The number of agents in many multiagent systems in the real world, such as storage robots and drone cluster systems, continually changes. Still, most current multiagent reinforcement learning (RL) algorithms are limited to fixed network dimensions, and prior knowledge is used to preset the number of agents in the training phase, which leads to a poor generalization of the algorithm. In addition, these algorithms use centralized training to solve the instability problem of multiagent systems. However, the centralized learning of large-scale multiagent RL algorithms will lead to an explosion of network dimensions, which in turn leads to very limited scalability of centralized learning algorithms. To solve these two difficulties, in this article propose a group centralized training and decentralized execution-unlimited dynamic agent-number network (GCTDE-UDAN). First, since we use the attention mechanism to select several leaders and establish a dynamic number of teams, and the UDAN performs a nonlinear combination of all agents' Q values when performing value decomposition, it is not affected by changes in the number of agents. Moreover, our algorithm can unite any agent to form a group and conduct centralized training within the group, avoiding network dimension explosion caused by the global centralized training of large-scale agents. Finally, we verified on the simulation and experimental platform that the algorithm can learn and perform cooperative behaviors in many dynamic multiagent environments.

Quadruped robots have superior terrain adaptability and flexible movement capabilities than traditional robots. In this article, we innovatively apply it in person-following tasks, and propose an efficient motion planning scheme for quadruped robots to generate a flexible and effective trajectory in confined spaces. The method builds a real-time local costmap via onboard sensors, which involves both static and dynamic obstacles. And we exploit a simplified kinodynamic model and formulate the friction pyramids formed by ground reaction forces' inequality constraints to ensure the executable of the optimized trajectory. In addition, we obtain the optimal following trajectory in the costmap completely based on the robot's rectangular footprint description, which ensures that it can walk through the narrow spaces avoiding collision. Finally, a receding horizon control strategy is employed to improve the robustness of motion in complex environments. The proposed motion planning framework is integrated on the quadruped robot JueYing and tested in simulation as well as real scenarios. It shows that the execution success rates in various scenes are all over 90%. © 1996-2012 IEEE.

The number of agents in many multiagent systems in the real world, such as storage robots and drone cluster systems, continually changes. Still, most current multiagent reinforcement learning (RL) algorithms are limited to fixed network dimensions, and prior knowledge is used to preset the number of agents in the training phase, which leads to a poor generalization of the algorithm. In addition, these algorithms use centralized training to solve the instability problem of multiagent systems. However, the centralized learning of large-scale multiagent RL algorithms will lead to an explosion of network dimensions, which in turn leads to very limited scalability of centralized learning algorithms. To solve these two difficulties, in this article propose a group centralized training and decentralized execution-unlimited dynamic agent-number network (GCTDE-UDAN). First, since we use the attention mechanism to select several leaders and establish a dynamic number of teams, and the UDAN performs a nonlinear combination of all agents' Q values when performing value decomposition, it is not affected by changes in the number of agents. Moreover, our algorithm can unite any agent to form a group and conduct centralized training within the group, avoiding network dimension explosion caused by the global centralized training of large-scale agents. Finally, we verified on the simulation and experimental platform that the algorithm can learn and perform cooperative behaviors in many dynamic multiagent environments.

Quadruped robots have superior terrain adaptability and flexible movement capabilities than traditional robots. In this article, we innovatively apply it in person-following tasks, and propose an efficient motion planning scheme for quadruped robots to generate a flexible and effective trajectory in confined spaces. The method builds a real-time local costmap via onboard sensors, which involves both static and dynamic obstacles. And we exploit a simplified kinodynamic model and formulate the friction pyramids formed by ground reaction forces' inequality constraints to ensure the executable of the optimized trajectory. In addition, we obtain the optimal following trajectory in the costmap completely based on the robot's rectangular footprint description, which ensures that it can walk through the narrow spaces avoiding collision. Finally, a receding horizon control strategy is employed to improve the robustness of motion in complex environments. The proposed motion planning framework is integrated on the quadruped robot JueYing and tested in simulation as well as real scenarios. It shows that the execution success rates in various scenes are all over 90%. © 1996-2012 IEEE.

Manual assembly is still an essential part in modern manufacturing. Understanding the actual state of the assembly process can not only improve quality control of products, but also collect comprehensive data for production planning and proficiency assessments. Addressing the rising complexity led by the uncertainty in manual assembly, this paper presents an efficient approach to automatically capture and analyze hand operations in the assembly process. In this paper, a detection-based tracking method is introduced to capture trajectories of hand movement from the camera installed in each workstation. Then, the actions in hand trajectories are identified with a novel temporal action localization model. The experimental results have proved that our method reached the application level with high accuracy and a low computational cost. The proposed system is lightweight enough to be quickly set up on an embedded computing device for real-time online inference and on a cloud server for offline analysis as well. © 1982-2012 IEEE.

自主移动机器人或AGV是智能技术的一个重要发展方向。在技术层面，智能AGV的最关键功能是自定位。在复杂多变的作业环境下，基于GPS的定位技术无法满足导航精度，因此高精度地图是机器人定位的必要。但目前，在港口环境下，高精度地图构建以及AGV定位仍然存在诸多问题。针对这些问题，本报告将介绍如下几方面的贡献:1、第一章介绍地图构建与定位的基本框架，重点介绍地图模型，该模型面向长期可靠运行目标设计，充分考虑了环境的缓变特性，并且在地图难以精确构建的前提下，对定位的方式进行了思考和设计，从而提升鲁棒性。2、第二章介绍面向三维激光传感器信息的地点重识别问题，该问题解决了地图构建中的闭环检测，减少地图构建的误差累积。重点介绍了基于神经网络的地点重识别方法，并与其它算法进行了对比，本课题的算法在准确性和效率上都获得了提升。3、第三章介绍了在大范围场景下，如何压缩激光点云地图，以实现更高效率的自主定位。重点介绍了基于整数规划方法和随机森林的压缩框架，并在数据集上进行了完整测试。4、第四章为本报告的总结和结论部分，总结本报告所介绍的各方面技术及所取得的效果。 Autonomous mobile robot or automated guided vehicle(AGV)is an important development direction of intelligent techniques.At the technical level,the most critical task of intelligent AGV is self-positioning.In the complex and changeable environments,the GPS based positioning techniques can not meet the accuracy requirement for robot navigation,therefore,the high precision map is the necessary componnent for robot localization.But so far,in the port environment,there are still some problems with AGV localization and mapping with high precision.To solve these problems,this report will introduce the contribitions as follows:1、The basic framework of localization and mapping is introduced in the first chapter,focusing on the mapping model.This proposed model is designed for long-term operations,and with the full consideration of changeable environments.And on the premise that the map is difficult to construct accurately,we design the localization method,thus improving the robustness.2、In the second chapter,we introduce the the 3D LiDAR sensor based place recognition.The place recognition solvesthe loop closure detection in mapping,which can reduce the accumulated error in map construction.This method focuses on the deep learning based place recognition,with the comparison with other methods.This method improves the efficiency and effictiveness of place recognition.3、In the third chapter,we introduce how to compress 3D LiDAR map in large scale environments,thus achieving self-positioning with higher efficiency.This method focuses on the linear programming and random forest based method,and is validated on the datasets thoroughly.4、The final conclusions are presented in the fourth chapter,which include the introduced methods in this report.

In this letter, we develop a low-cost stereo visual-inertial localization system, which leverages efficient multi-state constraint Kalman filter (MSCKF)-based visual-inertial odometry (VIO) while utilizing an a priori LiDAR map to provide bounded-error three-dimensional navigation. Besides the standard sparse visual feature measurements used in VIO, the global registrations of visual semi-dense clouds to the prior LiDAR map are also exploited in a tightly-coupled MSCKF update, thus correcting accumulated drift. This cross-modality constraint between visual and LiDAR pointclouds is particularly addressed. The proposed approach is validated on both Monte Carlo simulations and real-world experiments, showing that LiDAR map constraints between clouds created through different sensing modalities greatly improve the standard VIO and provide bounded-error performance.

We propose an active learning-based grasp method for accurate industrial manipulation that combines the high accuracy of geometrically driven grasp methods and the generalization ability of data-driven grasp methods. Our grasp sequence consists of pregrasp stage and grasp stage which integrates the active perception and manipulation. In pregrasp stage, the manipulator actively moves and perceives the object. At each step, given the perception image, a motion is chosen so that the manipulator can adjust to a proper pose to grasp the object. We train a convolutional neural network to estimate the motion and combine the network with a closed-loop control so that the end effector can move to the pregrasp state. In grasp stage, the manipulator executes a fixed motion to complete the grasp task. The fixed motion can be acquired from the demonstration with nonexpert conveniently. Our proposed method does not require the prior knowledge of camera intrinsic parameters, hand-eye transformation, or manually designed feature of objects. Instead, the training data sets containing prior knowledge are collected through interactive perception. The method can be easily transferred to new tasks with a few human interventions and is able to complete high accuracy grasp task with a certain robustness to partial observation condition. In our circuit board grasping tests, we could achieve a grasp accuracy of 0.8 mm and 0.6°. Note to Practitioners-The research in this paper is motivated by the following practical problem. Manipulators on industrial lines can complete high accuracy tasks with hand-crafting features of objects. The perception is only used for object detection and localization. It is not flexible since the prior knowledge differs from tasks, which takes a long time to deploy in a new task. Besides, only well-trained experts are qualified to complete the deployment process. Our grasp method uses a convolutional network to estimate the motion for manipulator directly from images. The camera is mounted on the manipulator and can perceive the object actively. The training data set of the network is specific for different objects that can be automatically collected with a few human interventions. Our method simplifies the deployment process and can be applied in 3C industry (computers, communications, and consumer electronics) where the products upgrade frequently. © 2019 IEEE.

自主移动机器人或AGV是智能技术的一个重要发展方向。目前，世界制造业的生产模式正在面临着从批量生产向用户定制的转变，对智能、个性化制造有着强劲的需求。拥有自主移动技术的机器人，必将成为未来智能制造中不可替代的重要装备和自动化手段。除了制造业以外，先进移动操作机器人技术还可广泛应用于医院、家政等服务领域，以及港口运输、载人航天、救灾等特殊领域，具有广阔发展前景。在技术层面，智能AGV的最首要任务是解决自定位。在复杂多变的作业环境下，基于GPS的定位技术无法满足导航精度，因此高精度地图是机器人定位技术的必要。本报告本报告将介绍如下几方面的贡献: \t第一章介绍面向三维激光传感器信息的地点重识别问题，该问题解决了地图构建中的闭环检测，减少地图构建的误差累积。重点介绍了基于神经网络的地点重识别方法，并与其它算法进行了对比，本课题的算法在准确性和效率上都获得了提升。 \t第二章介绍面向异构地图的定位方法，该方法利用异构传感器信息之间的关联性使视觉能够在激光地图上定位，从而屏蔽视觉传感器在实际环境中存在的大量环境变化因素，从而使基于视觉的长期定位方法成为可能。该算法在室外环境中实现了时间跨度一年的定位。 \t第三章为本报告的总结和结论部分，总结本报告所介绍的各方面技术及所取得的效果。 Autonomous mobile robots or AGVs are an important development direction of intelligent technology.At present,the production mode of the world's manufacturing industry is facing a shift from mass production to customization,and there is a strong demand for intelligent and personalized manufacturing.Robots with autonomous mobile technology will become an irreplaceable important equipment and automation in the future of intelligent manufacturing.In addition to manufacturing,advanced mobile robot technology can also be widely used in hospitals,home economics and other service areas,as well as special areas such as port transportation,manned spaceflight,disaster relief,etc.,with broad development prospects.At the technical level,the most important task of intelligent AGV is to solve self-positioning.In the complex and varied working environment,GPS-based positioning technology can not meet the navigation accuracy,so high-precision map is necessary for robot positioning technology.This report will present the following contributions in this report: The first chapter introduces the location re-identification problem for 3D laser sensor information.This problem solves the closed-loop detection in map construction and reduces the error accumulation of map construction.The method of location re-identification based on neural network is introduced,and compared with other algorithms,the algorithm of this subject has been improved in accuracy and efficiency. The second chapter introduces the location method for heterogeneous maps,which uses the correlation between heterogeneous sensor information to enable visual positioning on the laser map,thereby shielding the large number of environmental changes that the visual sensor has in the actual environment.Make vision-based long-term positioning methods possible.The algorithm achieves a one-year positioning of time span in an outdoor environment. Chapter III is the summary and conclusion part of this report.It summarizes the various technologies and effects achieved in this report.

We propose an end to end deep learning approach for generating real-time facial animation from just audio. Specifically, our deep architecture employs deep bidirectional long short-term memory network and attention mechanism to discover the latent representations of time-varying contextual information within the speech and recognize the significance of different information contributed to certain face status. Therefore, our model is able to drive different levels of facial movements at inference and automatically keep up with the corresponding pitch and latent speaking style in the input audio, with no assumption or further human intervention. Evaluation results show that our method could not only generate accurate lip movements from audio, but also successfully regress the speaker's time-varying facial movements. © 2019 IEEE.