Driver Profiling Using Long Short Term Memory (LSTM) and Convolutional Neural Network (CNN) Methods

Cura, Aslihan; Küçük, Halûk; Ergen, Erdem; Oksuzoglu, Ismail Burak

doi:10.1109/tits.2020.2995722

Cited by 48 publications

(15 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For identifying specific driving style, Boyce et al have summarized the similar characteristics of radical drivers and have pointed out that young drivers often have a radical style and are prone to speeding violations [93]; Musselwhite have also made a detailed classification for radical drivers. They have combined driving with the environment and given four sub-dimensions: unintentional adventure, environmental feedback adventure, computational adventure, and continuous adventure [94]; Cura et al have developed NN models based on LSTM and CNN respectively, focusing on acceleration, deceleration, and lane change behavior, and have conducted more detailed style analysis on aggressive bus drivers [95]; Gatteschi et al have conducted a further subdivision study on the aggressiveness of driving behavior to non-motor vehicles [96]. For classification methodologies of driving styles, Abdelrahman et al have proposed a framework for calculating driver risk status judgment based on baseline driving events and risk probability prediction, and have realized effective prediction of driving risk level with the help of machine learning (ML) model [97]; also based on ML algorithms, Yuksel et al have proposed a high-precision and low-cost driver risk assessment black box system based on the driving data that can be obtained from accelerometers and gyroscope sensors and have specifically pointed out that the system can be used in vehicle UBI products [98].…”

Section: Research On Differentiated Macro Driving Stylementioning

confidence: 99%

A Comprehensive Review of Driving Style Evaluation Approaches and Product Designs Applied to Vehicle Usage-Based Insurance

Nai¹,

Yang²,

Wei³

et al. 2022

Sustainability

View full text Add to dashboard Cite

Vehicle insurance is a very important source of income for insurance companies, and it is closely related to the driving style performed by driving behavior. Different driving styles can better reflect the driving risk than the number of violations, claims, and other static statistic data. Subdivide the vehicle insurance market according to the personal characteristics and driving habits of the insured vehicles, and studying the personalized vehicle insurance products, will help the insurance companies to improve their income, help the drivers to change their bad driving habits, and thus help to realize the healthy development of the vehicle insurance industry. In the past 20 to 30 years, more and more insurance companies around the world have launched vehicle usage-based insurance (UBI) products based on driving style analysis. However, up to now, there are few comprehensive reports on commercial vehicle UBI products and their core driving risk assessment methods. On the basis of literature indexing on the Web of Science and other academic platforms by using the keywords involved in vehicle UBI, over 100 relevant works of literature were screened in this paper, and a detailed and comprehensive discussion on the driving style evaluation methods and the design of commercial vehicle UBI products during the past 20 to 30 years has been made, hoping to get a full understanding of the possible factors affecting driving style and the collectible data that can reflect these factors, and to get a full grasp of the developing status, challenges and future trends in vehicle insurance branch of the Internet of Vehicles (IoV) industry.

show abstract

Section: Research On Differentiated Macro Driving Stylementioning

confidence: 99%

A Comprehensive Review of Driving Style Evaluation Approaches and Product Designs Applied to Vehicle Usage-Based Insurance

Nai¹,

Yang²,

Wei³

et al. 2022

Sustainability

View full text Add to dashboard Cite

show abstract

“…where h t is the summary of input xt and the past hidden layer state ht-1, U (z) , W (z) , U (r) , W (r) , U and W is a trainable parameter matrix. However, both the GRU and the LSTM network can only consider the information of the past time of the prediction point, but cannot consider the state of the future time, so the prediction accuracy cannot be further improved [23][24]. BiGRU adds a hidden layer on the basis of bidirectional GRU neural network, divides the prediction process into two directions: forward prediction and backward prediction, and determines the output result jointly by the hidden layer of the two directions.…”

Section: Figure 2 a Structure Of Gru Neural Network Circulationmentioning

confidence: 99%

Multi-Model Fusion Short-Term Load Forecasting Based on Random Forest Feature Selection and Hybrid Neural Network

Zhu

et al. 2021

IEEE Access

View full text Add to dashboard Cite

In an increasingly open electricity market environment, short-term load forecasting (STLF) can ensure the power grid to operate safely and stably, reduce resource waste, power dispatching, and provide technical support for demand-side response. Recently, with the rapid development of demand side response, accurate load forecasting can better provide demand side incentive for regional load of prosumer groups. Traditional machine learning prediction and time series prediction based on statistics failed to consider the non-linear relationship between various input features, resulting in the inability to accurately predict load changes. Recently, with the rapid development of deep learning, extensive research has been carried out in the field of load forecasting. On this basis, a feature selection algorithm based on random forest is first used in this paper to provide a basis for the selection of the input features of the load forecasting model. After the input features are selected, a hybrid neural network STLF algorithm based on multi-model fusion is proposed, of which the main structure of the hybrid neural network is composed of convolutional neural network and bidirectional gated recurrent unit (CNN-BiGRU). The input data is obtained by using long sliding time windows of different steps, then multiple CNN-BiGRU models are trained respectively. The forecasting results of multiple models are averaged to get the final forecasting load value. The load datasets come from a region in New Zealand and a region in Zhejiang, China, are used as load forecast examples. Finally, a variety of load forecasting algorithms are introduced for comparison. The experimental results show that our method has a higher accuracy than comparison models.

show abstract

“…Furthermore, the increase of vehicle fuel efficiency has become a key research field due to its substantial impact on the usage of fossil fuel and global carbon repercussions [86]. In the evaluation of driver behavior, various control systems, such as controller area network (CAN bus) of public transports or LSTM deep learning predictor controller, cloud computing platform, and sensors have been developed to categorize and evaluate the eco-driving performance of bus drivers and to characterize driver behavior by engine speed, deceleration, pedaling, corner turn, and lane change endeavors [72,73,103].…”

Section: ) Benefits Related To Increasing Energy Conservation Awareness Of Drivermentioning

confidence: 99%

A Landscape of Research on Bus Driver Behavior: Taxonomy, Open Challenges, Motivations, Recommendations, Limitations, and Pathways Solution in Future

et al. 2021

View full text Add to dashboard Cite

Driver behavior is a concerning issue in the area of intelligent transportation system (ITS). Driver behavior is a significant key player in a wide range of unpleasant events during the ride, such as accidents or crashes, traffic congestion, harsh braking, and acceleration/deceleration. Influencing factors of driver behavior have been explored in several studies. It is imperative to investigate these factors in order to provide a comprehensive analysis and to categorize them on the basis of a coherent taxonomy. With that, this study conducted a systematic review on prior studies that focused on bus driver behavior, particularly in the ITS. This study also established a taxonomy on the topic of driver behavior in multiple areas of ITS and their classifications. Different databases, namely ScienceDirect, Web of Science, and IEEE Explore, were utilized to obtain relevant articles from 2008 to 2021 (15 April). Several filtering and scanning stages were performed according to the exclusion/inclusion criteria on all 2,803 articles obtained; however, only 87 articles met the criteria. The final set of articles were categorized into a taxonomy. The first part of the taxonomy focuses on five main factors that influence driver behavior: environmental, demographic, habit, vehicle, and on-road routine factors. The second part of the taxonomy discusses the mapping of data collection methods on the basis of four categories: real-time data collection, survey, simulation, and benchmark. Discussion and analysis were provided to highlight the critical literature gaps on bus driver behavior in the ITS, involving the use of real-time data collection, which is imperative for acquiring highly accurate and sophisticated data. This multi-field systematic review has exposed new research opportunities, motivations, challenges, limitations, and recommendations and highlighted the need for the synergistic integration of interdisciplinary works. Overall, this study presented pathways solution in future direction on the basis of three sequenced phases, namely design, labeling and validation, and machine learning. This study can serve as a guide for future researchers, as it addressed the ambiguities in the ITS-driver behavior domain and provided valuable information on these ITS-driver behavior trends.

show abstract

Driver Profiling Using Long Short Term Memory (LSTM) and Convolutional Neural Network (CNN) Methods

Cited by 48 publications

References 32 publications

A Comprehensive Review of Driving Style Evaluation Approaches and Product Designs Applied to Vehicle Usage-Based Insurance

A Comprehensive Review of Driving Style Evaluation Approaches and Product Designs Applied to Vehicle Usage-Based Insurance

Multi-Model Fusion Short-Term Load Forecasting Based on Random Forest Feature Selection and Hybrid Neural Network

A Landscape of Research on Bus Driver Behavior: Taxonomy, Open Challenges, Motivations, Recommendations, Limitations, and Pathways Solution in Future

Contact Info

Product

Resources

About