基于光纤微弯曲效应的滑觉传感器研制

发布时间：2018-12-16 17:57

【摘要】：随着机器人技术的发展,机械手被广泛应用于物体拾放。利用滑觉能准确检测抓取物体所需的力或力矩,提高机械手抓取的可靠性。现有的滑觉传感器主要采用加速度、视觉识别、电压检测等原理。但由于受到技术水平和制造成本的限制,滑觉传感器的研究总体仍处于探索阶段。论文提出了一种基于光纤微弯效应的滑觉传感器,用来检测抓取物体时的相对滑动。传感器基于光纤微弯损耗原理,通过“力信号——位移信号——光通量——电压信号”的转化从而测量物体与机械手之间的摩擦力。设计了一种悬臂弹力齿板结构实现力信号——位移信号的转化,优点在于形变量较小且有较好的重复精度。进一步分析位移信号——光通量损耗之间的关系,确定了悬臂弹力齿板的重要参数,包括:齿距、齿廓半径、弹性悬臂组的胡克系数等。通过光电转换电路把光通量转换成电压信号。以上系统组成一个测量单元。传感器由两个相互正交的测量单元组成,分别用来测量X/Y方向的两个摩擦力。为精确地获得传感器的力信号和电压信号之间的关系,需要进行标定试验。首先分别对X/Y方向的两个单元单独进行标定,结果得到力信号和电压呈现明显的线性相关,相关系数分别为0.99963与0.99869。系数不同的原因是两传感器单元弹性悬臂组的胡克系数不同。接着对组合的传感器在各个方向上进行标定,发现由于摩擦力的存在,实验结果存在误差。进一步地对摩擦力进行补偿后,实验表明,负载与X轴夹角在30°~60°范围内时,总体误差在±2%以内,能满足传感器检测需要,表明该传感器设计方案是有效的。搭建试验平台模拟机械手抓取试验装置,测量传感器与物体间的摩擦力,包括:静摩擦力和滑动摩擦力。试验结果显示当传感器与物体间从静摩擦力转变为滑动摩擦力时,电压信号存在一个明显的下降沿。当加载载荷为16N时,下降沿的电压降约为1V。因此可以很容易地被检测出物体和传感器之间发生滑动的时刻。
[Abstract]:With the development of robot technology, manipulator is widely used in object picking. The slip sense can accurately detect the force or torque needed to grab the object and improve the reliability of the manipulator. The existing slip sensor mainly adopts the principles of acceleration, vision recognition, voltage detection and so on. However, due to the limitation of technology level and manufacturing cost, the research of slide sensor is still in the exploration stage. In this paper, a slide-sensing sensor based on optical fiber micro-bending effect is proposed to detect the relative slippage of an object. Based on the principle of optical fiber micro-bending loss, the sensor measures the friction between the manipulator and the object through the conversion of "force signal, displacement signal, luminous flux voltage signal". A cantilever elastic-tooth plate structure is designed to realize the transformation of force signal, displacement signal, which has the advantages of small deformation and good repetition accuracy. Based on the analysis of the relationship between the displacement signal and the luminous flux loss, the important parameters of the cantilever elastic tooth plate are determined, including the tooth pitch, the radius of the tooth profile, the Hook coefficient of the elastic cantilever group, and so on. The flux of light is converted into a voltage signal by the photoelectric conversion circuit. The above system constitutes a measuring unit. The sensor consists of two orthogonal measuring units used to measure two friction forces in the X / Y direction. In order to obtain the relationship between the force signal and the voltage signal accurately, calibration experiments are needed. First, the two units in the X / Y direction are calibrated separately, and the results show that the force signal and the voltage show obvious linear correlation, the correlation coefficients are 0.99963 and 0.99869 respectively. The difference in coefficient is due to the difference of Hook coefficient in the elastic cantilever group of two sensor elements. Then the combined sensors are calibrated in various directions and the experimental results are found to be error due to the existence of friction. After the friction is compensated further, the experiment shows that the overall error is within 卤2% when the angle between the load and the X axis is 30 掳~ 60 掳, which can meet the needs of the sensor detection, which shows that the sensor design scheme is effective. A test platform was built to simulate the grab test device of manipulator to measure the friction between the sensor and the object, including static friction and sliding friction. The experimental results show that when the static friction force is changed to the sliding friction force between the sensor and the object, there is an obvious drop edge in the voltage signal. When the load is 16 N, the voltage drop along the drop edge is about 1 V. Therefore, it is easy to detect the time of sliding between the object and the sensor.
【学位授予单位】：浙江理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP212

【参考文献】