A mobile robot navigation system fusing vision and multi-line laser

CN117760444BActive Publication Date: 2026-05-29SOUTH CHINA UNIV OF TECH
2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
SOUTH CHINA UNIV OF TECH
Filing Date
2024-01-08
Publication Date
2026-05-29

AI Technical Summary

Technical Problem

In existing mobile robot navigation systems, a single sensor cannot effectively utilize multimodal information for environmental perception and cognition, resulting in insufficient navigation task planning and inflexible navigation relying on target points.

Method used

By employing a vision-multi-line laser fusion method, combined with natural language processing, feature extraction, and a cross-modal fusion network, a grid map is built using multi-line lasers and visual information is recorded. The cross-modal fusion network is then used to process natural language commands and target detection results, enabling accurate positioning and navigation of the target object.

Benefits of technology

It improves the mobile robot's perception and cognition of the environment, enabling it to efficiently complete navigation tasks based on natural language commands, adapt to complex environmental changes, and enhance the accuracy and flexibility of navigation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117760444B_ABST
    Figure CN117760444B_ABST
Patent Text Reader

Abstract

The application discloses a kind of visual and multi-line laser fusion mobile robot navigation system, comprising: mapping and positioning module, responsible for establishing grid map, setting navigation point and recording the environmental image of navigation point;Natural language processing module, pre-process natural language instruction, additional information of the target-related article in the environment is extracted through association model;Feature extraction module provides target detection information of target point for fusion network;Cross-modal fusion module, obtain candidate target navigation point sequence through fusion feature map;Target verification module, check and output final target sequence;Navigation and update module, execute navigation and update the weight of association model and the environmental image of navigation point.The application has both perception and cognitive ability, so that mobile robot can complete navigation task according to natural language instruction.
Need to check novelty before this filing date? Find Prior Art