Obstacle avoidance tracking control with antiswing and tracking errors constraint for underactuated automated lifting robots with load hoisting/lowering

The existing automated lifting robot technology focuses merely on motion control and ignores the surrounding environment. In practice, obstacles inevitably exist in the movement path of the automated lifting robot, which affects construction safety. Furthermore, due to the underactuated characteristics of the automated lifting robot, the load can be difficult to control when it swings violently, which undoubtedly poses huge challenges to obstacle avoidance trajectory planning and controller design. In this paper, an obstacle avoidance trajectory and its tracking controller with antiswing and tracking errors constraint are proposed. To ensure accurate load positioning and effective obstacle avoidance, the proposed control method introduces a four-segment polynomial trajectory interpolation curve to construct an obstacle avoidance trajectory based on analyzing the geometric relationship between variables. To improve the transient coupling control performance of the system, combined with the passive analysis of the automated lifting robot system, this method constructs a potential function that limits the tracking error and a coupling signal that enhances the coupling relationship between the system variables. Barbalat's lemma and Lyapunov techniques are used to analyze the stability of the system. Simulation and experimental results show that the proposed control method can significantly suppress or even eliminate load oscillation, accurately locate the load, avoid obstacles, improve the safety and efficiency of the working automated lifting robot, and have strong robustness to changes in system parameters and the addition of external disturbances.     

基于三阶超螺旋滑模观测器算法的机器人力矩估计方法

本文研究了六自由度(6-DOF)机器人关节力矩估计问题.为了提高机器人关节力矩估计精度,本文设计了一种三阶超螺旋滑模关节力矩观测器.在对观测器系统进行稳定性分析时,本文设计了一种新的Lyapunov函数证明了观测器闭环系统的稳定性.与已有的机器人力矩观测器结果不同,该力矩观测器仅以机器人关节角度作为观测器输入信号,减少了对机器人系统信息的依赖.最后,以6-DOF工业机器人为对象,开展实验验证.实验结果表明,与传统线性力矩观测器相比,本文提出的滑模力矩观测器估计误差小,验证理论的正确性.     

融合深度学习与成像模型的水下图像增强算法

水下机器人的视觉感知功能因受到水下环境因素的影响,面临着图像质量降低的挑战,如图像颜色畸变、整体色调偏绿、偏蓝、对比度较低、细节较为模糊等。提出一种结合深度学习方法与物理成像模型的新型水下图像增强算法,通过构建包含扩张卷积和带参数激活函数的神经网络,进行背景散射光和直接传输映射的估计,并结合成像模型的数学表达进行重建运算得到增强后图像。实验结果表明,与UDCP、IBLA、GLNet等典型图像增强算法相比,该算法具有更快的运算速度,且能够消除水下环境因素带来的影响,丰富图像色彩的同时能增强各类细节,在峰值信噪比指标和结构相似度指标上取得了较大值。此外,增强后的图像在特征点匹配实验中获得了更好的匹配效果。     

地下管道检测移动机器人的发展现状

简要介绍了管道检测移动机器人的发展状况,并对管道检测移动机器人的发展趋势及关键技术进行了探讨。     

基于全向视觉的移动机器人实时全局地图构建

实时全局地图构建是实现移动机器人智能化的关键;研究了一种基于全向视觉的移动机器人全局地图的实时构建方法.首先介绍了全向视觉的体系结构,然后对图像处理软件的相关模块,包括基于颜色的阈值分割、区域连通、特征提取和目标识别等进行了说明,最后通过坐标的转换实现全局地图的构建;实验结果表明,由于充分利用了全向视觉的特性,构建的全局地图准确性高、实时性好,完全能够满足移动机器人对环境建模的需要.     

不确定移动机器人编队间接自适应模糊动力学控制

研究不确定非完整移动机器人Leader/Follower编队动力学控制问题,对Follower提出了自适应队形跟踪控制方法.首先,基于队形方程设计了运动学队形跟踪控制器,使用Backstepping技术扩展出队形跟踪误差动力学方程;然后,利用自适应模糊系统逼近其不确定项,构造了间接自适应模糊动力学队形跟踪控制器;最后,证明了队形跟踪误差可收敛到原点的小邻域内.仿真结果验证了所提出控制策略的有效性.     

Improving the authentic learning experience by integrating robots into the mixed-reality environment

The main aim of the modern popular teaching method of authentic learning has been to provide students with everyday-life challenges that develop knowledge and skills through problem solving in different situations. Many emerging information technologies have been used to present authentic environment in pedagogical purpose. However, there are few studies that have been discussed the sense of authenticity and characters in scene and how students interact with the characters involved in the task. We designed a system, RoboStage, with authentic scenes by using mixed-reality technology and robot to investigate the difference in learning with either physical or virtual characters and learning behaviors and performance through the system. Robots were designed to play real interactive characters in the task. The experiment of the study conducted with 36 junior high students. The results indicated that RoboStage significantly improved the sense of authenticity of the task and also positively affected learning motivation. Learning performance was conditionally affected by RoboStage.     

智能科学与技术专业毕业生情况分析与专业建设

针对南开大学"智能科学与技术"新兴本科专业,对首届学生的来源和去向进行整理和分析,根据毕业生的座谈与反馈情况,提出南开大学"智能科学与技术"专业的调整方向与后续建设目标。     

Switching fuzzy tracking control for mobile robots under curvature constraints

In this paper a switching fuzzy logic controller for mobile robots with a bounded curvature constraint is presented. The controller tracks piece-wise linear paths, which are an approximation of the feasible smooth reference path. The controller is constructed through the use of a map, which transforms the problem to a simpler one; namely the tracking of straight lines. This allows the use of an existing fuzzy tracker deployed in a previous work, and its simplification leading to a 70% rule reduction. Simulation results and a comparison analysis with existing trackers are also presented along with some stability considerations on the impulsive error dynamics which emerge.     

基于关键帧相似性的仿人机器人步态规划

为解决传统步态规划方法运算量大、运行时间长的问题,提出一种基于关键帧相似性的仿人机器人步态规划方法,通过对步态关键帧的相似性度量、髋关节的规划、零力矩点(ZMP)的实时校正,实现仿人机器人曲线行走步态的在线调整,调整后的步态能够较好地逼近实际ZMP曲线。实践结果表明,该方法在降低运算量的同时较好解决了仿人机器人的步态稳定及快速规划问题,并且可以减少步态选择不理想时的时间消耗。