Design of a Semi-Supervised Learning Strategy based on Convolutional Neural Network for Vehicle Maneuver Classification

Mammeri, Abdelhamid; Zhao, Yiheng; Boukerche, Azzedine; Siddiqui, Abdul Jabbar; Pekilis, Barry

doi:10.1109/wisee.2019.8920301

Cited by 7 publications

(8 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hence, we investigate a structure of stacked autoencoder that has a minimum number of layers and a minimum number of units in each layer to learn the hidden representation. We add 10 units in the encoded layer so that it will be the size of the feature vector of the supervised model and need less time during the supervised Though there are a few work done by adopting semisupervised approach, if we do a comparative analysis of [37] and [42] with the performance of our proposed Semi-Supervised method none of them have provided accuracy and error of each class. Hence, we can not do any direct comparison.…”

Section: ) Results Of Semi-supervised Modelmentioning

confidence: 99%

“…They found 86% accuracy using semi-supervised SVM which is a little lower than that of our semi-supervised model. However, classifying multiclass driving maneuvers increases the number of classifiers and computational complexity as well [42]. In [42], accuracy has been computed for motion, velocity, turning were 87.3%, 78.4% and 76.9%, respectively.…”

Section: ) Results Of Semi-supervised Modelmentioning

confidence: 99%

“…However, classifying multiclass driving maneuvers increases the number of classifiers and computational complexity as well [42]. In [42], accuracy has been computed for motion, velocity, turning were 87.3%, 78.4% and 76.9%, respectively. In motion category, sub classes were move and stop which were very easy to find out.…”

Section: ) Results Of Semi-supervised Modelmentioning

confidence: 99%

“…Our semi-supervised model provides lower performance than supervised model. A high number of training data causes data redundancy [42]. Besides, autoencoder provides a lossy output by nature; therefore the encoded vector is not the same as the feature vector of Supervised model.…”

Section: E Discussionmentioning

confidence: 99%

“…Mammeri et. al [42] proposed a deep semisupervised approach utilizing manually few labeled driving data considering three CAN bus parameters i.e., velocity, acceleration, and steering wheel angle. They proposed a coarse thresholding strategy to label data from video, re-fined the created labels to reduce error and trained a simple Convolutional Neural Network (CNN) to classify ten driving maneuvers subclasses which were stop, move, acceleration, deceleration, constant speed, left/right turning, left/right curving and constant direction under three groups (motion, velocity, turning).…”

Section: Deep Learning Approachesmentioning

confidence: 99%

See 4 more Smart Citations

Driving Maneuver Classification Using Domain Specific Knowledge and Transfer Learning

2021

View full text Add to dashboard Cite

With the increasing number of vehicles, the usage of technology has also been increased in the transportation system. Although automobile companies are using advanced technologies to develop high performing transports, traffic safety still remains to be a concerning issue. Drivers' driving behavior is considered as one of the key factors of the traffic safety, which could be monitored from their individual driving maneuvers. In this paper, we present a supervised learning model and a semi-supervised transfer learning model for the classification of driving maneuvers from the sensor fusion time series data. The semi-supervised model consists of an unsupervised long-short term memory (LSTM) autoencoder and a supervised LSTM classifier. The supervised model consists of a supervised LSTM model. Because of using LSTM, both of the models can analyze time-series data. In the semi-supervised model, the LSTM encoder learns from unlabeled data as a compressed low dimensional feature vector, which then transfers the learning to the supervised LSTM classifier to classify the driving maneuvers. With the proposed models, we use domain specific knowledge data of the driving environment, such as data changing rules of various driving maneuvers as well as the temporal features over time. We use class functions for seven driving maneuver types and convert those into binary feature vector to use with the LSTM models. We present a comparative analysis of the per class accuracy of the proposed semi-supervised and supervised models with and without using domain-specific knowledge, where the models with the domain specific knowledge outperform. Our proposed semi-supervised and supervised models are compared with the other existing approaches, where our models trained with the domain specific knowledge provide better performance. We also compare the per class accuracy for both the supervised and semi-supervised models, where all the maneuver class accuracy for supervised model is above 98% and semi-supervised model is above 95%. Although the supervised model outperforms the semi-supervised model, semi-supervised model would be more beneficial in applications where the labeled driving maneuvers data is hard to capture or insufficient.INDEX TERMS Driving maneuver classification, domain specific knowledge, LSTM autoencoder, semisupervised learning, transfer learning.

show abstract

Section: ) Results Of Semi-supervised Modelmentioning

confidence: 99%

Section: ) Results Of Semi-supervised Modelmentioning

confidence: 99%

Section: ) Results Of Semi-supervised Modelmentioning

confidence: 99%

Section: E Discussionmentioning

confidence: 99%

Section: Deep Learning Approachesmentioning

confidence: 99%

See 3 more Smart Citations

Driving Maneuver Classification Using Domain Specific Knowledge and Transfer Learning

2021

View full text Add to dashboard Cite

show abstract

Driving maneuver classification from time series data: a rule based machine learning approach

2022

View full text Add to dashboard Cite

Drivers’ improper driving behavior plays a vital role in road accidents. Different approaches have been proposed to classify and evaluate driving performance to ensure road safety. However, most of the techniques are based on neural networks which work like a black box and make the logical reasoning behind the classification decision unclear. In this paper, we propose a rule-based machine learning technique using a sequential covering algorithm to classify the driving maneuvers from time-series data. In the sequential covering algorithm, the impact of each rule is measured as the metrics of coverage and accuracy, where the coverage and accuracy indicate the amount of covered and correctly identified instances in a maneuver class, respectively. The final ruleset for each maneuver class is formed with only the significant rules. In this way, the rules are learned in an unsupervised manner and only the best performance of the rules are included in the ruleset. The set of rules is also optimized by pruning based on the performance of the test data. Application of the proposed system is beneficial compared to the traditional machine learning and deep learning approaches which typically require a larger dataset and higher computational time and complexity.

show abstract

A method based on Vision Transformer and multiple image information for vehicle lane-changing recognition in mixed traffic and connected environment

Ji,

Zhang,

Zhang

2024

Transportation Letters

View full text Add to dashboard Cite

Design of a Semi-Supervised Learning Strategy based on Convolutional Neural Network for Vehicle Maneuver Classification

Cited by 7 publications

References 15 publications

Driving Maneuver Classification Using Domain Specific Knowledge and Transfer Learning

Driving Maneuver Classification Using Domain Specific Knowledge and Transfer Learning

Driving maneuver classification from time series data: a rule based machine learning approach

A method based on Vision Transformer and multiple image information for vehicle lane-changing recognition in mixed traffic and connected environment

Contact Info

Product

Resources

About