In this paper, we propose a data-aided channel estimator, which can improve the performance of the linear minimum-mean-squared-error (LMMSE) by clustering and reinforcement-learning for multiple-input multiple-output (MI-MO) systems. For clustering-based data detection, we develop a system constrained Gaussian mixture model (SCGMM), in which the a posteriori probabilities (APPs) can be calculated by the expectation-maximization (EM) algorithm. The initial centroids of the SCGMM are sensitive to the channel estimation. To obtain robust channel estimation, we design initial pilots that can reduce the estimated error of the LMMSE. To further improve the quality of channel estimation, we propose a data-aided channel estimation algorithm, which exploits the techniques of coding and reinforcement-learning to obtain soft symbol decisions. Numerical results show that the proposed method can approach the bit-error-rate (BER) performance with perfect channel state information (CSI) in the high signal-to-noise (SNR) region.

The fish-like robot is one of the typical underwater robots, which has the advantage of high maneuverability with low noise due to its bioinspired structure and biomimetic locomotion. However, it is challenging to efficiently design motion controllers for such robots to achieve satisfactory performance on specific control tasks in the real underwater environment, since the complex fluid-structure interaction exists during their swimming and exact dynamic models are absent. In this article, we propose a learning framework, incorporating a simulation system and a training methodology, to autonomously and fast train in simulation to create control policies that are capable of directly applying to a type of physical fish-like robots to perform motion control tasks. First, we construct a simulation system combining a data-driven environment and a computational fluid dynamics (CFD)-based environment, thus well balancing the simulation accuracy and the calculation speed. Second, we design a training methodology to train deep reinforcement learning (DRL)-based policies for the robot in our constructed simulation system to perform a specific control task. Then, we use two typical motion control tasks to verify our proposed framework. One is the path-following control task, which is a one-objective problem with dense rewards, while the other is the pose control task which is a two-objective problem with sparse rewards. For each task, the DRL-based control policy trained by our learning framework is directly deployed on the physical fish-like robot to perform the task in the real world. Experimental results show that the policies trained in simulation still work well in the real world, and perform even better in terms of control accuracy and stability compared with the traditional control methods, thus demonstrating the effectiveness of our learning framework.

Mobile gaming has experienced explosive growth in the past few years. Meanwhile, the gaming scenes become more complicated, but the mobile devices have limited computing resources and inadequate energy, which can not satisfy the increasing requirements of mobile games. In recent years, edge computing flourishes as a promising solver to offload computation-intensive tasks to edge devices. However, how to schedule the various rendering tasks to heterogeneous edge servers is challenging because of the variety and unknown workload of the rendering tasks. To overcome these challenges, we design a rendering scheduling framework in edge computing and formulate the rendering scheduling problem as a congestion game, whose objective is to minimize the maximal completion time. We also design a rendering scheduling algorithm based on our congestion game model and derive the Nash Equilibrium and convergence. In addition, we explore a case study called redundant object elimination in rendering scenarios and adopt machine learning tools for workload estimation. Finally, we conduct extensive simulations based on our algorithm, which show significant performance compared with alternative strategies.