Intelligent Transportation Monitoring Through Teager–Kaiser Energy Feature Enhancement and Metaheuristic Optimization

In next-generation smart city infrastructure, monitoring emotional stability in real time is a crucial component. This paper introduces a powerful Edge-AI monitoring system suitable for all children, with a specific focus on those displaying special behavioral patterns or sensory sensitivities. The proposed model combines improved Frilled Lizard Optimization (FLO), the Histogram of Orientated Gradients (HOG), and the Teager–Kaiser Energy Operator (TKEO) to enhance the analysis of spontaneous behaviors. Then, the model is called FLHTKEO. The model demonstrates high accuracy in interpretating children’s facial expressions in real-time recorded datasets. To address the non-stationary noise and varying brightness typical of school bus interior, the TKEO is employed as a preprocessing tuning stage to amplify the immediate energy of features extracted via HOG. Additionally, the proposed model is implemented on a System-on-Chip (SoC) using a Raspberry Pi 4 to achieve real-time monitoring and generate automatic alarms when abnormal or uncertain behaviors are detected among children in the school vehicle. Real-time datasets were collected, simulating diverse school bus scenarios under varying lighting and behavioral conditions. The proposed model achieves a superior AUC-ROC ≅ 0.9613, proving a high degree of the precision recognition of children’s facial expressions. The system will help with some healthy measures during ongoing and future emergencies. Simulations and experimental results indicate that the proposed method is effective for the model, and automatic compliance monitoring systems are set to advance in the future. Finally, it is compared with the last piece of previous work.