Driving Maneuvers Prediction Based on Cognition-driven and Data-driven Method

Zhou, Dong; Ma, Huimin; Dong, Yuhan

doi:10.1109/vcip.2018.8698695

Cited by 11 publications

(18 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The number of valid video samples for training the whole framework relatively to the covered time period before a maneuver is shown in Table III. We use a 5-fold cross-validation for all the experiments in this work, which also aligns with other previous works using the Brain4cars dataset [1], [9], [10], [11], [12].…”

Section: B Maneuver Anticipation Frameworkmentioning

confidence: 87%

“…As previously mentioned, there are multiple works aiming at the driver maneuver anticipation [1], [9], [10], [11], [12]. However, none of the previous work solved driver intention prediction with information from both video (in and out of the car) streams, since the traffic on road is too complex for hand-crafting explicit features.…”

Section: Related Workmentioning

confidence: 99%

“…However, none of the previous work solved driver intention prediction with information from both video (in and out of the car) streams, since the traffic on road is too complex for hand-crafting explicit features. Therefore, several works, such as [1], [9], [10], [12], use manual-encoded feature vectors. On the other hand, training CNNs with outside videos in an end-to-end fashion did not show satisfactory results [11], since there was not enough on-road video data related to maneuver anticipation for training a CNN-based deep network.…”

Section: Related Workmentioning

confidence: 99%

“…Images convey massive information, and much of the literature shows the possibility to predict driver intention according to the drivers' videos, since the drivers turn their heads to glance in the side mirrors. Previous work based on the Brain4cars dataset, such as [1], [9], [10], [11], [12], have all achieved maneuver prediction. Although the reported results are quite impressive, there are still some issues that deserve scrutiny.…”

mentioning

confidence: 99%

“…Additionally, video frames of driver observations are used to extract features e.g. head postures [1], [9], [10], [12]. However, in these works, the traffic information is manually encoded into a vector with four elements, where the first two Boolean values indicate whether a lane exists on the right or left side of the vehicle, the third bit (also Boolean) implies if an intersection or turn exists in 15 meters, and the last value represents the current speed of the car.…”

mentioning

confidence: 99%

See 4 more Smart Citations

Driver Intention Anticipation Based on In-Cabin and Driving Scene Monitoring

Akata

Kasneci

2020

Preprint

View full text Add to dashboard Cite

Numerous car accidents are caused by improper driving maneuvers. Serious injuries are however avoidable, if such driving maneuvers are detected beforehand and the driver is assisted accordingly. In fact, various recent research has focused on the automated prediction of driving maneuver based on handcrafted features extracted mainly from in-cabin driver videos. Since the outside view from the traffic scene may also contain informative features for driving maneuver prediction, we present a framework for the detection of the drivers' intention based on both in-cabin and traffic scene videos. More specifically, we (1) propose a Convolutional-LSTM (ConvLSTM)-based autoencoder to extract motion features from the out-cabin traffic, (2) train a classifier which considers motions from both in-and outside of the cabin jointly for maneuver intention anticipation, (3) experimentally prove that the in-and outside image features have complementary information. Our evaluation based on the publicly available dataset Brain4cars shows that our framework achieves a prediction with the accuracy of 83.98% and F1-score of 84.3%. I. INTRODUCTIONAccording to the World Health Organization [2], about 1.35 million people die in car accidents every year worldwide. These statistics, however, do not include non-fatal injuries from traffic accidents. Most of these accidents are caused by improper driver behavior: Based on the statistics from the Department for Transport (DfT) in Great Britain, a survey [6] revealed that there were 15,560 accidents reported due to poor turn or maneuver, which ranked top 5 in causes of road accidents in 2017. As automated vehicle technology emerges, it promised to be safer than human driving [3], [4], [5]. However, there is still much research to be conducted in order to reach to the fully automated level working at any possible traffic situation and weather conditions. On the half way to autonomous driving vehicles, it is therefore necessary to provide already existing Advanced Driver Assistance Systems (ADAS) the functionality for collaboration with the human driver in the most efficient way, for example to alert the driver in case of a dangerous maneuver.Recently, many researchers focused on detecting maneuver intention of the driver before execution. For example, Brain4cars [1] and Honda Research Institute Driving Dataset (HDD) [7] are two datasets specifically designed for learning driver behaviors. HDD for example [7] uses three highresolution video cameras, GPS, signals from LiDAR sensor and vehicle CAN-Bus to record the traffic scenes. Brain4cars[1] provides videos from inside and outside of the car. GPS and vehicle dynamics are also recorded with the videos.

show abstract

Section: B Maneuver Anticipation Frameworkmentioning

confidence: 87%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

mentioning

confidence: 99%

mentioning

confidence: 99%

See 3 more Smart Citations

Driver Intention Anticipation Based on In-Cabin and Driving Scene Monitoring

Akata

Kasneci

2020

Preprint

View full text Add to dashboard Cite

show abstract

Driver intention prediction based on multi-dimensional cross-modality information interaction

Xue,

Xu,

Qiao

et al. 2024

Multimedia Systems

View full text Add to dashboard Cite

Driver intention prediction allow drivers to perceive possible dangers in the fastest time and has become one of the most important research topics in the field of self-driving in recent years. In this study, we propose a driver intention prediction method based on multi-dimensional crossmodality information interaction. First, an efficient video recognition network is designed to extract channel-temporal features of in-side (driver) and out-side (road) videos respectively, in which we design a cross-modality channel-spatial weight mechanism to achieve information interaction between the two feature extraction networks corresponding respectively to the two modalities, and we also introduce a contrastive learning module by which we force the two feature extraction networks to enhance structural knowledge interaction. Then, the obtained representations of in-and out-side videos are fused using a Res-Layer based module to get a preliminary prediction which is then corrected by incorporating the GPS information to obtain a final decision. Besides, we employ a multi-task framework to train the entire network. We validate the proposed method on the public dataset Brain4Car, and the results show that the proposed method achieves competitive results in accuracy while balancing performance and computation.

show abstract

DIPNet: Driver intention prediction for a safe takeover transition in autonomous vehicles

Bonyani

Rahmanian

Jahangard

et al. 2023

IET Intelligent Trans Sys

View full text Add to dashboard Cite

Following the successful development of advanced driver assistance systems (ADAS), the current research directions focus on highely automated vehicles aiming at reducing human driving tasks, and extending the operational design domain, while maintaining a higher level of safety. Currently, there are high research demands in academia and industry to predict driver intention and understating driver readiness, e.g. in response to a “take‐over request” when a transition from automated driving mode to human mode is needed. A driver intention prediction system can assess the driver's readiness for a safe takeover transition. In this study, a novel deep neural network framework is developed by adopting and adapting the DenseNet, long short‐term memory, attention, FlowNet2, and RAFT models to anticipate the diver maneuver intention. Using the public “Brain4Cars” dataset, the driver maneuver intention will be predicted up to 4 s in advance, before the commencement of the driver's action. The driver intention prediction is assessed based on 1) in‐cabin 2) out‐cabin (road) and 3) both in‐out cabin video data. Utilizing K‐fold cross‐validation, the performance of the model is evaluated using accuracy, precision, recall, and F1‐score metrics. The experiments show the proposed DIPNet model outperforms the state‐of‐the‐art in the majority of the driving scenarios.

show abstract

Driving Maneuvers Prediction Based on Cognition-driven and Data-driven Method

Cited by 11 publications

References 14 publications

Driver Intention Anticipation Based on In-Cabin and Driving Scene Monitoring

Driver Intention Anticipation Based on In-Cabin and Driving Scene Monitoring

Driver intention prediction based on multi-dimensional cross-modality information interaction

DIPNet: Driver intention prediction for a safe takeover transition in autonomous vehicles

Contact Info

Product

Resources

About