Advanced monitoring technologies that can instantly assess the surrounding environment and driver behavior are required to solve this problem. In this study, we present a hybrid deep learning framework for improving real-time driver safety by leveraging convolutional neural networks (CNNs) and convolutional long short-term memory (ConvLSTM) networks. This framework is designed to process information obtained from an in-vehicle dashcam, enabling the system to keep an eye on both the driver and the surrounding road conditions. The hybrid CNN module extracts spatial information from the dashcam vision-based footage, enabling it to recognize objects such as lane markers, cars, and pedestrians. These characteristics are essential for comprehending possible environmental risks. The ConvLSTM network, in the meantime, analyzes frame sequences, collecting temporal dependencies, including changes in vehicle speed, lane position, and closeness to obstacles over time. The system can now continually evaluate the driving behavior and the external road circumstances thanks to this combined spatial–temporal model, which improves its ability to identify risky scenarios, including lane deviations, unsafe following distances, and distracted driving. The proposed hybrid model offers a thorough safety monitoring system that incorporates both behavioral and environmental elements, and it can produce real-time alerts to warn drivers of potential hazards. The dashcam module is synchronized with vehicles, and the network analysis is incorporated with NVIDIA GPU (RTX 3080) real-time hybrid deep learning hardware driver safety monitoring system presented.
Devaraj et al. (2026) studied this question.