Telerehabilitation is a cost-effective alternative to in-clinic rehabilitation. Although convenient, it lacks immersive and free-viewpoint patient visualization. Current research explores two solutions to this issue. Mesh-based methods use 3D models and motion capture for AR visualization. However, they are labor-intensive and less photorealistic than 2D images. Microsoft's Holoportation generates photorealistic 3D models with eight RGBD cameras in real time. However, it requires complex setups, high GPU power, and high-speed communication infrastructure, making deployment challenging. This article presents a Real-Time Free-Viewpoint Holographic Patient Rendering (RT-FVHP) system for telerehabilitation. Unlike traditional methods that require manually crafted assets such as 3D meshes, texture maps, and skeletal rigging, our data-driven approach eliminates the need for explicit asset definitions. Inspired by the HumanNeRF framework, we retarget dynamic human poses to a canonical pose and leverage 3D Gaussian Splatting to train a neural network in canonical space for patient representation. The trained model generates 2D RGB$\sigma$σ outputs via Gaussian Splatting rasterization, guided by camera parameters and human pose inputs. Compatible with HoloLens 2 and web-based platforms, RT-FVHP operates effectively under real-world conditions, including handling occlusions caused by treadmills. Occlusion handling is accomplished using our Shape-Enforced Gaussian Density Control (SGDC), which initializes and densifies 3D Gaussians in occluded regions using estimated SMPL human body priors. This approach minimizes manual intervention while ensuring complete body reconstruction. With efficient Gaussian rasterization, the model delivers real-time performance of up to 400 FPS at 1080p resolution on a dedicated RTX6000 GPU.

In this article, we propose a distributed medium access control (MAC) protocol for the unmanned aerial vehicle (UAV) network with directional antennas. It uses a distributed learning algorithm to control each node's communication parameters (such as sending rate) according to the observations of only 1-hop neighbors. Because of the asynchronous and decentralized nature of multi-agent deep distributed reinforcement learning (MADDRL), we adopt Multi Agent Deep Deterministic Policy Gradient (MADDPG) [3] to generate discrete actions for the MAC layer protocol. Under the individual and distributed observations of the environment, each node/agent can update their own states and rewards locally. Moreover, two major communication quality evaluation metrics - throughput (THR) and delay (DEL) are used to show the effectiveness and scalability of the distributed algorithm. From the results, the protocol using MADDPG shows more advantages than the general MAC schemes, including higher end-to-end THR, lower queuing DEL, and lower communication message exchange.