基于多传感器信息融合的机器人障碍物检测
发布时间：2018-11-09 12:03
【摘要】：针对单一传感器在机器人避障过程中不能全面且准确定位障碍物的缺点,提出基于多传感器信息融合的障碍物检测方法。第一阶段使用视觉传感器检测未知环境中的障碍物,通过Zernike矩边缘检测方法提取障碍物图像边缘,然后采用Hough变换原理提取障碍物的直线特征,获得障碍物大概位置;第二阶段使用超声波传感器和红外传感器对障碍物进行检测,获得障碍物准确位置;最后使用联合卡尔曼滤波对3种传感器获得的信息进行融合,得出融合后的障碍物位置信息。实验结果表明:该方法克服视觉传感器、超声波传感器和红外传感器的局限性,可以精确感知机器人周围的未知环境信息,并能够检测和定位机器人路径上的障碍物,定位误差6 cm,满足机器人避障的实时性和可靠性需求。
[Abstract]:Aiming at the shortcoming that single sensor can not locate obstacles completely and accurately in the course of robot obstacle avoidance, an obstacle detection method based on multi-sensor information fusion is proposed. In the first stage, the obstacle in unknown environment is detected by visual sensor, and the edge of obstacle image is extracted by Zernike moment edge detection method. Then the linear feature of obstacle is extracted by Hough transform principle, and the approximate position of obstacle is obtained. In the second stage, ultrasonic sensors and infrared sensors are used to detect the obstacles and obtain the exact position of the obstacles. Finally, the information obtained from the three sensors is fused by using the combined Kalman filter, and the position information of the obstacle after fusion is obtained. The experimental results show that the method overcomes the limitations of vision sensors, ultrasonic sensors and infrared sensors, and can accurately perceive the unknown environment information around the robot, and can detect and locate obstacles on the robot path. The positioning error of 6 cm, can meet the requirements of real-time and reliability of robot obstacle avoidance.
【作者单位】： 太原理工大学信息工程学院;
【分类号】：TP242;TP391.41
[Abstract]:Aiming at the shortcoming that single sensor can not locate obstacles completely and accurately in the course of robot obstacle avoidance, an obstacle detection method based on multi-sensor information fusion is proposed. In the first stage, the obstacle in unknown environment is detected by visual sensor, and the edge of obstacle image is extracted by Zernike moment edge detection method. Then the linear feature of obstacle is extracted by Hough transform principle, and the approximate position of obstacle is obtained. In the second stage, ultrasonic sensors and infrared sensors are used to detect the obstacles and obtain the exact position of the obstacles. Finally, the information obtained from the three sensors is fused by using the combined Kalman filter, and the position information of the obstacle after fusion is obtained. The experimental results show that the method overcomes the limitations of vision sensors, ultrasonic sensors and infrared sensors, and can accurately perceive the unknown environment information around the robot, and can detect and locate obstacles on the robot path. The positioning error of 6 cm, can meet the requirements of real-time and reliability of robot obstacle avoidance.
【作者单位】： 太原理工大学信息工程学院;
【分类号】：TP242;TP391.41
【相似文献】
相关期刊论文 前10条
1 陈之中;吴训成;;倒车环境下的网状障碍物识别技术[J];上海工程技术大学学报;2012年01期
2 向荣;蒋荣欣;;铁路机车快速超视距障碍物识别算法[J];湘潭大学自然科学学报;2013年02期
3 匡佩;;一种基于运动模板的障碍物检测算法[J];科技传播;2012年03期
4 马赓宇,林学，
本文编号：2320360
本文链接：https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/2320360.html
