A review: Machine learning for strain sensor-integrated soft robots

Compliant and soft sensors that detect machinal deformations become prevalent in emerging robots for closed-loop feedback control. In contrast to conventional sensing applications, the stretchy body of robot enables programmable actuating behaviors automated manipulations across a wide strain range, which poses high requirements integrated customized sensor characteristics, high-throughput data processing, timely decision-making. As various robotic (strain, pressure, shear, etc.) meet similar challenges, this perspective, we choose as representative example summarize latest advancement sensor-integrated design driven by machine learning techniques, including materials optimization, signal analyses, in-sensor computing. These implementations greatly accelerate automation, reduce resource consumption, expand working scenarios robots. We also discuss prospects fusing technology creating next-generation intelligent