Xueqian Wang

Deep Multi-Agent Reinforcement Learning (MARL) excels in co-operative tasks but often struggles with local optima in high - dimensional joint action spaces. In contrast, Evolutionary Algorithms (EAs) offer robust global exploration capabilities. Although hybrid approaches seek to combine the strengths of both paradigms, they typically face a critical bottleneck: the prohibitive sample cost of evaluating large populations via environment rollouts. To address this challenge, we propose Surrogate-assisted Evolutionary Multi-Agent Reinforcement Learning (SEMARL), a unified framework that synergizes gradient-based refinement with surrogate-assisted evolutionary search. SEMARL employs a cooperative co-evolutionary architecture to maintain diverse agent policies and injects gradient-refined parameters into the population to accelerate convergence. Crucially, we leverage the centralized critic from the gradient learner as a computationally efficient surrogate for fitness estimation. To prevent misleading guidance from an inaccurate critic, we introduce an adaptive reliability control mechanism based on temporal difference (TD) error, which dynamically regulates the surrogate's influence. Experiments on the Multi-Agent MuJoCo benchmark demonstrate that SEMARL significantly outperforms other algorithms, achieving superior asymptotic performance with substantially higher sample efficiency.

As drone-based applications proliferate, paramount contactless sensing of airborne drones from the ground becomes indispensable. This work demonstrates concentrating on propeller rotational speed will substantially improve drone sensing performance and proposes an event-camera-based solution, EventPro. EventPro features two components: Count Every Rotation achieves accurate, real-time propeller speed estimation by mitigating ultra-high sensitivity of event cameras to environmental noise. Every Rotation Counts leverages these speeds to infer both internal and external drone dynamics. Extensive evaluations in real-world drone delivery scenarios show that EventPro achieves a sensing latency of 3 ms and a rotational speed estimation error of merely 0.23%. Additionally, EventPro infers drone flight commands with 96.5% precision and improves drone tracking accuracy by over 22% when combined with other sensing modalities. Demo: https://eventpro25.github.io/EventPro/.

Holographic displays deliver realistic visual experiences by reconstructing three-dimensional (3D) light fields. However, widely used RGB-D data offers reliable geometry only in focused regions, lacking light field information in defocused areas, particularly with transparent or reflective surfaces. Existing methods typically utilize coherent diffraction to approximate defocus effects in incoherent perception, introducing ringing artifacts and unnatural depth transitions. We propose a method generating physically consistent complex-valued light fields from multi-layer RGB-D data. Our approach integrates neural rendering and multi-branch phase prediction networks to compute amplitude and phase, while explicitly enforcing wave propagation to preserve coherence and physical consistency. Our simulations achieve an average peak signal-to-noise ratio of 32.35 dB and an average structural similarity index of 0.938. Experiments demonstrate correct defocus relationships and natural defocus blur in scenes with glasses and mirrors. Our approach bridges coherent computation and incoherent perception, addressing physical consistency in 3D holography for next-generation human-computer interfaces.

Advanced sensing is rapidly reshaping intelligent robotic systems by enabling robots to perceive, reason, and interact with the physical world under real-world uncertainty

Real-world decision-making systems often suffer from non-negligible observation and action delays arising from sensing and actuation, which break the standard Markov assumption and degrade reinforcement learning (RL) performance. We present Delay-resilient Encoder-Enhanced RL (DEER), a practical framework that systematically integrates well-established techniques (sequence-to-sequence representation learning, information-state construction, and offline-to-online RL) to enable robust performance in delayed environments. DEER first pretrains an encoder on offline delay-free trajectories and then uses it online to map delayed observations with recent action histories into fixed-length context vectors for standard RL agents. We further provide a theoretical decomposition that links reconstruction error, distributional mismatch, latent-policy generalization, and environment mismatch to overall performance, offering interpretable guidance for design choices. Empirically, integrating DEER with Soft Actor-Critic yields competitive and stable performance across a range of constant and random delaysettings on six continuous-control benchmarks. The contribution is therefore twofold: (i) a practical, deployable engineering solution for delayed RL that leverages existing building blocks effectively; and (ii) a quantitative analysis that explains when and why this integration works.

Human upper-limb movement is produced through the antagonistic action of tendons and is controlled in a joint space-oriented manner. Inspired by this functionality, a fully decoupled tendon-driven humanoid arm (FDTDH-Arm) is proposed, in which joint space decoupling is achieved at the mechanical level via humanoid antagonistic actuation and joint regulation rather than complex modeling-based compensation. To characterize the motion behavior introduced by rolling constraints, joint-level and whole-arm kinematic models are established. A prototype of the proposed arm is developed and experimentally validated. The results demonstrate effective mechanical joint space decoupling, passive joint stiffness of the same order of magnitude as that reported for the human upper limb, a mean positioning error of 0.40 mm, and rapid whole-arm motion with a maximum end effector velocity of 3.62 m/s. The proposed design provides a mechanical implementation and biomimetic solution for humanoid manipulation in human-interactive environments.

Compared with general Federated Learning (FL), Decentralized FL (DFL) has diminished central communication burdens and lower risks of disruption. To be compatible with real-world scenarios, some existing works produce several local personalized models rather than a universal model for all edge devices. However, they still suffer from inferior performance due to the full model aggregation in heterogeneous datasets. Therefore, we propose a DFL framework DFedMDC through partial model personalization, which can adapt to resource-heterogeneous environments (e.g., the Internet of Things (IoT)). Specifically, it personalizes the “right” components of local models and trains the shared model u_i and personal model v_i alternately in each client i. To further accelerate the convergence process, we propose DFedSMDC with properly directed noise perturbation into the gradient update. Theoretically, we provide convergence analysis of both algorithms in the general non-convex setting. It can shed light on how vital factors affect the convergence rate, such as these alternate updates of partial models u_i, v_i, data heterogeneity δ² as well as various communication topologies (characterized by the spectral gap 1 − λ). Empirically, we confirm the state-of-the-art (SOTA) superiority of the proposed methods on several real-world datasets with various data distributions, relative to both SOTA personalized FL (PFL) and DFL baselines.