Soft robots powered by sustainable energy abundantly available on Earth, such as heat, humidity, sunlight, osmotic potential, pH variation, triboelectricity, and wind, represent a promising shift toward eco-friendly and autonomous robotic systems. Efficiency depends on selecting and engineering responsive materials that directly transform environmental stimuli into mechanical actuation and motion, or harvest and store environmental energy to power actuators. Thermo-responsive materials undergo shape changes with temperature variations, while hygroscopic materials leverage moisture adsorption to induce actuation. Photothermal materials convert sunlight into heat and can combine thermal or hygroscopic actuators for controlled deformation. Osmotic processes drive movement through fluidic interactions, whereas pH-sensitive hydrogels respond to chemical gradients, facilitating controlled motion. Triboelectric materials generate electricity via contact-induced charge transfer, enabling self-powered sensing and actuation, while wind-dispersed structures exploit aerodynamic forces for unique movements. This review explores the critical roles of chemical, physical, mechanical, and environmental properties of materials in designing soft robots for sustainable and autonomous operation. Importantly, the review distinguishes between the broad concept of environmental energy and operation that is energetically sustainable. It systematically evaluates reported actuators and soft robotic systems based on whether their required energy sources and operating conditions are naturally occurring and regenerable, or instead depend on restricted environmental ranges, auxiliary inputs, or laboratory-controlled conditions. By examining material behavior, integration into multifunctional composites, and mechanism design for exploiting sustainable energy, this review identifies both established and emerging pathways toward environmentally realistic, autonomous, and long-lived soft robotic systems, with potential applications in environmental monitoring, reforestation, and other robotic domains.
Mariani et al. (2026) studied this question.