An autonomous obstacle avoidance method and system for industrial robots

By using a stacked, separate convolutional unit obstacle recognition model and Riemannian metric tensor technology, the problem of obstacle avoidance delay in industrial robots in complex dynamic environments is solved, achieving safe and efficient path planning and obstacle avoidance.

CN122308434APending Publication Date: 2026-06-30SHANGHAI SENIOR TECH SCHOOL
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
SHANGHAI SENIOR TECH SCHOOL
Filing Date
2026-04-28
Publication Date
2026-06-30

AI Technical Summary

Technical Problem

Existing autonomous obstacle avoidance methods for industrial robots rely on simple grid maps or heuristic path planning, which cannot continuously optimize paths in complex dynamic environments. This leads to obstacle avoidance delays, frequent path corrections, and even obstacle avoidance failures, affecting work safety and production efficiency.

Method used

An obstacle recognition model employing stacked separable convolutional units analyzes environmental images in real time, constructs a dynamic risk density function and maps it to the configuration space of an industrial robot, establishes geodesic equations through Riemannian metric tensors, transforms the path planning problem into a geodesic solution problem, and obtains the optimal obstacle avoidance path.

Benefits of technology

It accurately identifies obstacles and their dynamic movement trends with low computational load, achieving smooth and continuous movement, significantly reducing obstacle avoidance delay and path correction frequency, and improving safety and production efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122308434A_ABST
    Figure CN122308434A_ABST
Patent Text Reader

Abstract

This invention provides an autonomous obstacle avoidance method and system for industrial robots, relating to the field of industrial control technology. The method includes: acquiring multiple frames of environmental images of the industrial robot along a preset travel path; determining whether obstacles exist in each frame of environmental images; if so, retaining multiple consecutive environmental images containing obstacles; based on the retained environmental images, constructing a dynamic risk density function that fuses the current position and predicted displacement of the obstacle; mapping the dynamic risk density function to the configuration space of the industrial robot; establishing a Riemannian metric tensor that dynamically evolves over time; establishing a geodesic equation to transform the path planning problem in the autonomous obstacle avoidance process of the industrial robot into a geodesic solution problem; acquiring the current configuration and target configuration of the industrial robot; using the current configuration and target configuration as boundary conditions, solving the geodesic equation and outputting the optimal obstacle avoidance path; guiding the industrial robot to avoid obstacles according to the optimal obstacle avoidance path; otherwise, re-acquiring images.
Need to check novelty before this filing date? Find Prior Art