چكيده به لاتين
Today, chronic diseases are the leading cause of mortality worldwide. Preventing the exacerbation of chronic illnesses significantly reduces the mortality rate. In most cases, sudden worsening of a disease is identified through changes in certain initial medical parameters of patients, such as heart rate, respiration rate, body temperature, blood oxygen saturation, and more. Detecting these early warning signs through continuous monitoring provides an opportunity to reduce the risk. Recent advancements in cyber-physical systems in the healthcare domain have enabled the use of wearable devices for remote patient monitoring. However, continuous and sustained service in remote monitoring is highly influenced by the energy consumption of wearable sensors and battery size. Efficient sensor deployment, utilizing sensors in low-power modes with reduced performance, noise reduction, and periodic signal recording, are some proposed energy management methods to prolong device operation. Nonetheless, each of these methods has its drawbacks. Decreasing the number of sensors and using them in low-power modes leads to increased ambiguity and reduced signal quality. Noise reduction methods come with substantial overhead and complexity. Moreover, periodic signal recording, instead of continuous data collection, increases the risk of missing certain health status changes or early signs of deterioration. To address these challenges, a method called Collaborative Sensing (CS) has been introduced. The primary aim of this research is to enhance the precision of sensor selection and reduce energy consumption. This method, by considering data overlap and sensor interdependency, adjusting sensor sampling rates, and dynamically turning sensors on and off, can reduce energy consumption and enhance the reliability of cyber-physical systems. The results obtained from the implementation indicate that the CS method has reduced power consumption by 42.6 percent and 60 percent in comparison to the baseline method and the most recent and closest existing method, respectively. Furthermore, the CS method has improved detection accuracy by 35.4 percent and 5 percent compared to the baseline method and the most recent and closest existing method, respectively.
