No abstract

Soft robots have been extensively researched due to their flexible, deformable, and adaptive characteristics. However, compared to rigid robots, soft robots have issues in modeling, calibration, and control in that the innate characteristics of the soft materials can cause complex behaviors due to non-linearity and hysteresis. To overcome these limitations, recent studies have applied various approaches based on machine learning. This paper presents existing machine learning techniques in the soft robotic fields and categorizes the implementation of machine learning approaches in different soft robotic applications, which include soft sensors, soft actuators, and applications such as soft wearable robots. An analysis of the trends of different machine learning approaches with respect to different types of soft robot applications is presented; in addition to the current limitations in the research field, followed by a summary of the existing machine learning methods for soft robots.

We demonstrate a record-high extinction-ratio of 50.4 dB in a 2 × 2 silicon Mach-Zehnder switch equipped with a variable splitter as the front 3-dB splitter. The variable splitter is adjusted to compensate for the splitting-ratio mismatch between the front and rear 3-dB splitters. The high extinction ratio does not rely on waveguide crossings and meets a strong demand in applications to multiport circuit switches. Large fabrication tolerance will make the high extinction ratio compatible with a volume production with standard complementary metal-oxide semiconductor fabrication facilities.

Conserving and enhancing freshwater biodiversity are global issues to ensure ecosystem integrity and sustainability. To meet this, it is critical to understand how the biological assemblages are determined by environmental gradients in different spatial scales. Nevertheless, information on their large-scale environmental relationships remains scarce in Korea. We aimed to understand nationwide spatial distribution patterns of benthic macroinvertebrates and important environmental factors affecting their distribution in 388 streams and rivers across Korea. A total of 340 taxa, belonging to 113 families in 23 orders of five phyla, were identified. Assemblage composition in most Korean streams included a few predominant colonizers and a majority of rare taxa. Cluster analysis based on benthic macroinvertebrates classified a total of 720 sampling sites into five clusters according to the pollution levels from fast-flowing less polluted streams with low electrical conductivity to moderately or severely polluted streams with high electrical conductivity and slow water velocity. Canonical correspondence analysis revealed that altitude, water velocity and streambed composition were the most important determinants, rather than watershed and water chemistry variables, for explaining the variation in macroinvertebrate assemblage patterns. The results provide basic information for establishing the conservation and restoration strategies of macroinvertebrate biodiversity against anthropogenic disturbances and developing more confident bio-assessment tools for diagnosing stream ecosystem integrity.

Few-layer tungsten diselenide (WSe2) is investigated using circularly polarized Raman spectroscopy with up to eight excitation energies. The main E2g 1 and A1g modes near 250 cm -1 appear as a single peak in the Raman spectrum taken without consideration of polarization but are resolved by using circularly polarized Raman scattering. The resonance behaviors of the E2g 1 and A1g modes are examined.Firstly, both the E2g 1 and A1g modes are enhanced near resonances with the exciton states. The A1g mode exhibits Davydov splitting for trilayers or thicker near some of the exciton resonances. The low-frequency Raman spectra show shear and breathing modes involving rigid vibrations of the layers and also exhibit strong dependence on the excitation energy. An unidentified peak at ~19 cm -1 that does not depend on the number of layers appears near resonance with the B exciton state at 1.96 eV (632.8 nm). The strengths of the intra-and inter-layer interactions are estimated by comparing the mode frequencies and Davydov splitting with the linear chain model, and the contribution of the next-nearest-neighbor interaction to the inter-layer interaction turns out to be about 34% of the nearest-neighbor interaction. Fano resonance is observed for 1.58-eV excitation, and its origin is found to be the interplay between two-phonon scattering and indirect band transition.