Detection of road traffic participants using cost-effective arrayed ultrasonic sensors in low-speed traffic situations

Abstract

Effective detection of traffic participants is crucial for driver assistance systems. Traffic safety data reveal that the majority of preventable pedestrian fatalities occurred at night. The lack of light at night may cause dysfunction of sensors like cameras. This paper proposes an alternative approach to detect traffic participants using cost-effective arrayed ultrasonic sensors. Candidate features were extracted from the collected episodes of pedestrians, cyclists, and vehicles. A conditional likelihood maximization method based on mutual information was employed to select an optimized subset of features from the candidates. The belonging probability to each group along with time was determined based on the accumulated object type attributes outputted from a support vector machine classifier at each time step. Results showed an overall detection accuracy of 86%, with correct detection rate of pedestrians, cyclists and vehicles around 85.7%, 76.7% and 93.1%, respectively. The time needed for detection was about 0.8 s which could be further shortened when the distance between objects and sensors was shorter. The effectiveness of arrayed ultrasonic sensors on objects detection would provide all-around-the-clock assistance in low-speed situations for driving safety.

Hyperparameter Optimization for Improving Recognition Efficiency of an Adaptive Learning System

“…The key solutions include the solution of using sensors and the image processing one. For the solution of using sensors, object detection and tracking like vehicles, pedestrian and other obstacles are focused on [2], [3]. Whereas, the image processing method typically utilizes CNN models, including AlexNet [4], GoogleNet [5], Microsoft ResNet [6], R-CNN [7], Fast R-CNN [8], Faster R-CNN [9], and so on.…”

Gaze and Eye Tracking: Techniques and Applications in ADAS

“…Finally, it is notable that a lack of light at night time reduces the functionality of human vision and cameras, which results in increased pedestrian fatalities occurring at night. The authors of [132] proposed an approach which utilized cost-effective arrayed ultrasonic sensors to detect traffic participants in low-speed situations. The results show an overall detection accuracy of 86%, with correct detection rates of cyclists, pedestrians, and vehicles at around 76.7%, 85.7%, and 93.1%, respectively.…”

Detection of Road Objects With Small Appearance in Images for Autonomous Driving in Various Traffic Situations Using a Deep Learning Based Approach

“…Effectively detecting road objects in various environments would significantly improve driving safety and travel efficiency of autonomous vehicles (AVs) [1]. However, due to the diversification of time of day, weather, complex environment and objects occlusion, road object detection has reached a bottleneck for further improvement [2,3].…”

“…Treating object detection as a regression task, the one-stage methods directly predict the location of an object in an image and classifies the object accordingly. The two-stage methods divide the detection task into two steps: (1) extracting regions of interest from an image where the objects would probably be, (2) analyzing and recognizing the candidate regions for final detection. The one-stage methods run faster, but have a slightly lower accuracy compared with the two-stage methods.…”