Computational intelligence for soft strain sensor sustainability

The ever-growing demand of soft strain sensors in ubiquitous electronics poses urgent need of sustainability innovations to reduce resource waste. Besides the efforts toward sustainable device manufacture, it is as important to build performance sustainability of soft sensor with reliable sensing signals over its lifecycle, yet facing pervasive predicaments of nonlinearity, hysteresis, cycling attenuation, and batch inconsistency. Herein, instead of empirical experiments via material or structural engineering, we propose an efficient computational sustainability framework to comprehensively address these signal predicaments by utilizing hierarchical domain-constrained machine learning (ML) models. Using an eco-friendly carbon waste-based strain sensor as a case study, domain intuited ML models are developed and show both high computation efficiency and learning accuracy in terms of real-time sensing signal calibration and signal compensation of the developed sensor, which automatically enhances its unsatisfied sensing performance to equivalently linear, non-hysteresis, long-term stable, and batch consistent one without trial-and-error experiments. As final demonstrations, the ML-driven computational models are capable of expanding sensor’s reliable working lifetime for > 3,000 times than its own counterpart for multiple robotic tasks, facilitating their long-term usages during practical applications to attain the sustainable development goal.