartphone gait data in order to detect intoxication. Prior work had explored using the BiCNN to classify natural images or explored other gait-related tasks but not intoxication Our complete intoxication classification pipeline consists of several important pre-processing steps carefully adapted to the BAC classification task, including step detection and segmentation, data normalization to account for inter-subject variability, data fusion, GAF image generation from time-series data, and a BiCNN classification model. In rigorous evaluation, our BiCNN model achieves an accuracy of 83.5%, outperforming the previous state-of-the-art and demonstrating the feasibility of our approach.

Otokodlayıcı Tabanlı Boyut Azaltma ve Akıllı Saat Tabanlı Giyilebilir Hareket Algılayıcıları Kullanarak Yaşlılarda Düşme Tespiti

Falling is a serious health risk that can even result in death, especially for the elderly. For this reason, it 
 is crucial to prevent falls and, in cases where prevention is not possible, to detect and intervene as soon 
 as possible. Smartwatches are an ideal tool for fall detection due to their constant presence, rich sensor 
 resources, and communication capabilities. The aim of this study is to detect falls in elderly people with 
 high accuracy using motion sensor data obtained from smartwatches. To achieve this, a dataset was 
 created consisting of falls and daily activities. Then, the feature vector was extracted which has 
 provided successful results in signal processing studies. Afterward, the dimensionality of the dataset 
 was reduced using an autoencoder-based approach in order to decrease the workload on smartwatches 
 and ensure more accurate and faster classification. The dataset was classified using machine learning 
 methods including naive Bayes, logistic regression, and C4.5 decision tree, and successful results were 
 obtained. Their performances were then compared. It was observed that reducing the dimensionality 
 had positive effects on both the classification accuracy and the computation time.

Unsupervised Domain Adaptation for Mitigating Sensor Variability and Interspecies Heterogeneity in Animal Activity Recognition.

Animal activity recognition (AAR) using wearable sensor data has gained significant attention due to its applications in monitoring and understanding animal behavior. However, two major challenges hinder the development of robust AAR models: domain variability and the difficulty of obtaining labeled datasets. To address this issue, this study intensively investigates the impact of unsupervised domain adaptation (UDA) for AAR. We compared three distinct types of UDA techniques: minimizing divergence-based, adversarial-based, and reconstruction-based approaches. By leveraging UDA, AAR classifiers enable the model to learn domain-invariant features, allowing classifiers trained on the source domain to perform well on the target domain without labels. We evaluated the effectiveness of UDA techniques using dog movement sensor data and additional data from horses. The application of UDA across sensor positions (neck and back), sizes (middle-sized and large-sized), and gender (female and male) within the dog data, as well as across species (dog and horses), exhibits significant improvements in the classification performance and reduced the domain discrepancy. The results highlight the potential of UDA to mitigate the domain shift and enhance AAR in various settings and for different animal species, providing valuable insights for practical applications in real-world scenarios where labeled data is scarce.

A Novel Heuristic Fall-Detection Algorithm Based on Double Thresholding, Fuzzy Logic, and Wearable Motion Sensor Data

A Novel Heuristic Fall-Detection Algorithm Based on Double Thresholding, Fuzzy Logic, and Wearable Motion Sensor Data

Tank test results of synthetic aperture sonar image reconsturction using two frequency bands

Synthetic aperture experiments have been done in water tanks using two different frequency bands. Their center frequencies are around 25 kHz and 300 kHz, respectively. The lower frequency system has eight equally spaced receiver channels with one wavelength distance and the higher frequency one has sixteen receiver channels with about seven wavelength spacing. Five different types of structures were selected as sonar targets. They are sphere, cylinder, octahedron-shape structure, star-shape structure, and an in-plane grid target. The sphere and cylinder were floating in the water column while the others were located on the tank bottoms. As the transmitting signals linear frequency modulation was applied. Sonar images were reconstructed by synthetic aperture processing and the deviated motion from the perfect rectilinear trajectory due to the vibration of the vertical support was compensated by using motion sensor data and receiver signals. Synthetic aperture images were experimentally generated using the two different frequency bands and compared. [This research was supported by KIMST funded by the Agency of Korea Coast Guard (KIMST-20210547).]

In-situ empirical validation of common indoor climate parameters in an inhabited multizone dwelling

Combined building indoor climate and energy simulation models only recently gained vast popularity and their application has been moving from the research community to a broader audience. Yet, in-situ empirical validation of this new generation of complex multi-purpose dynamic simulation models has lagged behind. Using a dynamic multizone building indoor climate and energy simulation model in Modelica with the IDEAS library and buoyancy driven airflow components (validated with CONTAM), this research presents model validation results and lessons learned from an in-situ empirical validation study of common indoor climate parameters (i.e., indoor air temperature (), relative humidity (RH) and CO2 concentration (CO2)) for an inhabited and mechanically ventilated case study dwelling in The Netherlands. The simulation results show that the latest generation of building indoor climate and energy models in Modelica have great ability to accurately predict common indoor climate parameters in multizone inhabited dwellings (provided that user behavior info is available). Evaluation metrics for the three studied parameters show excellent calibration criteria (i.e., MAE between 0.60–0.78 °C (), 3.5–4.6% (RH) and 88–181 ppm (CO2)) and the accompanying graphs corroborate the findings. In the event that no motion sensor data is available, statistically generated occupancy profiles prove good representative alternatives on the condition that basic info is available about the number of inhabitants and the inhabitants’ lifestyle. In-situ monitoring for empirical model validation proves to be a real challenge full of (un)foreseen obstacles.

A new CNN-LSTM architecture for activity recognition employing wearable motion sensor data: Enabling diverse feature extraction

Extracting representative features to recognize human activities through the use of wearables is an area of on-going research. While hand-crafted features and machine learning (ML) techniques have been sufficiently well investigated in the past, the use of deep learning (DL) techniques is the current trend. Specifically, Convolutional Neural Networks (CNNs), Long Short Term Memory Networks (LSTMs), and hybrid models have been investigated. We propose a novel hybrid network architecture to recognize human activities through the use of wearable motion sensors and DL techniques. The LSTM and the 2D CNN branches of the model that run in parallel receive the raw signals and their spectrograms, respectively. We concatenate the features extracted at each branch and use them for activity recognition. We compare the classification performance of the proposed network with three single and three hybrid commonly used network architectures: 1D CNN, 2D CNN, LSTM, standard 1D CNN-LSTM, 1D CNN-LSTM proposed by Ordóñez and Roggen, and an alternative 1D CNN-LSTM model. We tune the hyper-parameters of six of the models using Bayesian optimization and test the models on two publicly available datasets. The comparison between the seven networks is based on four performance metrics and complexity measures. Because of the stochastic nature of DL algorithms, we provide the average values and standard deviations of the performance metrics over ten repetitions of each experiment. The proposed 2D CNN-LSTM architecture achieves the highest average accuracies of 95.66% and 92.95% on the two datasets, which are, respectively, 2.45% and 3.18% above those of the 2D CNN model that ranks the second. This improvement is a consequence of the proposed model enabling the extraction of a broader range of complementary features that comprehensively represent human activities. We evaluate the complexities of the networks in terms of the total number of parameters, model size, training/testing time, and the number of floating point operations (FLOPs). We also compare the results of the proposed network with those of recent related work that use the same datasets.

Road conditions monitoring using semantic segmentation of smartphone motion sensor data

Road conditions monitoring using semantic segmentation of smartphone motion sensor data

Improved electrode configuration for Transcutaneous Afferent Patterned Stimulation for Essential Tremor and Parkinson’s Disease (P7-11.009)

<h3>Objective:</h3> This study evaluated a modified electrode configuration for Transcutaneous afferent patterned stimulation (TAPS) intended to reduce skin irritation without compromising efficacy. <h3>Background:</h3> TAPS is a FDA-cleared, non-invasive, wrist-worn therapy for hand tremor in Essential Tremor (ET) also granted Breakthrough Designation for action tremor in Parkinson’s disease (PD). The most prevalent adverse event of TAPS therapy is dorsal wrist skin irritation. <h3>Design/Methods:</h3> Finite element modeling suggested a 6-electrode configuration provided similar median and radial nerve excitation with lower stimulation amplitudes than standard TAPS 3-electrode configuration. Nineteen action tremor subjects with ET or PD were enrolled in a study comparing telemedicine clinical ratings (TETRAS or UPDRS), activities of daily living (BF-ADL), and motion sensor data (tremor power improvement ratio, TPIR) over two weeks of home-use per configuration. <h3>Results:</h3> Mean stimulation amplitude was reduced from 4.0mA to 2.7mA from 3- to 6-electrodes, respectively (p=0.0001). Adverse events (e.g., skin irritation) were 71% (5 of 7) and 29% (2 of 7) with 3- and 6-electrodes (p=0.3). Efficacy was similar between electrode configurations. Median TPIR was 2.2 (IQR: 1.2–4.3) and 2.2 (1.2–2.7) for 3- and 6-electrodes (p&lt;0.0001 per configuration, p=0.3 between configurations). Responder analysis showed 87% improved (TPIR&gt;1) with both configurations, while 53% and 60% improved 2-fold (TPIR&gt;2) for 3- and 6-electrodes. Mean BF-ADL improved 0.5 (SD: 0.3) and 0.5 (0.6) with 3- and 6-electrodes (p=1.0), while mean TETRAS improved 0.4 (0.3) and 0.4 (0.5) (p=0.8, N=10) and mean UPDRS improved 0.9 (0.6) and 0.6 (0.2) respectively (N=6). <h3>Conclusions:</h3> The 6-electrode configuration reduced stimulation amplitudes while maintaining efficacy of the 3-electrode configuration. Future studies will compare the configurations with extended duration stimulation. <b>Disclosure:</b> Dr. Kent has received personal compensation for serving as an employee of Cala Health. Dr. Kent has stock in Cala Health. Dr. Kent has received intellectual property interests from a discovery or technology relating to health care. Dr. Kent has received intellectual property interests from a discovery or technology relating to health care. Ms. Aratia has nothing to disclose. Dr. Khosla has received personal compensation for serving as an employee of Cala Health. Dr. Khosla has received personal compensation in the range of $100,000-$499,999 for serving as a Consultant for Cala Health. Ruta Deshpande has nothing to disclose. Kathryn Rosenbluth has received personal compensation for serving as an employee of Cala Health. Kathryn Rosenbluth has received stock or an ownership interest from Cala Health.

Research and Implementation of Low-Power Anomaly Recognition Method for Intelligent Manhole Covers

This paper addresses the difficulty of balancing a real-time response and low power consumption in intelligent manhole cover application scenarios. It proposes a method to distinguish normal and abnormal events by segmenting the boundary at which the acceleration of the intelligent manhole cover deviates from a set threshold and lasts for a certain period, based on the difference in the intelligent manhole cover’s vibration patterns when a normal event and an abnormal event occur. This paper uses the autonomous data fusion of digital output motion sensor data to implement a pattern recognition algorithm for the above-mentioned pattern, which reduces the MCU computing and working time and the overall power consumption of the system while meeting real-time response requirements. The test results demonstrate that the method has a high rate of anomaly recognition accuracy. The method ensures the system’s real-time response capability, and the actual low power consumption test demonstrates that the device can operate continuously for 9.5 years. The low power consumption index exceeds the requirements of the existing national standard, thereby resolving the issue that it is challenging to balance intelligent manhole cover abnormality recognition and low power consumption.