Actuators that use soft materials such as polymers (soft actuators) have various advantages such as small size, light weight, quietness, and water resistance, and can also make use of various power sources such as heat, electricity, and light. Soft actuators can also make curved and delicate motions as muscles, so an active role is expected in places closer to our lives, and application is especially anticipated to wearable machines for work and power assist in the rehabilitation and nursing care fields, and remote-controlled machines for medical surgery support, etc. However, the material development process thus far has relied on repeated trials and errors based on the "intuition and experience" of engineers, and the enormous costs and time needed has become an issue. Various simulation technologies based on computational science have been developed as a solution, and with recent improvements in computer performance and development of various software, these technologies are developing to a level that can be expected as practical means of material design and search.

The researchers have developed a method that accelerates development of materials that exhibit large deformation without resistance, and are needed for soft actuators and other applications, as part of "Ultra High-Throughput Design and Prototyping Technology for Ultra Advanced Materials Development Project" of the New Energy and Industrial Technology Development Organization (NEDO).

The developed method was applied to liquid crystal elastomers, which are the subject of active R&D competition as soft actuator materials. The many parameters that represent molecular structures and the corresponding simulation results of material deformation were compiled into a database and analysis was performed using machine learning. This enabled identification of the molecular structure parameters that determine large deformation characteristics, and the number of required parameters is reduced to approximately 1/10. Thus it is possible to propose likely candidates for molecular structures that exhibit "Soft-Elasticity," or large deformation without resistance (output loss), which is a major feature of soft actuator materials, in a very short time. This can be expected to greatly shorten the development period for innovative soft actuator materials. The developed method is also expected to be applied to development of various materials with large deformation characteristics, such as elastomers and gels.