Kangle Yuan | arXiv Science

arXiv2026

Band structure control in the altermagnetic candidate MnTe by temperature and strain

The temperature and strain dependences of the optical conductivity spectrum of hexagonal manganese telluride (MnTe) were measured, revealing absorption in the terahertz (THz) region from spin-split bands to acceptor levels. The temperature dependence of the THz absorption peak is consistent with that of a ferromagnetic phase transition, even though MnTe exhibits no net magnetism. The temperature dependence was attributed to a change in the altermagnetic electronic structure. A Fano-like antisymmetric line shape in the optical phonon absorption was observed, which originates from the interaction between optical phonons and the spin-split bands. Additionally, under negative uniaxial pressure, the THz peak shifts away from the Fermi level (EF), suggesting that spin-splitting bands at energies away from EF, consistent with the theoretical prediction that the spin-splitting angle decreases. The observed behavior of the THz peak clearly shows that MnTe has the altermagnetic electronic structure.

cond-mat.str-el↗

arXiv2025

Learning Agile Tensile Perching for Aerial Robots from Demonstrations

Perching on structures such as trees, beams, and ledges is essential for extending the endurance of aerial robots by enabling energy conservation in standby or observation modes. A tethered tensile perching mechanism offers a simple, adaptable solution that can be retrofitted to existing robots and accommodates a variety of structure sizes and shapes. However, tethered tensile perching introduces significant modelling challenges which require precise management of aerial robot dynamics, including the cases of tether slack & tension, and momentum transfer. Achieving smooth wrapping and secure anchoring by targeting a specific tether segment adds further complexity. In this work, we present a novel trajectory framework for tethered tensile perching, utilizing reinforcement learning (RL) through the Soft Actor-Critic from Demonstrations (SACfD) algorithm. By incorporating both optimal and suboptimal demonstrations, our approach enhances training efficiency and responsiveness, achieving precise control over position and velocity. This framework enables the aerial robot to accurately target specific tether segments, facilitating reliable wrapping and secure anchoring. We validate our framework through extensive simulation and real-world experiments, and demonstrate effectiveness in achieving agile and reliable trajectory generation for tensile perching.

cs.RO↗