This study proposes a spatial multiplexing wireless communication method based on coding metasurfaces, achieving efficient electromagnetic wave manipulation through optimized coding matrices to enable independent information transmission in different directions. By employing the weighting theorem, we optimize the reflection characteristics of the metasurface, suppressing sidelobe gain and enhancing the gain difference between the main and sidelobes. This improves beam directionality and reduces energy leakage in non-target directions. In the experiment, we constructed a dual-channel direct digital transmission system, where an FPGA dynamically modulates the coding metasurface to generate specific reflected beams based on input data, enabling independent transmission of a 6.1 GHz carrier signal in two asymmetric directions. Experimental results show that, with Chebyshev-weighted optimized coding sequences, the gain difference between the main lobe and the strongest sidelobe increased by 4.99 dB, 1.5 dB, and 7.33 dB at three positive angles (+18°, +28°, and +45°), and by 4.94 dB, 4.03 dB, and 4.22 dB at three negative angles (−15°, −35°, and −40°), respectively. Furthermore, a wireless communication system was constructed based on the metasurface, enabling simultaneous transmission of digital information at two asymmetric angles, −15° and +18°. The proposed spatial multiplexing strategy can be further extended to multi-channel wireless transmission. This study verifies the feasibility of using metasurfaces to manipulate communication signals, providing a foundational reference for future expansion and applications in complex channel environments.

Manipulating human poses based on natural language is an emerging research field that has traditionally focused on coarse commands such as “walking” or “dancing.” However, fine-grained pose manipulation, like instructing “put both hands in front of the stomach,” remains underexplored. In this paper, we introduce PoseLLaVA, a pioneering model that integrates SMPL-based pose representations into the multimodal LLaVA framework. Through a novel pose encoder decoder mechanism, PoseLLaVA achieves precise alignment between pose, textual, and visual modalities, enabling detailed control over pose manipulation tasks. PoseLLaVA excels in three key tasks: pose estimation, generation, and adjustment, all driven by detailed language instructions. We further introduce a fine-grained pose adjustment dataset PosePart, where each sample contains an initial pose and a target pose, along with specific instructions for adjustments, mimicking the guidance a human instructor might provide. Extensive evaluations across these tasks demonstrate significant improvements over existing methods, including metrics such as MPJPE and PA-MPJPE, which measure SMPL reconstruction errors, and Recall rates, which assess feature alignment across modalities. Specifically, PoseLLaVA reduces MPJPE errors by more than 20% compared to state-of-the-art methods in pose adjustment and generation tasks. Additionally, we demonstrate the feasibility of combining PoseLLaVA with generative models, such as diffusion, for pose image editing, highlighting its potential applications in language-controlled pose manipulation.