Deep Learning and IoT Based Robotics to Monitor the Traffic

In densely populated areas, managing traffic laws has become more difficult due to the rising number of cars on the road. One of the most critical components of effective traffic management and improved mobility is real-time traffic monitoring systems. Consequently, precise and dependable real-time traffic information has always been essential for cars and drivers. Many other approaches have been suggested as potential answers to the issues plaguing traffic conditions recently. A different approach is VCC or vehicle cloud computing. In addition, an advanced model called an Iot-aided robotic (IoRT), which uses cameras and Internet of Things (IoT) detector nodes to gather actual traffic information, has been created. Two deep learning methods, one based on improved LeNet-5 for real-time signs for traffic identification and the other on the Inception-V3 model for detecting and identifying traffic lights, are the critical contributions provided by this research effort. Additionally, to decrease road accidents, the ideal distance between barriers and ultrasonic sensors was determined by analyzing the timing and speed of ultrasonic waves. Data acquired by cameras and sensors is analyzed using a variety of image processing algorithms before being uploaded to the cloud and made accessible to commuters and drivers via a mobile app. According to the test findings, the suggested models are far more accurate. The expanded GTSRB (EGTSRB) dataset achieved 98.89% accuracy while the German Traffic Sign Recognition Benchmark (GTSRB) dataset achieved 98.23% accuracy using the modified LeNet-5. The second model achieved a precision of 99.7 percent using data from the Laboratory for the Intelligent and Safe Automobiles (LISA). The results of this study beat previous research on comparable traffic monitoring systems by 2.13% for traffic light recognition and detection and by 4.89% for traffic sign identification.