Machine Learning for Recognizing Driving Patterns of Drivers of Large Commercial Trucks

2015 ◽  
Vol 2517 (1) ◽  
pp. 18-27 ◽  
Author(s):  
Chen Chen ◽  
Yuanchang Xie
Keyword(s):  
Machine Learning ◽  
Activity Patterns ◽  
Early Morning ◽  
Crash Risk ◽  
Late Afternoon ◽  
Duty Hours ◽  
Logistic Regression Models ◽  
Complementary Method ◽  
Driving Patterns ◽  
Rest Breaks

Commercial large truck crashes are more likely to involve fatalities and significant costs than passenger vehicle crashes are. To reduce fatigue-related crashes of large trucks caused by drivers' irregular work schedules, FMCSA has enforced hours-of-service rules to regulate the activities of drivers of commercial large trucks. The complex influence of drivers' multiday driving activity patterns on crash risk was examined with data collected from two national truckload carriers. A machine learning approach, k-means clustering, was used to classify large truck drivers into 10 clusters according to their 15-min driving activities over multiple days. Then, the crash risk and driving activity pattern were identified for each cluster. Discrete-time logistic regression models were used to quantify the relationships between driving activity patterns and crash risk. Results indicated that the driving pattern with the lowest crash risk could be daytime driving between 4:00 a.m. and noon, with rest breaks in the late afternoon (4:00 to 6:00 p.m.). Drivers with high proportions of afternoon on-duty time after a long off-duty period experienced significantly higher crash risk. A representative day concept is proposed as a complementary method to identify relationships between driving patterns and crash risk. Moreover, on-duty hours can be a useful indicator of crash risk for drivers of large trucks. High proportions of on-duty hours in the early morning and late afternoon often are associated with high crash risk.

scholarly journals Thermal ecology and activity patterns of the lizard community of the Restinga of Jurubatiba, Macaé, RJ

2001 ◽  
Vol 61 (2) ◽  
pp. 287-294 ◽  
Author(s):  
F. H. HATANO ◽  
D. VRCIBRADIC ◽  
C. A. B. GALDINO ◽  
M. CUNHA-BARROS ◽  
C. F. D. ROCHA ◽  
...  
Keyword(s):  
Body Temperature ◽  
Rio De Janeiro ◽  
Activity Patterns ◽  
Early Morning ◽  
Thermal Ecology ◽  
Congeneric Species ◽  
Late Afternoon ◽  
Hemidactylus Mabouia ◽  
Environmental Temperatures

We analyzed the thermal ecology and activity patterns of the lizard community from the Restinga of Jurubatiba, Rio de Janeiro, Brazil. The broadest activity was that of Tropidurus torquatus, a sit-and-wait forager, while the active foraging teiid Cnemidophorus littoralis had the shortest activity. The nocturnal gekkonid Hemidactylus mabouia was found active during the day only during early morning and late afternoon, when environmental temperatures are low. Body temperature was highest for Cnemidophorus littoralis and lowest for the two Mabuya species. The patterns found here are discussed and compared to those of congeneric species in other habitats in Brazil.

scholarly journals Physical activity around the clock: objectively measured activity patterns in young children of the GECKO Drenthe cohort

2019 ◽  
Vol 19 (1) ◽  
Author(s):  
Rikstje Wiersma ◽  
Congchao Lu ◽  
Esther Hartman ◽  
Eva Corpeleijn
Keyword(s):  
Physical Activity ◽  
Young Children ◽  
Activity Patterns ◽  
Early Morning ◽  
Time Segment ◽  
Physically Active ◽  
Late Afternoon ◽  
Preventive Child Healthcare ◽  
Measured Activity ◽  
Definition Of

Abstract Background Given the widespread problem of physical inactivity, and the continued growth in prevalence of childhood and adolescent obesity, promotion of regular physical activity (PA) among young people has become a public priority. A greater understanding of children’s PA patterns throughout the day is needed to effectively encourage children to be more physically active. Hence this study looking at the distribution of PA in young children throughout the day and its relevance to overweight. Methods Accelerometers (ActiGraph GT3X, weartime > 600 min/day, ≥3 days) were used to measure the PA of 958 children (aged 5.7 ± 0.8 years, 52% boys) enrolled in the GECKO Drenthe cohort. Levels of sedentary time (ST), light PA (LPA) and moderate-to-vigorous PA (MVPA) were recorded throughout the day and analysed in segments (07:00–09:00, 09:00–12:00, 12:00–15:00, 15:00–18:00, 18:00–21:00). Body mass index was measured by Preventive Child Healthcare nurses and Cole’s (2012) definition of overweight was used. General linear mixed models, adjusted for age, sex and season, were used to analyse patterns of PA and ST throughout the day. Results Children were most sedentary in the early morning (07:00–09:00) and evening (18:00–21:00), and exhibited the most time spent engaged in LPA and MVPA in the afternoon (12:00–15:00) and late afternoon (15:00–18:00). The greatest inter-individual variation in ST, LPA and MVPA among the children occurred in the late afternoon and evening (approximately 40, 30 and 15 min difference per time segment between 25th and 75th percentile, respectively). The most active children (highest quartile of MVPA) were found to be more active and less sedentary throughout the entire day than the least active children (lowest quartile of MVPA). Furthermore, children with overweight were no less active than children without overweight. Conclusions At this young age, the relevance of different PA patterns to childhood overweight was minimal. Children were most active in the afternoon and late afternoon. To encourage PA in general, ST can be reduced and PA increased in the early morning and evening. Targeted PA interventions to specifically stimulate the least active children could take place in the late afternoon or evening.

scholarly journals Meals for two: foraging activity of the butterflyfish Chaetodon Striatus (perciformes) in southeast Brazil

2005 ◽  
Vol 65 (2) ◽  
pp. 211-215 ◽  
Author(s):  
R. M. Bonaldo ◽  
J. P. Krajewski ◽  
I. Sazima
Keyword(s):  
Benthic Invertebrates ◽  
Early Morning ◽  
Southeastern Brazil ◽  
Foraging Activity ◽  
Feeding Rates ◽  
Western Atlantic ◽  
Late Afternoon ◽  
Sea Floor ◽  
Rocky Reefs ◽  
Feeding Activities

The banded butterflyfish (Chaetodon striatus) from the tropical and subtropical western Atlantic is a territorial, diurnal forager on benthic invertebrates. It is usually seen moving singly or in pairs, a few meters above the sea floor. We studied the foraging activity of C. striatus on rocky reefs in southeastern Brazil. This fish spent about 11 h and 30 min per day on feeding activities, and preferred colonies of non-scleratinian anthozoans over sandy and rocky substrata while foraging. The lowest feeding rates were recorded in the early morning and late afternoon, but we found no further differences between feeding rates throughout the day. We also found no differences between the feeding rates of paired and single individuals.

scholarly journals Seasonal Variation of the Diurnal Cycle of Rainfall in Southern Contiguous China

2008 ◽  
Vol 21 (22) ◽  
pp. 6036-6043 ◽  
Author(s):  
Jian Li ◽  
Rucong Yu ◽  
Tianjun Zhou
Keyword(s):  
Seasonal Variation ◽  
Diurnal Cycle ◽  
Radiative Forcing ◽  
Early Morning ◽  
Rain Gauge ◽  
Rainfall Events ◽  
Late Afternoon ◽  
Long Duration ◽  
China Region ◽  
Afternoon Peak

Abstract Hourly station rain gauge data are employed to study the seasonal variation of the diurnal cycle of rainfall in southern contiguous China. The results show a robust seasonal variation of the rainfall diurnal cycle, which is dependent both on region and duration. Difference in the diurnal cycle of rainfall is found in the following two neighboring regions: southwestern China (region A) and southeastern contiguous China (region B). The diurnal cycle of annual mean precipitation in region A tends to reach the maximum in either midnight or early morning, while precipitation in region B has a late-afternoon peak. In contrast with the weak seasonal variation of the diurnal phases of precipitation in region A, the rainfall peak in region B shifts sharply from late afternoon in warm seasons to early morning in cold seasons. Rainfall events in south China are classified into short- (1–3 h) and long-duration (more than 6 h) events. Short-duration precipitation in both regions reaches the maximum in late afternoon in warm seasons and peaks in either midnight or early morning in cold seasons, but the late-afternoon peak in region B exists during February–October, while that in region A only exists during May–September. More distinct differences between regions A and B are found in the long-duration rainfall events. The long-duration events in region A show dominant midnight or early morning peaks in all seasons. But in region B, the late-afternoon peak exists during July–September. Possible reasons for the difference in the diurnal cycle of rainfall between the two regions are discussed. The different cloud radiative forcing over regions A and B might contribute to this difference.

45 Application of Machine Learning Models to Thermal Burn Patient Outcome Predictions in the Aftermath of a Nuclear Event

2021 ◽  
Vol 42 (Supplement_1) ◽  
pp. S33-S34
Author(s):  
Morgan A Taylor ◽  
Randy D Kearns ◽  
Jeffrey E Carter ◽  
Mark H Ebell ◽  
Curt A Harris
Keyword(s):  
Machine Learning ◽  
Logistic Regression ◽  
Length Of Stay ◽  
Regression Models ◽  
Large Scale ◽  
Prediction Models ◽  
Burn Patients ◽  
Thermal Burn ◽  
Logistic Regression Models ◽  
Burn Patient

Abstract Introduction A nuclear disaster would generate an unprecedented volume of thermal burn patients from the explosion and subsequent mass fires (Figure 1). Prediction models characterizing outcomes for these patients may better equip healthcare providers and other responders to manage large scale nuclear events. Logistic regression models have traditionally been employed to develop prediction scores for mortality of all burn patients. However, other healthcare disciplines have increasingly transitioned to machine learning (ML) models, which are automatically generated and continually improved, potentially increasing predictive accuracy. Preliminary research suggests ML models can predict burn patient mortality more accurately than commonly used prediction scores. The purpose of this study is to examine the efficacy of various ML methods in assessing thermal burn patient mortality and length of stay in burn centers. Methods This retrospective study identified patients with fire/flame burn etiologies in the National Burn Repository between the years 2009 – 2018. Patients were randomly partitioned into a 67%/33% split for training and validation. A random forest model (RF) and an artificial neural network (ANN) were then constructed for each outcome, mortality and length of stay. These models were then compared to logistic regression models and previously developed prediction tools with similar outcomes using a combination of classification and regression metrics. Results During the study period, 82,404 burn patients with a thermal etiology were identified in the analysis. The ANN models will likely tend to overfit the data, which can be resolved by ending the model training early or adding additional regularization parameters. Further exploration of the advantages and limitations of these models is forthcoming as metric analyses become available. Conclusions In this proof-of-concept study, we anticipate that at least one ML model will predict the targeted outcomes of thermal burn patient mortality and length of stay as judged by the fidelity with which it matches the logistic regression analysis. These advancements can then help disaster preparedness programs consider resource limitations during catastrophic incidents resulting in burn injuries.

Implementation of Machine Learning Algorithms for Prediction of Fluidelastic Instability in Tube Arrays

2021 ◽  
Vol 143 (2) ◽  
Author(s):  
Joaquin E. Moran ◽  
Yasser Selima
Keyword(s):  
Machine Learning ◽  
Logistic Regression ◽  
Learning Algorithm ◽  
Machine Learning Algorithms ◽  
Two Phase ◽  
Factors Affecting ◽  
Logistic Regression Models ◽  
Number Of Factors ◽  
Tube Arrays ◽  
Fluidelastic Instability

Abstract Fluidelastic instability (FEI) in tube arrays has been studied extensively experimentally and theoretically for the last 50 years, due to its potential to cause significant damage in short periods. Incidents similar to those observed at San Onofre Nuclear Generating Station indicate that the problem is not yet fully understood, probably due to the large number of factors affecting the phenomenon. In this study, a new approach for the analysis and interpretation of FEI data using machine learning (ML) algorithms is explored. FEI data for both single and two-phase flows have been collected from the literature and utilized for training a machine learning algorithm in order to either provide estimates of the reduced velocity (single and two-phase) or indicate if the bundle is stable or unstable under certain conditions (two-phase). The analysis included the use of logistic regression as a classification algorithm for two-phase flow problems to determine if specific conditions produce a stable or unstable response. The results of this study provide some insight into the capability and potential of logistic regression models to analyze FEI if appropriate quantities of experimental data are available.

scholarly journals Replay in minds and machines

2021 ◽  
Author(s):  
Lennart Wittkuhn ◽  
Samson Chien ◽  
Sam Hall-McMaster ◽  
Nicolas W. Schuck
Keyword(s):  
Machine Learning ◽  
Brain Activity ◽  
Activity Patterns ◽  
Representation Learning ◽  
Parallel Machine ◽  
Artificial Agents ◽  
Active Behavior ◽  
Performance Improvements ◽  
Experience Replay ◽  
Data Efficiency

Experience-related brain activity patterns have been found to reactivate during sleep, wakeful rest, and brief pauses from active behavior. In parallel, machine learning research has found that experience replay can lead to substantial performance improvements in artificial agents. Together, these lines of research have significantly expanded our understanding of the potential computational benefits replay may provide to biological and artificial agents alike. We provide an overview of findings in replay research from neuroscience and machine learning and summarize the computational benefits an agent can gain from replay that cannot be achieved through direct interactions with the world itself. These benefits include faster learning and data efficiency, less forgetting, prioritizing important experiences, as well as improved planning and generalization. In addition to the benefits of replay for improving an agent's decision-making policy, we highlight the less-well studied aspect of replay in representation learning, wherein replay could provide a mechanism to learn the structure and relevant aspects of the environment. Thus, replay might help the agent to build task-appropriate state representations.

Spot on: using camera traps to individually monitor one of the world’s largest lizards

2020 ◽  
Vol 47 (4) ◽  
pp. 326 ◽  
Author(s):  
Harry A. Moore ◽  
Jacob L. Champney ◽  
Judy A. Dunlop ◽  
Leonie E. Valentine ◽  
Dale G. Nimmo
Keyword(s):  
Activity Patterns ◽  
Image Data ◽  
Dorsal Surface ◽  
Camera Trap ◽  
Camera Traps ◽  
Identification Accuracy ◽  
Individual Identification ◽  
Lizard Species ◽  
Late Afternoon ◽  
Large Animals

Abstract ContextEstimating animal abundance often relies on being able to identify individuals; however, this can be challenging, especially when applied to large animals that are difficult to trap and handle. Camera traps have provided a non-invasive alternative by using natural markings to individually identify animals within image data. Although camera traps have been used to individually identify mammals, they are yet to be widely applied to other taxa, such as reptiles. AimsWe assessed the capacity of camera traps to provide images that allow for individual identification of the world’s fourth-largest lizard species, the perentie (Varanus giganteus), and demonstrate other basic morphological and behavioural data that can be gleaned from camera-trap images. MethodsVertically orientated cameras were deployed at 115 sites across a 10000km2 area in north-western Australia for an average of 216 days. We used spot patterning located on the dorsal surface of perenties to identify individuals from camera-trap imagery, with the assistance of freely available spot ID software. We also measured snout-to-vent length (SVL) by using image-analysis software, and collected image time-stamp data to analyse temporal activity patterns. ResultsNinety-two individuals were identified, and individuals were recorded moving distances of up to 1975m. Confidence in identification accuracy was generally high (91%), and estimated SVL measurements varied by an average of 6.7% (min=1.8%, max=21.3%) of individual SVL averages. Larger perenties (SVL of >45cm) were detected mostly between dawn and noon, and in the late afternoon and early evening, whereas small perenties (SVL of <30cm) were rarely recorded in the evening. ConclusionsCamera traps can be used to individually identify large reptiles with unique markings, and can also provide data on movement, morphology and temporal activity. Accounting for uneven substrates under cameras could improve the accuracy of morphological estimates. Given that camera traps struggle to detect small, nocturnal reptiles, further research is required to examine whether cameras miss smaller individuals in the late afternoon and evening. ImplicationsCamera traps are increasingly being used to monitor reptile species. The ability to individually identify animals provides another tool for herpetological research worldwide.

scholarly journals Analysis of Machine Learning Techniques Applied to Sensory Detection of Vehicles in Intelligent Crosswalks

Sensors ◽  
2020 ◽  
Vol 20 (21) ◽  
pp. 6019
Author(s):  
José Manuel Lozano Domínguez ◽  
Faroq Al-Tam ◽  
Tomás de J. Mateo Sanguino ◽  
Noélia Correia
Keyword(s):  
Machine Learning ◽  
Smart Cities ◽  
Machine Learning Techniques ◽  
Support Vector ◽  
Learning Models ◽  
Fuzzy Classifier ◽  
Logistic Regression Models ◽  
The Road ◽  
Learning Agent ◽  
Machine Learning Models

Improving road safety through artificial intelligence-based systems is now crucial turning smart cities into a reality. Under this highly relevant and extensive heading, an approach is proposed to improve vehicle detection in smart crosswalks using machine learning models. Contrarily to classic fuzzy classifiers, machine learning models do not require the readjustment of labels that depend on the location of the system and the road conditions. Several machine learning models were trained and tested using real traffic data taken from urban scenarios in both Portugal and Spain. These include random forest, time-series forecasting, multi-layer perceptron, support vector machine, and logistic regression models. A deep reinforcement learning agent, based on a state-of-the-art double-deep recurrent Q-network, is also designed and compared with the machine learning models just mentioned. Results show that the machine learning models can efficiently replace the classic fuzzy classifier.

Sign in / Sign up

Export Citation Format

Share Document