Managing Optical Sensor Data Security and Mitigating Data Variation in Real-Time Applications Using Fractional Analytical Method

Abstract

Optical sensors are employed in different real-time applications for an object, temperature, etc., sensing, improving the analysis accuracy. The optical sensors utilized in medical applications collect more sensitive data. Therefore, sensor data must be maintained in terms of providing security and privacy. The optical sensor recorded information influenced by the intermediate attacker that affects the data processing accuracy. The problem with the sensor input analysis is the intensity variation and failures in observed sequences. This article addresses the above problem using reinforcement learning, and the method is named Fractional Analytical Method with Feature Variation (FAM-FV). This method relies on data intensity features observed in sensing, transmission, and analysis time instances which are used to manage data security. Reinforcement learning identifies the analysis required based on ceasing intensity in different intervals. The required analysis is provided for restoring the intensity validations for reducing mean errors. This is performed based on the reinforced data attributes and matching features observed in the previous interval. The matching is performed from sensing to analysis interval to improve accuracy. The learning paradigm identifies and recommends the analysis level in different intervals. The proposed method is verified using accuracy, precision, mean error, analysis time, and complexity metrics.