Unlock instant, AI-driven research and patent intelligence for your innovation.

Autonomous mobile robot positioning method based on deep learning

An autonomous movement and deep learning technology, applied in the field of robotics, can solve problems such as ranging data noise, flying pixels, robot positioning loss, errors, etc., and achieve the effects of simplifying the calculation process, improving accuracy and reliability, and improving precision

Active Publication Date: 2022-03-25
HANGZHOU LANXIN TECH CO LTD
View PDF6 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] Low-cost lidar (such as triangulation lidar) is commonly used in the consumer market. At the same time, a considerable number of robots are equipped with other types of ranging sensors, such as TOF cameras. Because the measurement principle is different from typical lidar , the accuracy and consistency of these distance sensors have a certain gap compared with typical lidar, and the noise and flying pixels in the distance measurement data are more serious. Therefore, if these low-precision distance sensors are used If the data is combined with the ordinary laser positioning algorithm for positioning calculation, a large positioning error will be generated, which will affect the autonomous navigation and walking of the robot. walk in a straight line
In severe cases, the robot's positioning will be lost

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Autonomous mobile robot positioning method based on deep learning
  • Autonomous mobile robot positioning method based on deep learning
  • Autonomous mobile robot positioning method based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0050] like Figure 4 As shown, the embodiment of the present invention provides a positioning method for an autonomous mobile robot based on deep learning. The execution subject of the method of this embodiment may be the control device of the autonomous mobile robot. The following steps are all corresponding to the online identification stage in the second embodiment The steps; its specific implementation method includes the following steps:

[0051] S10. Obtain the detection data of the autonomous mobile robot on surrounding obstacles at the current moment, and obtain the initial estimated position information of the autonomous mobile robot in the laser map at the current moment. The laser map is a pre-established grid based on the driving path area information of the autonomous mobile robot A map in lattice form.

[0052] In this embodiment, the detection data of surrounding obstacles by the autonomous mobile robot at the current moment can be obtained by means of a low-p...

Embodiment 2

[0063] like figure 1 As shown, the deep learning-based autonomous mobile robot positioning method of this embodiment can be divided into two stages: mapping and positioning; the two stages will be described in detail below. The autonomous mobile robot in the following embodiments may be an AGV. In this embodiment, an AGV is used as an example for illustration, but it is not limited to be just an AGV.

[0064] Mapping stage

[0065] The mapping stage can be divided into two processes: building a laser map and training a deep learning neural network.

[0066] First of all, ordinary laser radar and low-precision distance sensor are installed on the AGV. In practical applications, the height of laser radar and distance sensor is the same, and the installation position is close, that is, the height and position are close.

[0067] During the mapping process, the radar data of the ordinary lidar and the detection data of the low-precision distance sensor are saved at the same ti...

Embodiment 3

[0092] An embodiment of the present invention also provides an autonomous mobile robot, including a control device and a low-precision sensor, the low-precision sensor is connected to the control device, and the control device performs the autonomous robot based on deep learning described in any one of the first aspects above. Localization methods for mobile robots. In this embodiment, the installation positions of the low-precision sensor and the laser radar are close, and the heights are consistent.

[0093] In a third aspect, an embodiment of the present invention also provides a control device for an autonomous mobile robot, which includes: a memory and a processor, the memory is used to store computer programs, and the processor is used to execute the computer programs stored in the memory And execute the steps of the method for positioning an autonomous mobile robot based on deep learning described in any one of the first aspects above.

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention relates to an autonomous mobile robot positioning method based on deep learning, and the method comprises the steps: S10, obtaining the detection data of an autonomous mobile robot for surrounding obstacles at the current moment, and obtaining the initial estimation position information of the autonomous mobile robot in a laser map at the current moment, the laser map is a map which is pre-established based on the driving path area information of the autonomous mobile robot and is in a grid map form; s20, inputting the detection data and the initial estimation position information into a pre-trained deep learning neural network to obtain corrected detection data output by the deep learning neural network; s30, matching the corrected detection data with a laser map, and obtaining the positioning information of the autonomous mobile robot at the current moment; the deep learning neural network is a model which is built based on a UNET architecture and can receive the position information and the detection data and correct the detection data. The positioning method can effectively improve the positioning precision of the autonomous mobile robot.

Description

technical field [0001] The invention relates to the technical field of robots, in particular to a positioning method for an autonomous mobile robot based on deep learning. Background technique [0002] At present, the positioning methods mainly used in the field of autonomous mobile robots (such as AGV and AMR) include laser positioning, two-dimensional code positioning, magnetic stripe positioning, visual positioning, etc. Among them, laser positioning uses laser radar to obtain laser data of the surrounding environment. The data is the distance information from the surrounding obstacles to the laser radar. By matching the obstacle information with the environmental map, the precise pose of the robot in the map can be obtained. Since it does not need to add artificial markers, laser positioning is a natural navigation method, so , which has become the mainstream localization method for autonomous mobile robots in recent years and has been used more and more widely. [0003...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G01S17/06G01S17/89G06N3/04G06N3/08
CPCG01S17/06G01S17/89G06N3/08G06N3/045
Inventor 钱誉钦张易学周玄昊
Owner HANGZHOU LANXIN TECH CO LTD