Exploiting the use of recurrent neural networks for driver behavior profiling

Carvalho, Eduardo Atem de; Ferreira, Bruno V.; Ferreira, Jair; Souza, Cleidson de; Carvalho, Hanna V.; Suhara, Yoshihiko; Pentland, Alex; Pessin, Gustavo

doi:10.1109/ijcnn.2017.7966230

Cited by 54 publications

(42 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Furthermore, Jozefowicz and colleagues 21 showed that for a great class of problems, GRU outperformed LSTM. Their study is also corroborated by Carvalho and colleagues 22 , were GRU units showed lower dispersion than LSTM on the results.…”

supporting

confidence: 72%

An Amazon stingless bee foraging activity predicted using recurrent artificial neural networks and attribute selection

Gomes

Suhara

Nunes‐Silva

et al. 2020

Sci Rep

Self Cite

View full text Add to dashboard Cite

show abstract

supporting

confidence: 72%

An Amazon stingless bee foraging activity predicted using recurrent artificial neural networks and attribute selection

Gomes

Suhara

Nunes‐Silva

et al. 2020

Sci Rep

Self Cite

View full text Add to dashboard Cite

show abstract

“…To verify the feasibility and effectiveness of the proposed driving error detection framework, we tested it on a high‐quality open dataset of driving errors created by Ferreira Júnior and utilized in many similar studies (Carvalho et al, ; Júnior et al, ). This dataset contains smartphone sensor measurements captured while performing seven different types of driving events.…”

Section: Experiments and Resultsmentioning

confidence: 99%

Enhancing driving safety: Discovering individualized hazardous driving scenes using GIS and mobile sensing

Goldberg

Chu

et al. 2019

Transactions in GIS

View full text Add to dashboard Cite

Traffic collisions have been well acknowledged as a significant threat to public health, closely related to human driving errors. This study introduces an innovative approach to investigate spatiotemporal distributions of individualized driving errors and to characterize hazardous driving scenes, in which drivers are more prone to make driving mistakes. We first create a multi‐feature‐fusion framework to extract driving errors using smartphone sensors. Then, the detected errors are geo‐statistically analyzed with road networks and driving trajectories to identify driving error hotspots. We next construct a “scenic tuple” for representing the occurrence of driving errors. Finally, the individualized hazardous driving scenes are extracted by mining a long‐term collection of scenic tuples. Results demonstrate that our proposed approach can effectively identify driving errors. Additionally, the spatiotemporal patterns of driving mistakes can be identified from the individualized hazardous driving scenes, which has the potential to aid in reducing driving risks.

show abstract

“…However, with larger databases, it can be expected that Deep Learning techniques offer higher recognition results. In fact, in a later work by the same authors (Carvalho et al [ 2 ]) they use the same database to apply Deep Learning and to compare the benefits with different RNN schemes. In their empirical evaluation, the gated recurrent unit (GRU) was the most reliable RNN to be deployed with the accelerometer data; the long short-term memory (LSTM) and the simple RNN have a greater difference depending on the numbers of neurons.…”

Section: Related Workmentioning

confidence: 99%

“…We must point out that the estimation of this principal direction of vehicle movement can be considered equivalent to a calibration process, which can be found in many driving characterization research papers, to map phone coordinates into vehicle reference coordinates. Carvalho et al [ 2 ] and Lu et al [ 3 ] perform this transformation from a smartphone’s coordinate system to a vehicle’s coordinate system. Also, it is mentioned in Kanarachos et al [ 4 ], where this reorientation is carried out by fusing the accelerometer, the gyroscope and the magnetometer signals and calculating the Euler rotation angles, or only with the accelerometer and magnetometer.…”

Section: Introductionmentioning

confidence: 99%

Estimating Vehicle Movement Direction from Smartphone Accelerometers Using Deep Neural Networks

Sánchez

Pozo

Gómez

2018

Sensors

View full text Add to dashboard Cite

Characterization of driving maneuvers or driving styles through motion sensors has become a field of great interest. Before now, this characterization used to be carried out with signals coming from extra equipment installed inside the vehicle, such as On-Board Diagnostic (OBD) devices or sensors in pedals. Nowadays, with the evolution and scope of smartphones, these have become the devices for recording mobile signals in many driving characterization applications. Normally multiple available sensors are used, such as accelerometers, gyroscopes, magnetometers or the Global Positioning System (GPS). However, using sensors such as GPS increase significantly battery consumption and, additionally, many current phones do not include gyroscopes. Therefore, we propose the characterization of driving style through only the use of smartphone accelerometers. We propose a deep neural network (DNN) architecture that combines convolutional and recurrent networks to estimate the vehicle movement direction (VMD), which is the forward movement directional vector captured in a phone’s coordinates. Once VMD is obtained, multiple applications such as characterizing driving styles or detecting dangerous events can be developed. In the development of the proposed DNN architecture, two different methods are compared. The first one is based on the detection and classification of significant acceleration driving forces, while the second one relies on longitudinal and transversal signals derived from the raw accelerometers. The final success rate of VMD estimation for the best method is of 90.07%.

show abstract

Exploiting the use of recurrent neural networks for driver behavior profiling

Cited by 54 publications

References 14 publications

An Amazon stingless bee foraging activity predicted using recurrent artificial neural networks and attribute selection

An Amazon stingless bee foraging activity predicted using recurrent artificial neural networks and attribute selection

Enhancing driving safety: Discovering individualized hazardous driving scenes using GIS and mobile sensing

Estimating Vehicle Movement Direction from Smartphone Accelerometers Using Deep Neural Networks

Contact Info

Product

Resources

About