Tri-axial Acceleration Research Articles

A supervised model to identify wolf behavior from tri-axial acceleration

BackgroundIn wildlife studies, animal behavior serves as a key indicator of the impact of environmental changes and anthropogenic disturbances. However, wild animals are elusive and traditional GPS studies only provide limited insight into their daily activities. To address this issue, behavior classification models have increasingly been used to detect specific behaviors in wildlife equipped with tri-axial accelerometers. Such models typically need to be trained on data from the target species. The present study focuses on developing a behavioral classification model tailored to the grey wolf (Canis lupus) and encompassing a variety of ecologically relevant behaviors.MethodsWe collected data from nine captive wolves equipped with collar-mounted tri-axial accelerometers recording continuous acceleration at 32 Hz (“fine-scale”) and averaged acceleration over 5-min intervals (“activity”). Using simultaneous video observations, we trained Random Forest models to classify wolf acceleration data into specific behaviors. We investigated the potential limits to the generalizability of these models to unlabeled data through individual-based cross-validation.ResultsWe present: (1) a model classifying fine-scale acceleration data (32 Hz) into 12 distinct behaviors (lying, trotting, stationary, galloping, walking, chewing, sniffing, climbing, howling, shaking, digging and jumping) with a class recall of 0.77–0.99 (0.01–0.91 in cross-validation), (2) a model classifying activity data (5-min averages) into 3 behavior categories (static, locomotion and miscellaneous) with a class recall of 0.43–0.91 (0.39–0.92 in cross-validation). Although classification performance decreased following cross-validation, recall scores for lying, trotting, stationary, galloping, walking and chewing individual behaviors (as well as static and locomotion categories) remained above 0.6. Classification performance was consistently poorer for rare behaviors, which constituted less than 1.1% of the training dataset.ConclusionsWe demonstrate the use of collar-mounted accelerometer to distinguish between 12 behaviors and 3 behavior categories in captive wolves, at fine-scale (32 Hz) and averaged 5-min resolutions, respectively. We also discuss the generalizability of the two models to free-ranging settings. These models can be employed to support future behavioral studies examining questions such as conflict mitigation, wolf responses to human disturbances, or specific activity budgets.

Cough recognition in pneumoconiosis patients based on a flexible patch with an embedded ACC sensor for remote monitoring

BackgroundMany respiratory diseases such as pneumoconiosis require to close monitor the symptoms such as abnormal respiration and cough. This study introduces an automated, nonintrusive method for detecting cough events in clinical settings using a flexible chest patch with tri-axial acceleration sensors.MethodsTwenty-five young healthy persons (hereinafter referred to as healthy adults) and twenty-five clinically diagnosed pneumoconiosis patients (hereinafter referred to as patients) participated in the experiment by wearing a flexible chest patch with an embedded ACC sensor. The top 56% of the highest scoring features were then combined using several feature selection algorithms to perform the cough classification task. The multicriteria decision making (MCDM) method was used to select the classifier with the highest scores.ResultsThe optimized classifier proposed in this paper achieved an accuracy of 87.1%, precision of 95%, recall of 79.1%, F1 score of 86.4%, and AUC of 95.4% for recognizing coughs in healthy adults; an accuracy of 96.1%, precision of 95%, recall of 97.4%, F1 score of 96.2%, and AUC of 98.7% for recognizing coughs in patients; and an overall accuracy of 92% for distinguishing coughs in the combined group of healthy adults and patients.ConclusionsOur study demonstrated the effectiveness of an automated cough recognition system in both pneumoconiosis patients and healthy adults. This approach facilitates daily remote monitoring of cough occurrence in individuals with pneumoconiosis, potentially enhancing the ability of physicians to evaluate clinical status.

Transformer-Driven Affective State Recognition from Wearable Physiological Data in Everyday Contexts

The rapid advancement in wearable physiological measurement technology in recent years has brought affective computing closer to everyday life scenarios. Recognizing affective states in daily contexts holds significant potential for applications in human–computer interaction and psychiatry. Addressing the challenge of long-term, multi-modal physiological data in everyday settings, this study introduces a Transformer-based algorithm for affective state recognition, designed to fully exploit the temporal characteristics of signals and the interrelationships between different modalities. Utilizing the DAPPER dataset, which comprises continuous 5-day wrist-worn recordings of heart rate, skin conductance, and tri-axial acceleration from 88 subjects, our Transformer-based model achieved an average binary classification accuracy of 71.5% for self-reported positive or negative affective state sampled at random moments during daily data collection, and 60.29% and 61.55% for the five-class classification based on valence and arousal scores. The results of this study demonstrate the feasibility of applying affective state recognition based on wearable multi-modal physiological signals in everyday contexts.

Depression Recognition Using Daily Wearable-Derived Physiological Data.

The objective identification of depression using physiological data has emerged as a significant research focus within the field of psychiatry. The advancement of wearable physiological measurement devices has opened new avenues for the identification of individuals with depression in everyday-life contexts. Compared to other objective measurement methods, wearables offer the potential for continuous, unobtrusive monitoring, which can capture subtle physiological changes indicative of depressive states. The present study leverages multimodal wristband devices to collect data from fifty-eight participants clinically diagnosed with depression during their normal daytime activities over six hours. Data collected include pulse wave, skin conductance, and triaxial acceleration. For comparison, we also utilized data from fifty-eight matched healthy controls from a publicly available dataset, collected using the same devices over equivalent durations. Our aim was to identify depressive individuals through the analysis of multimodal physiological measurements derived from wearable devices in daily life scenarios. We extracted static features such as the mean, variance, skewness, and kurtosis of physiological indicators like heart rate, skin conductance, and acceleration, as well as autoregressive coefficients of these signals reflecting the temporal dynamics. Utilizing a Random Forest algorithm, we distinguished depressive and non-depressive individuals with varying classification accuracies on data aggregated over 6 h, 2 h, 30 min, and 5 min segments, as 90.0%, 84.7%, 80.1%, and 76.0%, respectively. Our results demonstrate the feasibility of using daily wearable-derived physiological data for depression recognition. The achieved classification accuracies suggest that this approach could be integrated into clinical settings for the early detection and monitoring of depressive symptoms. Future work will explore the potential of these methods for personalized interventions and real-time monitoring, offering a promising avenue for enhancing mental health care through the integration of wearable technology.

Identification of Non-Restorative Sleep Associated with Depression Using Ambulatory Electrocardiogram and Triaxial Acceleration

Identification of Non-Restorative Sleep Associated with Depression Using Ambulatory Electrocardiogram and Triaxial Acceleration

Automatic Recognition of Motor Skills in Triathlon: A Novel Tool for Measuring Movement Cadence and Cycling Tasks.

Background/Objectives: The purpose of this research was to create a peak detection algorithm and machine learning model for use in triathlon. The algorithm and model aimed to automatically measure movement cadence in all three disciplines of a triathlon using data from a single inertial measurement unit and to recognise the occurrence and duration of cycling task changes. Methods: Six triathletes were recruited to participate in a triathlon while wearing a single trunk-mounted measurement unit and were filmed throughout. Following an initial analysis, a further six triathletes were recruited to collect additional cycling data to train the machine learning model to more effectively recognise cycling task changes. Results: The peak-counting algorithm successfully detected 98.7% of swimming strokes, with a root mean square error of 2.7 swimming strokes. It detected 97.8% of cycling pedal strokes with a root mean square error of 9.1 pedal strokes, and 99.4% of running strides with a root mean square error of 1.2 running strides. Additionally, the machine learning model was 94% (±5%) accurate at distinguishing between 'in-saddle' and 'out-of-saddle' riding, but it was unable to distinguish between 'in-saddle' riding and 'coasting' based on tri-axial acceleration and angular velocity. However, it displayed poor sensitivity to detect 'out-of-saddle' efforts in uncontrolled conditions which improved when conditions were further controlled. Conclusions: A custom peak detection algorithm and machine learning model are effective tools to automatically analyse triathlon performance.

Analysis and Decoupling Study of the Coupling Effects on Projectile Acceleration Parameters

Abstract Internal ballistic parameter testing is crucial for evaluating artillery performance and design, with projectile acceleration being a core parameter. During the firing process, projectile acceleration parameters can easily be coupled with the projectile’s modal effects and the transverse effects of triaxial accelerometers, leading to inaccurate measurements. Therefore, this paper innovatively proposes a data decoupling method to improve the accuracy of acceleration measurements. Firstly, the internal ballistic firing mechanism and the transverse effects of triaxial accelerometers are analyzed. Secondly, a simulation analysis of the projectile modal is conducted, combined with the projectile’s motion state within the barrel, to determine the modal orders and frequencies under axial stretching and compression vibration modes. Thirdly, shock tests are performed on triaxial accelerometers using a shock table to establish the relationships between triaxial acceleration signals. Finally, the projectile modal, Empirical Mode Decomposition (EMD), and the transverse effects of triaxial accelerometers are combined to decouple the acceleration signals. A comparison between the decoupled acceleration signals and theoretical data shows that the error is reduced from 8.1% to 3.2%, significantly improving the accuracy of the acceleration signals.

Innovative Approach to Fish Stress Analysis: AI-Enhanced Biosensing and Accelerometer Integration

Background and Objectives:In our collaborative efforts, we have pioneered the development of a wireless biosensor system that can monitor blood glucose levels in fish, serving as a indicator of stress response. However, while the response value of the biosensor detected the increase in the fish's blood glucose level, it was a challenge to pinpoint the exact factors causing it. To address this, we augmented our measurements with data from a triaxial acceleration sensor and posture estimation using deep learning. This integration of biosensors, physical sensors, and video analysis technology, a testament to the power of interdisciplinary research, allowed us to delve deeper into the state of the fish.Materials and Methods:The glucose biosensor, constructed using a Pt/Ir wire (φ: 0.178 mm) as the working electrode, Ag/AgCl as the reference electrode, and glucose oxidase immobilized on the working electrode surface, was connected to a data logger with a built-in triaxial acceleration sensor. This biosensor was then implanted in a Nile tilapia (Oreochromis niloticus), and the data logger was attached to the fish's side. After allowing the fish to swim freely in the experimental tank overnight, it was introduced to a larger individual, and its stress responses and acceleration due to fighting behavior were recorded. The confrontation was filmed from above the tank, and the positional coordinates of each individual in the filmed images were obtained using DeepLabCut.Results and Discussion:The response value of the biosensor of the individual that became subordinate in the tank from the middle of the confrontation increased. It is thought this was a stress response to being attacked by a larger individual and cornered against the wall in the tank. Acceleration also captured fluctuations derived from the intense fighting behavior during the confrontation. The positional coordinates of each individual reflected the power relationship between the larger and smaller individuals. The larger individual was often located near the center of the tank because of its dominance in the water tank.In contrast, the smaller, subordinate individual was often located near the wall and occasionally pushed against the wall by the larger individual. In this experiment, we analyzed one-hour-long videos during the confrontation to observe the positional relationship between the two individuals from the beginning to the end. Combined with detecting stress responses by biosensors, the possibility of exploring which fish behaviors are accompanied by changes in physiological conditions while the fish are swimming has become apparent.Conclusion:Data from physical sensors such as triaxial accelerometers and video analysis can enhance the information from biosensor response values. The fighting behavior of fish and the social relationships among fish in the tank were quantified by accelerometers and video analysis, shedding light on the causes of fluctuations in fish blood glucose levels. Given the close relationship between fish behavior and physiological state, the integration of physical sensor and video analysis technology with biosensor technology holds promise for improving fish health management in the field. Moreover, by experimentally correlating the fluctuations in biosensor values with the behavior of the fish in the video, it may be possible in the future to estimate the physiological state of the fish simply by monitoring the fish with a camera, a potential game-changer in fish health monitoring. Figure 1

Predicting Operational Events in Mechanized Weed Control Operations by Offline Multi-Modal Data and Machine Learning Provides Highly Accurate Classification in Time Domain

Monitoring of operations has become a critical activity in forestry, aiming to provide the data required by planning and production management. Conventional methods, on the other hand, come at a high expense of resources. A neural network was trained, validated, and tested in this study based on multi-modal data to classify relevant operational events in mechanized weed control operations. The architecture of a neural network was tuned in terms of the number of hidden layers and neurons, and the regularization term was set at various values to obtain optimally tuned models for three data modalities: triaxial acceleration data coupled with speed extracted from GNSS signals (AS), triaxial acceleration (A), and speed alone (S). In the training and validation phase, the models based on AS and A achieved a very high classification accuracy, accounting for 92 to 93% when considering four relevant events. In the testing phase, which was run on unseen data, the classification accuracy reached figures of 91 to 92%, indicating a good generalization ability of the models. The results point out that multimodal data are able to provide the features for distinguishing events and add spatial context to the monitored operations, standing as a suitable solution for offline, partly automated monitoring. Future studies are required to see how the capabilities of online, real-time technologies such as deep learning coupled with computer vision can add more context and improve classification performance.

Monitoring Multiple Behaviors in Beef Calves Raised in Cow-Calf Contact Systems Using a Machine Learning Approach.

The monitoring of pre-weaned calf behavior is crucial for ensuring health, welfare, and optimal growth. This study aimed to develop and validate a machine learning-based technique for the simultaneous monitoring of multiple behaviors in pre-weaned beef calves within a cow-calf contact (CCC) system using collar-mounted sensors integrating accelerometers and gyroscopes. Three complementary models were developed to classify feeding-related behaviors (natural suckling, feeding, rumination, and others), postural states (lying and standing), and coughing events. Sensor data, including tri-axial acceleration and tri-axial angular velocity, along with video recordings, were collected from 78 beef calves across two farms. The LightGBM algorithm was employed for behavior classification, and model performance was evaluated using a confusion matrix, the area under the receiver operating characteristic curve (AUC-ROC), and Pearson's correlation coefficient (r). Model 1 achieved a high performance in recognizing natural suckling (accuracy: 99.10%; F1 score: 96.88%; AUC-ROC: 0.999; r: 0.997), rumination (accuracy: 97.36%; F1 score: 95.07%; AUC-ROC: 0.995; r: 0.990), and feeding (accuracy: 95.76%; F1 score: 91.89%; AUC-ROC: 0.990; r: 0.987). Model 2 exhibited an excellent classification of lying (accuracy: 97.98%; F1 score: 98.45%; AUC-ROC: 0.989; r: 0.982) and standing (accuracy: 97.98%; F1 score: 97.11%; AUC-ROC: 0.989; r: 0.983). Model 3 achieved a reasonable performance in recognizing coughing events (accuracy: 88.88%; F1 score: 78.61%; AUC-ROC: 0.942; r: 0.969). This study demonstrates the potential of machine learning and collar-mounted sensors for monitoring multiple behaviors in calves, providing a valuable tool for optimizing production management and early disease detection in the CCC system.

Amplitude of Biceps Femoris Activation and Triaxial Acceleration in a 50-Meter Test in Sprinters: Pilot Study

Objective: To describe the relationship between bilateral electrical activity in the biceps femoris and the variation of triaxial acceleration in three 50 m sprints. Methods: Biceps femoris muscle activation and acceleration in the anterior–posterior, mediolateral, and superior–inferior axes were measured in three 50 m sprints in nine national-level sprinters. Results: There was significant asymmetry between both legs, and the variations between axes were significant between the anterior–posterior with respect to the lateral and superior–inferior, and between the lateral and superior–inferior (p < 0.05). Conclusions: Increased biceps femoris activation during running increases speed regardless of asymmetry in force application. In the maintenance of horizontal velocity, acceleration of the anterior–posterior axis is the most relevant and depends on the flexion-extension muscle actions contained in the lateral axis.

Classification of movements of collegiate female soccer players using inertial measurement units

This study aimed to classify the movements of female soccer players during matches using raw data measured by inertial measurement units (IMUs). Twelve collegiate female soccer players were equipped with IMUs (100 Hz), and raw triaxial acceleration data from eight official matches were analyzed. The measurement data were separated every 3 s, and a Fast Fourier Transform (FFT) was performed. After FFT, the Euclidean distances between the data when the players were in the stationary state and other states were calculated to classify the movements of the players using the k-means method. The data of the clustering numbers classified as the stationary state were eliminated after analyzing the movements of players by video filming the matches. After classification, the average Euclidean distances between the stationary state and other movements were calculated. Consequently, the results showed that the upward and downward directions of the raw data affected the classification. Using the methods of this study, it was also shown that the distribution of Euclidean distances differed from player to player. Our findings indicate that the method used in this study can be used to classify and characterize the movements of female collegiate soccer players.

Classification of sex-dependent specific behaviours by tri-axial acceleration in the tegu lizard Salvator merianae

Validated patterns of behaviour detected by tri-axial acceleration in the laboratory can be used for remote measurements of free-living animals. The tegu lizard naturally occupies diverse biomes in South America and presents ecological threats in regions where it was artificially introduced. We aimed to validate the use of tri-axial acceleration to distinguish among behaviours of male and female tegus in captivity by comparing observed behaviours to recorded acceleration data. Adult animals were externally fitted with an accelerometer fixed between their scapulae to quantify anteroposterior, lateral, and dorsoventral acceleration. Video recordings of cameras positioned on the walls of the animal-holding arena documented behaviours. Behaviour patterns, such as resting, walking, and eating, were identified for both sexes, and nest building in females and courtship and copulation in males. Random Forest algorithm was used to validate the behaviour patterns from accelerometry data based on two models, random split (70 % training-30 % validation; RS) and leave-one-out (divided by individual; LOO). Although LOO showed lower accuracies than RS for all the acceleration data, nest building in females and copulation in males had high accuracies in both models. In contrast, the lowest accuracies for walking and eating indicates they may involve more inconsistent movement patterns. Comparing the results from RS and LOO, female behaviours may be more identifiable in the field using triaxial accelerometry than males. The identification of behaviours by accelerometry, especially related to reproduction, without the necessity for direct observation of the tegus would be helpful for conservation purposes, for both natural and invasive populations.

Seasonal habitat use and diel vertical migration in female spurdog in Nordic waters

BackgroundStudying habitat use and vertical movement patterns of individual fish over continuous time and space is innately challenging and has therefore largely remained elusive for a wide range of species. Amongst sharks, this applies particularly to smaller-bodied and less wide-ranging species such as the spurdog (Squalus acanthias Linnaeus, 1758), which, despite its importance for fisheries, has received limited attention in biologging and biotelemetry studies, particularly in the North-East Atlantic.MethodsTo investigate seasonal variations in fine-scale niche use and vertical movement patterns in female spurdog, we used archival data from 19 pregnant individuals that were satellite-tagged for up to 365 days in Norwegian fjords. We estimated the realised niche space with kernel densities and performed continuous wavelet analyses to identify dominant periods in vertical movement. Triaxial acceleration data were used to identify burst events and infer activity patterns.ResultsPregnant females frequently utilised shallow depths down to 300 m at temperatures between 8 and 14 °C. Oscillatory vertical moments revealed persistent diel vertical migration (DVM) patterns, with descents at dawn and ascents at dusk. This strict normal DVM behaviour dominated in winter and spring and was associated with higher levels of activity bursts, while in summer and autumn sharks predominantly selected warm waters above the thermocline with only sporadic dive and bursts events.ConclusionsThe prevalence of normal DVM behaviour in winter months linked with elevated likely foraging-related activity bursts suggests this movement behaviour to be foraging-driven. With lower number of fast starts exhibited in warm waters during the summer and autumn months, habitat use in this season might be rather driven by behavioural thermoregulation, yet other factors may also play a role. Individual and cohort-related variations indicate a complex interplay of movement behaviour and habitat use with the abiotic and biotic environment. Together with ongoing work investigating fine-scale horizontal movement as well as sex- and age-specific differences, this study provides vital information to direct the spatio-temporal distribution of a newly reopened fishery and contributes to an elevated understanding of the movement ecology of spurdog in the North-East Atlantic and beyond.Graphical

A Machine Learning Model for Predicting Critical Minimum Foot Clearance (MFC) Heights

Tripping is the largest cause of falls, and low swing foot ground clearance during the mid-swing phase, particularly at the critical gait event known as Minimum Foot Clearance (MFC), is the major risk factor for tripping-related falls. Intervention strategies to increase MFC height can be effective if applied in real-time based on feed-forward prediction. The current study investigated the capability of machine learning models to classify the MFC into various categories using toe-off kinematics data. Specifically, three MFC sub-categories (less than 1.5 cm, between 1.5 and 2.0 cm, and higher than 2.0 cm) were predicted to apply machine learning approaches. A total of 18,490 swing phase gait cycles’ data were extracted from six healthy young adults, each walking for 5 min at a constant speed of 4 km/h on a motorized treadmill. K-Nearest Neighbor (KNN), Random Forest, and XGBoost were utilized for prediction based on the data from toe-off for five consecutive frames (0.025 s duration). Foot kinematics data were obtained from an inertial measurement unit attached to the mid-foot, recording tri-axial linear accelerations and angular velocities of the local coordinate. KNN, Random Forest, and XGBoost achieved 84%, 86%, and 75% accuracy, respectively, in classifying MFC into the three sub-categories with run times of 0.39 s, 13.98 s, and 170.98 s, respectively. The KNN-based model was found to be more effective if incorporated into an active exoskeleton as the intelligent system to control MFC based on the preceding gait event, i.e., toe-off, due to its quicker computation time. The machine learning-based prediction model shows promise for the prediction of critical MFC data, indicating higher tripping risk.

A high-performance dual-mode energy harvesting with nonlinear pendulum and speed-amplified mechanisms for low-frequency applications

Vibrations induced by human motions and vehicle operations feature low-frequency, broadband, and time-varying. However, enhancing the power density of energy harvesting for various low-frequency applications presents a significant challenge. This paper proposes a high-performance dual-mode energy harvester (DM-EH) incorporating coupled nonlinear pendulum and speed-amplified mechanisms with gears for the simultaneous scavenging vibration and rotation energy. The pendulum serves as the energy capture unit for detecting vibration energy while ensuring the effective operation of the generation unit in rotational environments. The gears amplify the speed of the excitation and enable frequency up-conversion, thereby enhancing the energy conversion. Experimental results substantiate the DM-EH's ability to extract power from vehicle operations at speeds ranging from 60 to 540 rpm, as well as from human motions at frequencies of 0.7–1.5 Hz with 30° amplitudes. In rotation mode, the prototype achieves a maximum average direct current power of 3.79 W, while in vibration mode, it attains 244 mW. The prototype successfully powers portable electronics and supports battery-free triaxial acceleration and temperature multi-sensors during human motions and railway simulation tests. This prototype showcases immense potential as a sustainable power source for portable electronics and self-powered monitoring applications.

High-impact dynamic loading method for calibration of triaxial acceleration sensors

Abstract In this study, a high-impact dynamic loading device was designed to generate a three-dimensional pulse excitation signal with high intensity shock acceleration and achieve triaxial synchronous calibration of a triaxial acceleration sensor. A light-gas gun interior ballistic model and a sensor mechanical response model were developed, and the relationships between the bullet impact velocity, barrel length, and initial chamber pressure were obtained. Additionally, the transformation relationship of the sensor’s triaxial acceleration in different coordinate systems was derived. Based on the stress wave theory and the finite element method, the influence of the bullet impact velocity on the variation pulse, and different slope and deflection angles on the triaxial acceleration were analyzed. By optimizing the parameter design, machining the prototype, and conducting high-impact dynamic loading tests, the results showed that the deviation between theoretical and measured values of the generated triaxial acceleration signal was small, and the maximum deviation was less than 4%. This indicated that the proposed high-impact dynamic loading device satisfied the calibration requirements for calibrating triaxial acceleration sensors, which can generate a three-dimensional acceleration with a peak value of not less than 700 000 m s−2.

Identifying prey capture events of a free-ranging marine predator using bio-logger data and deep learning.

Marine predators are integral to the functioning of marine ecosystems, and their consumption requirements should be integrated into ecosystem-based management policies. However, estimating prey consumption in diving marine predators requires innovative methods as predator-prey interactions are rarely observable. We developed a novel method, validated by animal-borne video, that uses tri-axial acceleration and depth data to quantify prey capture rates in chinstrap penguins (Pygoscelis antarctica). These penguins are important consumers of Antarctic krill (Euphausia superba), a commercially harvested crustacean central to the Southern Ocean food web. We collected a large data set (n = 41 individuals) comprising overlapping video, accelerometer and depth data from foraging penguins. Prey captures were manually identified in videos, and those observations were used in supervised training of two deep learning neural networks (convolutional neural network (CNN) and V-Net). Although the CNN and V-Net architectures and input data pipelines differed, both trained models were able to predict prey captures from new acceleration and depth data (linear regression slope of predictions against video-observed prey captures = 1.13; R 2 ≈ 0.86). Our results illustrate that deep learning algorithms offer a means to process the large quantities of data generated by contemporary bio-logging sensors to robustly estimate prey capture events in diving marine predators.

Overtaking maneuvers on two-lane highways under the microscope: Exploration of a multidimensional framework for the analysis of safety, comfort and efficiency using simulator data

An unreasonable overtaking attempt on two-lane highways could cause drivers to suffer in terms of driving safety, comfort, and efficiency. Several external factors related to the traffic environment (e.g., speed and car type of surrounding vehicles), were found to be the significant factors in drivers’ overtaking performance in the previous studies. However, the microscopic decision-making (e.g., the moments of the occupation of the opposite lane) mechanisms during overtaking, by means of which drivers react to changes in the external traffic environment and adjust their overtaking trajectories, are still need to be explored. Hence, this study had three goals: (i) To explore the spatial characteristics of micro-decisions (MDs) (such as the start and end point) in overtaking trajectories; (ii) To measure three types of performance indicators (i.e., safety, comfort, and efficiency) for the execution of overtaking maneuvers; (iii) To quantitatively explain the microscopic decision-making mechanism in overtaking. Data for overtaking trajectories were collected from driving a simulation experiment where 52 Chinese student drivers completed a series of overtaking maneuvers on a typical two-lane highway under different traffic conditions. Two analyses were conducted: firstly, the distributions of the relative distance between the ego and surrounding vehicles at four key points (i.e., the start, entry, back, and end) in the overtaking trajectory were investigated and clustered to uncover the spatial characteristics of the MDs. Secondly, the safety, comfort, and efficiency of the overtaking were measured by the aggregations of multi-targets collision risks, triaxial acceleration variances, and spatial consumptions respectively based on the Data Envelopment Analysis (DEA), which were further applied in a two-stage SEM model to reveal the quantitative interrelationships among the external factors, microscope decisions and performances in overtaking. We confirmed that the MDs could be considered as the mediating variables between the external factors and overtaking performances. In the presence of the more hazardous traffic environment (e.g., faster traffic flow and impeded by a truck), the safety, comfort and efficiency of overtaking would be deteriorated inevitably. But drivers would execute the overtaking under the longer passing sight distance, migrate their trajectories forward, and shorten the spatial duration to significantly improve the overtaking performances. Based on this mechanism, a overtaking trajectory optimization strategy for the advanced or automatic driving system, was confirmed and concluded that 1) the passing gap should be firstly planned according to the sight distance acceptance of different drivers, which directly determine the upper limit of the safety performance in the overtaking; 2) the trajectory forward migration and shortening the whole duration in overtaking could be effective to enhance the overtaking performances of the overtaking on the two-lane highway; 3) the guidance of the stable control of the steering wheel and gas/brake pedals is essential in the overtaking.

Accurate fall risk classification in elderly using one gait cycle data and machine learning

BackgroundFalls among the elderly are a major societal problem. While observations of medium-distance walking using inertial sensors identified potential fall predictors, classifying individuals at risk based on single gait cycles remains elusive. This challenge stems from individual variability and step-to-step fluctuations, making accurate classification difficult. MethodsWe recruited 44 participants, equally divided into high and low fall-risk groups. A smartphone secured on their second sacral spinous process recorded data during indoor walking. Features were extracted at each gait cycle from a 6-dimensional time series (tri-axial angular velocity and tri-axial acceleration) and classified using the gradient boosting decision tree algorithm. FindingsMean accuracy across five-fold cross-validation was 0.936. “Age” was the most influential individual feature, while features related to acceleration in the gait direction held the highest total relative importance when aggregated by axis (0.5365). InterpretationCombining acceleration, angular velocity data, and the gradient boosting decision tree algorithm enabled accurate fall risk classification in the elderly, previously challenging due to lack of discernible features. We reveal the first-ever identification of three-dimensional pelvic motion characteristics during single gait cycles in the high-risk group. This novel method, requiring only one gait cycle, is valuable for individuals with physical limitations hindering repetitive or long-distance walking or for use in spaces with limited walking areas. Additionally, utilizing readily available smartphones instead of dedicated equipment has potential to improve gait analysis accessibility.