ROS交流群
ROS Group
产品服务
Product Service
开源代码库
Github
官网
Official website
技术交流
Technological exchanges
激光雷达
LIDAR
ROS教程
ROS Tourials
深度学习
Deep Learning
机器视觉
Computer Vision

The visual navigation algorithm was first applied to the production environment.



  • 2016.7.19
    Recently received a robot order from the user. The user needs to make an exhibition robot and put it in his showroom. The entire exhibition hall is about 30 meters long, and the products of the user exhibition are arranged in turn on the side of the exhibition hall. Users need robots to stop at the products of each exhibition and introduce the product. At present, this user’s robot uses a tracking method, which puts a black route on the floor and indicates the first few products at that time. However, the user is very dissatisfied with these black lines and feels that the atmosphere of the entire exhibition hall is destroyed.

    This situation must use our visual positioning system. However, our visual positioning system is not yet mature, and the program can be accurately and stably operated. But since users have such a demand, they can only try it.

    The video above is a visual display of visual navigation. It can be seen that the visual positioning is highly accurate and works very stable. The program sets the car to run in the same direction as the black line. No matter how humans interfere, the car can return to its original direction. This is not possible with inertial navigation using gyroscopes and encoders. Inertial Navigation Once the wheels slip, the positioning creates an unrepairable error.

    Above is a broader test video. So far, Xiaoqiang’s visual navigation system has finally been used in the production environment.