论文部分内容阅读
针对现有抓取系统中双目视觉定位精度较差,易受环境影响,使得抓取的成功率较低,鲁棒性不强的现象,本文提出采用视觉、红外测距传感器、触觉传感器和编码器等多传感器数据融合的方法,设计并实现了一种可靠的、鲁棒性强的、能自动调整抓取力的抓取系统.通过双目视觉辅以单目相机和红外测距传感器来精确定位,改善抓空情况;通过集成触觉传感器和编码器,对抓取过程中的力-位进行实时监测,减少目标物体破碎和滑落的现象,并通过实验证明了相对于单传感器,多传感器数据融合能大大改善抓取的成功率,提高系统的性能.
In view of the phenomenon that the binocular vision positioning accuracy is poor and vulnerable to the environment in the existing grasping system, the success rate of grasping is low and the robustness is not strong. In this paper, a visual, infrared ranging sensor, a tactile sensor, Encoder and other multi-sensor data fusion method, a reliable and robust grab system that can automatically adjust the grab power is designed and implemented.By binocular vision combined with monocular camera and infrared ranging sensor To accurately locate and improve the situation of grasping the void. By integrating the touch sensor and the encoder, the force-bit in the grasping process can be monitored in real time to reduce the crushing and slipping of the target object. Experiments show that compared with the single sensor, Sensor data fusion can greatly improve the success rate of crawling, improve system performance.