AI-driven framework for text neck syndrome detection using non-contact software-defined RF sensing and sequential deep learning

Text neck syndrome is a rapidly growing health concern in today’s society, largely caused by the excessive use of mobile devices. Text neck syndrome has a significant impact on the musculoskeletal health of the broader population, particularly among frequent users of mobile devices. These types of health issues require treatment at an early stage, as they tend to worsen over time and become more difficult to manage. To address this issue, this study presents an innovative non-contact posture monitoring system using software-defined radio (SDR) technology to detect and analyse postural patterns associated with text neck syndrome for early interventions. The non-contact software-defined radio sensing system is developed using Universal Software Radio Peripheral (USRP) devices equipped with antennas. The experiments are conducted in a controlled lab environment to collect a dataset of distinct neck tilt angles ( $$0^o$$ 0 o , $$15^o$$ 15 o , $$30^o$$ 30 o , $$45^o$$ 45 o , and $$60^o$$ 60 o ). The collected dataset is processed using advanced signal processing techniques to clean and smooth the postural patterns. The machine learning (ML) and deep learning (DL) algorithms are used to categorise postural patterns and identify deviations indicative of text neck syndrome. The performance of these models was subsequently evaluated. The results demonstrate the ML and DL model’s ability to detect healthy and unhealthy postures with a maximum accuracy of 99.97% by using the random forest ML model and 99.89% by using the Bidirectional Long-Term Memory (Bi-LSTM) DL model. This system represents an accessible, contactless, and portable solution with the potential to revolutionise the early detection of text neck syndrome.