This work presents a driver assistance system to detect brake and parking signals. Vehicle present ahead of the host vehicle performs sudden action such as applying brakes. The host vehicle’s driver must respond in real time to avoid accidents or collisions. Detection of action performed by a leading vehicle using its taillights is implemented with computer vision and machine learning techniques. Features are extracted using scale invariant feature transform (SIFT) and accelerated-KAZE (AKAZE). The dimensions are reduced using K-means clustering and then by principal component analysis (PCA). Five classification models have been trained to evaluate the performance. Random forest classifier provided the highest accuracy of 82% among all classifiers. Voting classifier provides final prediction based on five classification model’s output. It provided 81% accuracy.
Driver Assistance System for Recognition of Brake and Parking Signal
Lect. Notes in Networks, Syst.
Proceedings of Third International Conference on Advances in Computer Engineering and Communication Systems ; Chapter : 21 ; 237-245
2023-03-18
9 pages
Article/Chapter (Book)
Electronic Resource
English
Method for driver assistance system to release electro-hydraulic parking brake
European Patent Office | 2024
|METHOD FOR A DRIVER ASSISTANCE SYSTEM FOR RELEASING AN ELECTRO-HYDRAULIC PARKING BRAKE
European Patent Office | 2024
|