Autonomous robots, particularly in the field of delivery, play a crucial role in enhancing the safety and efficiency of intelligent transportation systems. This study introduces a delivery robot designed to recognise traffic signs, distinguishing them from other types of signs, to navigate autonomously and safely in urban environments. To achieve this, Convolutional Neural Networks (CNN) were employed for traffic sign classification. A diverse image dataset was used for training and evaluation. The model achieved an accuracy of $93.35 \%$, a precision of $\mathbf{9 4 . 0 1 \%}$, a recall of $\mathbf{9 2 . 6 0 \%}$, and an F1-score of $\mathbf{9 3 . 2 9 \%}$, demonstrating its strong ability to differentiate traffic signs from other objects. The confusion matrix analysis highlights the need to refine the handling of false positives and false negatives to enhance system reliability. These findings validate the effectiveness of CNNs in traffic sign recognition and emphasize the significance of expanding and diversifying training data to improve the model’s adaptability across different environments.
Implementation of a Traffic Sign Recognition System Based on CNN for Autonomous Robots
01.07.2025
1416482 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Implementation of Traffic Sign Recognition System on Raspberry Pi
Springer Verlag | 2022
|