Biao Hu

Humanoid soccer poses a representative challenge for embodied intelligence, requiring robots to coordinate agile locomotion with unreliable visual perception in dynamic environments. However, existing systems typically rely on modular pipelines that separate perception from control or assume ideal sensing, making it difficult to achieve coherent and reactive behavior under real-world perceptual limitations. In this work, we present a unified reinforcement learning–based controller that enables humanoid robots to learn vision-driven reactive soccer skills by directly coupling visual perception with locomotion control. The robot is trained in simulation to acquire soccer behaviors, and adversarial motion priors guide policy learning toward natural motion patterns. To support robust performance under imperfect sensing, we introduce an encoder-decoder architecture together with a virtual perception system that models key characteristics of onboard vision, exposing the policy to perceptual noise and detection failures during training. This design encourages the policy to internalize perceptual uncertainty and continuously adapt its motion in a closed loop. The resulting controller produces coordinated soccer behaviors using only onboard vision, including ball searching, chasing, and multidirectional kicking. It reduces ball position estimation error by 46% and shortens time-to-kick by up to 64% compared with a rule-based baseline, achieving around 90% kicking success in frontfield positions. Experiments across diverse environments and dynamic scenarios, including real RoboCup competitions, further demonstrate the robust performance of the controller. These results highlight the practical effectiveness of integrating perceptual uncertainty directly into policy learning for achieving reliable vision-driven behaviors in humanoid robots operating under real-world conditions.

Flexible spines are critical to the remarkable agility and speed of animals. Translating this biological advantage to quadruped robots presents a significant control challenge, particularly in coordinating the spine and limbs for maximal velocity. In this work, we utilize reinforcement learning (RL) to develop high-speed locomotion for a bioinspired mouse robot with a lateral flexible spine. The resulting controller achieves motor performance that demonstrably surpasses non-spined and model-based methods. More importantly, our analysis reveals the principles behind this performance: the emergence of two distinct motion templates. For high-speed walking, the robot learns a “whip-like” spinal oscillation to increase leg swing frequency, while for agile turning, it adopts a dynamic “bend-and-straighten” pattern. These findings demonstrate the capability of RL to not only generate high-performance controllers but also to produce emergent strategies that, upon analysis, reveal underlying principles of high-speed, spine-driven locomotion.

The integration of AI tasks into Artificial Intelligence of Things (AIoT) devices has gained significant attention for its ability to reduce transmission latency and enhance data privacy compared to traditional cloud-based solutions. However, heterogeneous AIoT platforms face the challenge of balancing stringent latency constraints with energy efficiency under dynamic AI workloads. To address this issue, we propose BOLT-PM, a Bayesian Optimization framework for Latency-sensitive Task and Power Management, which dynamically allocates computing resources and optimizes power consumption in AIoT devices. We design a multi-branch neural network to accurately predict task execution times under varying resource constraints, enabling precise and adaptive performance estimation. We further enhance the Bayesian optimization process by embedding a variational autoencoder (VAE) to construct a smooth latent representation of the search space, thereby accelerating convergence and improving optimization stability. We implement BOLT-PM as a lightweight, platform-agnostic runtime framework that continuously adapts to workload fluctuations, maintaining latency guarantees while minimizing energy consumption. We evaluate BOLT-PM on several commercial AIoT boards, including RK3588, Jetson TX2, and Raspberry Pi, and compare it against classical heuristic methods and state-of-the-art approaches. Experimental results show that BOLT-PM achieves substantial energy savings while ensuring low-latency AI task execution, demonstrating its effectiveness as a robust and energy-efficient solution for power-aware AIoT applications.

Scheduling applications in vehicular edge computing (VEC) systems poses significant challenges due to strict timing constraints and varying levels of criticality. This paper presents a three‐stage scheduling framework designed to efficiently manage the execution of mixed‐criticality applications. The proposed method introduces scheduling policies that reduce the complexity of scheduling dual‐criticality DAG (Directed Acyclic Graph) applications on servers by transforming them into equivalent uniprocessor scheduling problems. To further enhance performance, a population‐based evolutionary algorithm is employed to optimize virtual machine configurations on each server, while a game‐theoretic approach assigns DAG applications to servers. Experimental results show that the proposed scheme outperforms both state‐of‐the‐art dynamic programming (DP) and particle swarm optimization (PSO) methods. The proposed MCS approach achieves a strong balance between scheduling quality and computational efficiency, with an of 0.87, an 80% success rate, and a low computation time (310 s), making it well‐suited for real‐time edge systems. Compared to other methods like PSO+, DP, and OneVM, MCS offers near‐optimal performance while avoiding the high computational cost and scalability limitations faced by those alternatives.

With the rapid growth of IoT technology and the rise of smart devices, edge computing, particularly federated learning (FL), has gained importance for preserving user data privacy. However, FL faces challenges like non-independent identically distributed data and device heterogeneity, leading to model disparities and reduced precision. Our research proposes a novel adaptive FL framework specifically engineered to synchronize computational capacities within heterogeneous edge computing landscapes. Building upon the proof of convergence boundaries for local aggregation model, this algorithm adapts the number of iterations for local updates by considering the resource consumption relationship between local aggregation model and the local updated model by various clients. This method exhibit adaptability within an environment where disparities in edge device computational capacities exist, effectively balancing computational prowess among diverse devices and enhancing the output performance of federated learning Experiments on MNIST and PlantVillage datasets show that in heterogeneous environments, our algorithm outperforms existing methods, improving the loss function by at least 16.87% and the convergence speed by at least 2 times, in various environments (MobileNet, AlexNet).

SCADA systems are widely used in industrial environments for centralized data monitoring and control, but their effectiveness is often limited to equipment with modern sensors. Many critical instruments, such as analog gauges and digital displays, still rely on manual readings due to the lack of digital interfaces, leading to inefficiencies and potential human error. Retrofitting these legacy instruments with IoT-enabled solutions is costly and disruptive. To address this challenge, we propose a cost-effective real-time data acquisition system that integrates legacy equipment into the SCADA systems using edge devices and deep learning. Our approach utilizes cameras and low-cost sensors to capture instrument readings, leveraging machine vision and indirect measurement techniques for accurate data extraction. Implemented on a Raspberry Pi 4B, the system employs the Fast-SCNN semantic segmentation model to identify key elements of analog gauges, achieving an average reading error of 1.5% at 2–3 frames per second. For digital meters, the PP-OCRv4 model achieves 95% accuracy on a custom dataset and 85.6% accuracy on Raspberry Pi 4B, operating at nearly 10 frames per second to enhance the reading efficiency. Additionally, an LSTM-based indirect temperature estimation model predicts liquid temperatures from external pipe wall readings, achieving the RMSEs of 2.92^∘C and 3.77^∘C on the validation and test sets, respectively. These findings highlight a practical and scalable solution for industrial automation, showcasing the potential of deep learning in edge computing for interfaceless instrument data acquisition.

Deploying deep neural networks (DNNs) on resource-constrained Internet of Things (IoT) devices is challenging due to limited processing power, energy constraints, and stringent latency requirements. This article introduces a new method to split DNN tasks in IoT systems, focusing on saving power and reducing delays. We use Gaussian process regression (GPR) to predict how long tasks will take under different conditions. We also use a simple linear regression model to estimate how much power IoT devices use based on their CPU usage. These predictive models are integrated into a Bayesian optimization framework to determine the optimal DNN task partitioning point, balancing latency and energy efficiency. The system adjusts to changes in the network and device conditions, ensuring it works well in different situations. Experiments on a heterogeneous IoT testbed demonstrate that GPR accurately predicts execution latencies, and the linear regression model provides reliable power consumption estimates. The Bayesian optimization algorithm efficiently explores the tradeoff space, offering low power consumption and high latency satisfaction rates. Our code is shared for public usehttps://github.com/nucleusbiao/Time-Sensitive-and-Power-Efficient-DNN-Task-Partitioning.