赵可可, 柴志雷, 吴东. 一种基于Zynq的ROS软硬件协同计算架构设计与实现[J]. 微电子学与计算机, 2017, 34(9): 87-91.
引用本文: 赵可可, 柴志雷, 吴东. 一种基于Zynq的ROS软硬件协同计算架构设计与实现[J]. 微电子学与计算机, 2017, 34(9): 87-91.
ZHAO Ke-ke, CHAI Zhi-lei, WU Dong. Design and Implementation of a ROS Software-hardware Co-processing Compute Architecture Based on Zynq[J]. Microelectronics & Computer, 2017, 34(9): 87-91.
Citation: ZHAO Ke-ke, CHAI Zhi-lei, WU Dong. Design and Implementation of a ROS Software-hardware Co-processing Compute Architecture Based on Zynq[J]. Microelectronics & Computer, 2017, 34(9): 87-91.

一种基于Zynq的ROS软硬件协同计算架构设计与实现

Design and Implementation of a ROS Software-hardware Co-processing Compute Architecture Based on Zynq

  • 摘要: 针对复杂机器人算法在嵌入式系统上运行时间过长的问题, 设计并实现了一种在Zynq平台上通过软硬件协同机制实现机器人操作系统(ROS) 节点的方法, 利用FPGA加速ROS中的复杂图像算法.ARM端软件进行接口间的转换, 可以和其他ROS节点正常通讯.随后测试了双目匹配算法census在系统中的实现.实验证明, 在处理复杂图像算法时相对于纯软件实现具有明显优势.

     

    Abstract: Aiming to solve the robot complex algorithm running time is too long in the embedded system, a node method designed by software-hardware co-processing in the Zynq for robot operating system (ROS), Complex image algorithm of ROS in FPGA accelerated and ARM handle software interface between the ROS and other normal communication node.Then the stereo matching algorithm census is tested in the system.Experimental results show that in dealing with complex image algorithm relative to the pure software implementation has obvious advantages.

     

/

返回文章
返回