Design, analysis and control for an antarctic modular manipulator

Antarctic scientific expedition has important strategic significance. It is an inevitable trend to apply robots to assist researchers during the Antarctic expedition. However, the robot manipula- tors at present have a series of problems and unable to meet the requirements of the Antarctic expe- dition. In this paper, a novel Antarctic modular robot manipulator is proposed, which has a compact structure with modular joints. The robot manipulator has high reliability, and quick assembling-and- disassembling ability. Through well wires arranging and thermal controlling, the manipulator can better adapt to the Antarctic environment. In addition, the work space of the manipulator is serious- ly analyzed, and a new hybrid position/force control method is adopted to make the manipulator per- form better. Simulation results validate the control method and show that the robot manipulator has a good performance to meet the requirements of Antarctic expedition.     

一种基于机器人的航天器大型部件自主装配方法

航天器大型部件在狭小空间中装配存在着视野受限、目标位置不可见等问题,传统的吊装方法无法精确调整位姿、磕碰风险极高．本文提出一种基于双目视觉定位的机器人辅助装配路径规划方法．该方法以航天器、待安装部件以及机器人等装配要素的3维模型为基础,采用双目视觉系统对航天器上安装位置的几何特征进行精确定位．利用测量结果确定机器人和航天器之间的相对位姿关系,进而构建航天器装配现场的虚拟环境,识别装配过程中的几何约束,然后利用轴对齐包围盒方法进行碰撞检测,采用随机路图法规划出一条无干涉的装配路径,最后利用离线编程技术生成机器人可执行的装配序列．以某型号航天器激光测高仪为对象开展装配试验,完成了大型部件在狭小凹舱内的无磕碰安装．结果表明该方法可以实现对螺纹孔精准定位,快速规划出无干涉装配路径,能够控制机器人安全、高效完成大型部件在狭小空间下的安装作业．     

Study on singularity analysis with differential transform method for 3-RRUR parallel robot based on Jacobian matrix

Parallel robot is used in many different fields nowadays,but the singularity of 3-RRUR parallel robot is more complicated,so a method to analyze the singularity of the 3-RRUR parallel robot is very necessary.First,the Jacobian matrix was built based on the differential transform method through the transfer matrixes between the poles.The connection between the position parameters and singularity condition was built through the analysis of the Jacobian matrix.Second,the effect on the singularity from the position parameters was analyzed,and then the singularity condition was confirmed.The effect on the singularity condition from position parameters was displayed by the curved surface charts to provide a basic method for the designing of the parallel robot.With this method,the singularity condition could be got when the length of each link is firmed,so it can be judged that if a group of parameters are appropriate or not,and the method also provides warrant for workspace and path planning of the parallel robot.     

基于动力学前馈的空间机器人多销孔装配力柔顺控制

多组销孔对接并固定是空间站舱外可移动仪器设备的最常用安装方式,其中多组销孔的自动对接是空间服务机器人能否代替航天员完成这一任务的关键。针对这一问题,首先对多销孔配合结构的特殊受力情况进行分析,提出适合的力柔顺控制方法,并针对微重力环境下机器人控制所面临的特殊问题,采用基于动力学前馈的控制方法,直接对关节进行力矩补偿,再基于所建立的机器人关节的数学模型,对前馈算法的补偿量进一步映射到电机电流环,从而进一步加快关节响应速度,保证力柔顺装配效果,最后分别针对所提出的力柔顺装配方法和前馈算法进行了模拟微重力试验和仿真试验,试验验证了所提出的针对空间多销孔对接的力柔顺装配控制方法的有效性。该方法为空间服务机器人安全完成舱外仪器设备安装任务提供了技术支撑,并可为空间服务机器人其他任务的安全稳定控制提供借鉴。     

Design and modeling for a kind of humanoid dexterous hand with elastic palm

With dexterous hands, robots can improve the work scope and work ability significantly. As palms of the existing multi-hand robots are made of steel plates that have small contact area, the robots cannot grab firmly. In this study, a new five-fingered dexterous robot hand is developed. Having flexible palm with 17 degree of freedoms （ DOFs）, the hand can grasp more stably and firm- ly. First, the forward kinematics and inverse kinematics of the fingers and the hand are calculated. Then, the connection between the force exerting on the end effectors and the torque exerting on the joint is set up, laying the foundation for the following control. Finally, through the analysis and sim- ulation of the position, velocity and acceleration, the trajectory planning has a better performance.     

基于改进Elman神经网络的红外被动测距算法研究

为了满足现代战争的需求,对于空中目标的距离精确估计显得愈发重要,由于目标的红外辐射在大气传播过程中的衰减,目标在不同波长的辐射会随着传输距离的改变而发生变化,故不同的波长的辐射里包含了目标的距离信息。基于以上原理,本文利用美国空军大气传输软件Modtran生成空中目标在不同波段的大气透过率数据,利用经纬仪获得目标的天顶角,最后建立基于改进Elman神经网络的被动测距模型。仿真结果表明本文算法能够有效提高对于空中目标距离估计的精确性。     

基于侧抑制网络的二维Otsu阈值分割算法

传统二维Otsu阈值分割算法未考虑人类视觉特性,分割结果不符合人眼视觉感受。为此,提出一种二维Otsu算法与侧抑制网络相结合的分割算法。该算法从基于人类视觉系统的侧抑制网络出发,利用侧抑制网络增强中心,抑制周围的特性,通过侧抑制网络处理原始图像,得到侧抑制图像,构建基于像素的灰度信息和侧抑制信息的二维直方图,并采用类间最大方差作为最佳阈值的选取准则。实验结果表明,与传统的Otsu算法和二维Otsu算法等相比,该算法具有较好的对比度、光照强度适应性和间断拟合能力,并能提高对图像噪声的鲁棒性,获得更理想的分割结果。     

基于柔顺控制的航天器大部件机器人装配技术

针对航天器大部件装配需求,给出一套基于力/位控制的机器人柔顺装配方法。基于装配界面作用力方向及机器人工具坐标系方向,实时计算用于力/位控制的笛卡尔坐标系。为避免力控制过程中的机器人抖动,设计了力与机器人速度的"S"型关系曲线。针对自由空间力控制、销钉导向等典型应用需求给出了力/位控制策略。针对机器人负载发生变化情况下的界面分离问题,对机器人在装配不同阶段下负载的质量特性进行预先测量,在界面分离时进行负载质量特性参数的切换,实现对应负载的重力补偿,并采用柔顺控制方法释放待分离界面的作用力,最终实现界面安全分离。试验结果表明,采用所设计的力/位控制方法,在工件安装的各主要过程中,工件可以稳定顺应外部约束,最终安装到位,在界面分离阶段,所述方法可以有效释放界面作用力,实现界面安全分离。