Fully stretchable hydrogel-based moisture-electric generators (FSHMEGs) are promising power sources for wearable and implantable electronics. Current FSHMEGs are constrained by low electrical output and mechanical fragility, mainly due to weak interfacial adhesiveness within multilayered archi- tectures. Here, we introduce an intrinsically adhesive hydrogel that forms robust hydrogel-electrode interfaces, enabling efficient transfer of both electrical charges and mechanical loads during deformation. As a result, the device delivers an open-circuit voltage of 0.94 V and a current density of 141 µA cm⁻² at 85% relative humidity, and maintains stable output for more than 220 h. The reinforced interface also imparts exceptional mechanical durability, exhibiting only negligible performance degradation after 8000 folding cycles and 1000 stretching cycles at 80% strain. Benefiting from rapid humidity responsiveness and continuous power delivery, the device enables non-invasive

respiration monitoring and can directly power wearable electronics (e.g., electrocardiogram (ECG) sensors). This interfacial-engineering holds promise for advancing the development of next-generation fully stretchable and flexible energy systems.

By electrochemically introducing phosphonate ester groups into conductive polymer films, researchers at Science Tokyo have addressed a fundamental trade-off between electronic charge transport and ion transport, overcoming a key performance limitation in organic electrochemical transistors (OECTs). The method enables precise tuning of polymer properties and can be applied to semicrystalline materials without redesigning monomers, supporting the development of improved biosensors and flexible electronic devices.

SUTD researchers have developed a reinforcement-learning-based safety system that teaches a stair-traversing service robot to brace itself mid-fall, addressing one of the biggest barriers to deploying autonomous robots on staircases.

A new study in Artificial Intelligence and Autonomous Systems shows that neural networks pre-trained to predict continuous gait cycle progression can be reused to improve gait phase classification from foot-mounted inertial sensor data. The approach achieved an F1-score of 0.9788, CPU inference latency below 0.07 ms with a compact DNN, and 92.3% accuracy on an independent young-adult dataset.