Robot Boundary-Walking Partition Planning Without Prior Map Storage

Solution Overview

Problem

Current laser SLAM sweeper cleaning methods face issues such as large differences between the outlined cleaning area and actual terrain, leading to inefficient navigation paths, excessive small areas, and slow cleaning due to the lack of a prior map, resulting in suboptimal coverage.

Innovation Solution

A cleaning partition planning method where a robot walks along the boundary, using laser map pixel information to identify and expand outline boundary segments, dividing the area into preset room cleaning partitions without requiring a pre-stored global map, and iteratively processing uncleaned areas to match the actual room boundary.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a rectangular frame area with M×N grid size is used for cleaning operation, then no prior map is needed, but the difference between the outline of the framed area and the actual terrain is large, causing more navigation paths and excessive small areas

Engineering Contradiction:

Improveability to operate without prior mapVSAvoidcleaning speed

Core Design Contradiction:

Adaptability or versatilityVSProductivity

Solution Approach 1:

The patent segments the cleaning area into multiple partitions based on actual terrain boundaries detected during operation. Instead of using a single fixed M×N grid, the system divides the space into smaller manageable sections that conform to actual walls and obstacles, allowing the robot to clean each partition efficiently while adapting to the real environment structure

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent implements dynamic area expansion where the cleaning boundary is not fixed but continuously adjusted based on detected terrain features. The robot starts with an initial framed area and dynamically expands or contracts the cleaning boundaries as it encounters actual walls and obstacles, transforming the static grid approach into a flexible adaptive system

Inventive Principle:

Principle #15Dynamics

2Adaptability or versatility

If a rectangular frame area with M×N grid size is used for cleaning operation, then no prior map is needed, but the outline of the framed area differs significantly from actual terrain, causing excessive small areas

Engineering Contradiction:

Improveability to operate without prior mapVSAvoidnumber of navigation paths

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent segments the cleaning area into multiple partitions based on actual terrain boundaries detected during operation. Instead of using a single fixed M×N grid, the system divides the space into smaller manageable sections that conform to actual walls and obstacles, allowing the robot to clean each partition efficiently while adapting to the real environment structure

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary boundary detection and partitioning before executing the main cleaning task. The robot first identifies actual terrain boundaries and creates an initial partition structure, then uses this pre-established framework to guide subsequent cleaning operations, avoiding the need to generate numerous small navigation paths during cleaning

Inventive Principle:

Principle #10Preliminary action

3Manufacturing precision

If outline boundary line segments are located according to pixel point statistical information, then the cleaning partition aligns with actual room boundary, but iterative expansion processing is required

Engineering Contradiction:

Improvealignment with actual room boundaryVSAvoidprocessing time for partition division

Core Design Contradiction:

Manufacturing precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary boundary detection and partitioning before executing the main cleaning task. The robot first identifies actual terrain boundaries and creates an initial partition structure, then uses this pre-established framework to guide subsequent cleaning operations, avoiding the need to generate numerous small navigation paths during cleaning

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent implements an iterative feedback mechanism where the robot continuously compares detected boundary features with the current partition model, adjusts the partition boundaries accordingly, and re-evaluates the configuration. This feedback loop ensures high alignment accuracy with actual room boundaries while systematically reducing the number of iterations needed through learned adaptation

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS11914391B2Cleaning partition planning method for robot walking along boundry, chip and robot

Publication Date: 2024.02.27 AMICRO SEMICONDUCTOR CO LTD

AI summary

The invention discloses a cleaning partition planning method for robot walking along the boundary, a chip and a robot. According to the cleaning partition planning method, a complete global map does not needed to be prestored in advance, but an initial room cleaning partition of the robot is divided in real time in a predefined cleaning area according to map image pixel information obtained by laser scanning in the process of walking along the boundary, meanwhile, the initial room cleaning partition of the robot is expanded by repeated iterative processing of the wall boundary of an uncleaned area in the same predefined cleaning area.