Noise suppression-embedded transfer component analysis for intelligent cockpit body sensor networks

The application of body sensor network technology has demonstrated promising potential in the realm of intelligent cockpit body domain cognitive services.Nevertheless, as the Internet of Things (IoT) continues to evolve, users of smart cabins have become increasingly attuned to their personalized service needs and the critical importance of protecting human domain data privacy.Addressing the challenge of diminishing the domain discrepancy between limited prior data and target user data has emerged as a pressing issue in the local model learning of smart cabin systems.In this paper, we propose a Noise Suppression-embedded Transfer Component Analysis (NS-TCA) method, which is based on a knowledge transfer strategy.By combining distribution shift elimination with sample instance screening, we introduce a noise sample suppression strategy in the transfer component analysis process.This ensures the accuracy and effectiveness of the intelligent vehicle domain task analysis model by improving the quality of training data.Experimental results on multiple public wireless body sensor network datasets show that NS-TCA can significantly improve the performance of intelligent vehicle body domain cognitive tasks.This work can provide technical support for the design of intelligent cockpit local information system frameworks.