Abstract
Human–vehicle classification is an essential component to avoiding accidents in autonomous driving. The classification technique based on the automotive radar sensor has been paid more attention by related researchers, owing to its robustness to low-light conditions and severe weather. In the paper, we propose a hybrid support vector machine–convolutional neural network (SVM-CNN) approach to address the class-imbalance classification of vehicles and pedestrians with limited experimental radar data available. A two-stage scheme with the combination of feature-based SVM technique and deep learning-based CNN is employed. In the first stage, the modified SVM technique based on these distinct physical features is firstly used to recognize vehicles to effectively alleviate the imbalance ratio of vehicles to pedestrians in the data level. Then, the residual unclassified images will be used as inputs to the deep network for the subsequent classification, and we introduce a weighted false error function into deep network architectures to enhance the class-imbalance classification performance at the algorithm level. The proposed SVM-CNN approach takes full advantage of both the locations of underlying class in the entire Range-Doppler image and automatical local feature learning in the CNN with sliding filter bank to improve the classification performance. Experimental results demonstrate the superior performances of the proposed method with the score of and area under the curve (AUC) of the receiver operating characteristic (ROC) of over several state-of-the-art methods with limited experimental radar data available in a 77 GHz automotive radar.
Highlights
In recent years, autonomous vehicles have received widespread attention from the academic research community and the general public
To address the classification in the class-imbalance case, we introduce a weighted false error (WFE) function that can be deployed readily in deep network architectures, and take full advantage of the powerful ability of automatic feature learning in the convolutional neural networks (CNNs) to improve the classification performance in the algorithm level
A novel two-stage hybrid support vector machine–convolutional neural network (SVM-CNN) method will be introduced to acquire enhanced classification performance with mitigating the class-imbalance issue. Unlike these methods, which directly utilize Range-Doppler images or micro-Doppler images as input images into the neural network for the classification [12,13,16,18,19], we proposed a two-stage procedure with joint exploitation of the support vector machine (SVM) and the CNN techniques
Summary
Autonomous vehicles have received widespread attention from the academic research community and the general public. Further works focused on using different domains of information to achieve vehicle and pedestrian classification, such as the phase characteristic of object signature [7], and the defined parameter, root radar cross section (RRCS) [8] Those methods mentioned above required ‘handcrafted’ extraction process on the radar data in order to obtain the most suitable combination of features to maximize classification accuracy [9,10]. A hybrid SVM-CNN approach is proposed to address the class-imbalance classification of vehicles and pedestrians with limited experimental radar data available in an automotive radar sensor. To address the classification in the class-imbalance case, we introduce a weighted false error (WFE) function that can be deployed readily in deep network architectures, and take full advantage of the powerful ability of automatic feature learning in the CNN to improve the classification performance in the algorithm level.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
