Deep iterative fuzzy pooling in unmanned robotics and autonomous systems for Cyber-Physical systems

Unmanned robotics and autonomous systems (URAS) are integral components of contemporary Cyber-Physical Systems (CPS), allowing vast applications across many domains. However, due to uncertainties and ambiguous data in realworld environments, ensuring robust and efficient decision-making in URAS is difficult. By capturing and reasoning with linguistic data, fuzzy logic has emerged as a potent tool for addressing such uncertainties. Deep Iterative Fuzzy Pooling (DIFP) is a novel method proposed in this paper for improving decision-making in URAS within CPS. The DIFP integrates the capabilities of deep learning and fuzzy logic to effectively pool and aggregate information from multiple sources, thereby facilitating more precise and trustworthy decision-making. This research presents the architecture and operational principles of DIFP and demonstrates its efficacy in various URAS scenarios through extensive simulations and experiments. The proposed method demonstrated a high-performance level, with an accuracy of 98.86%, precision of 95.30%, recall of 97.32%, F score of 96.26%, and a notably low false positive rate of 4.17%. The results show that DIFP substantially improves decisionmaking performance relative to conventional methods, making it a promising technique for enhancing the autonomy and dependability of URAS in CPS.