Excessive vehicle speed is a leading cause of traffic accidents, prompting the strategic installation of speed breakers on roads to mitigate risks in vulnerable areas. These road features play a crucial role in curbing vehicle speeds, ultimately bolstering pedestrian safety. Recognizing the urgency in addressing this concern, the real-time detection of speed breakers becomes imperative for providing timely alerts to drivers. The proposed solution relies on machine learning (ML), demonstrating effectiveness in identifying both marked and unmarked speed breakers under challenging conditions such as faded images, dust, tree shadows and variable street lighting, which are especially crucial during nighttime. By integrating the Arduino Nano 33 BLE and Tiny ML, this system ensures speed breaker recognition, autonomous data collection and delivers instantaneous alerts. This holistic implementation markedly improves on-road safety for drivers. The proposed work is compared with several algorithms and compared to performancemeasures, such as accuracy, precision, recall the results are satisfactory. Finally, a maximum accuracy of 92% was achieved.Using this approach, it was possible to detect speed breaker with a precision of 82% and recall of 94.%.
Real-Time Speed Breaker Detection with an Edge Impulse
SN COMPUT. SCI.
SN Computer Science ; 5 , 6
08.08.2024
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Edge impulse , Arduino nano 33 BLE , MobileNetV2 96 × 96 0.35 model Computer Science , Computer Science, general , Computer Systems Organization and Communication Networks , Software Engineering/Programming and Operating Systems , Data Structures and Information Theory , Information Systems and Communication Service , Computer Imaging, Vision, Pattern Recognition and Graphics
Deep Learning-Based Speed Breaker Detection
Springer Verlag | 2024
|Speed Breaker and Vehicle Accident Detection with Alert Sensors
Springer Verlag | 2025
|Springer Verlag | 2022
|Vision based speed breaker detection for autonomous vehicle [10696-107]
British Library Conference Proceedings | 2018
Edge detection on real time using LOG filter
IEEE | 1994
|