Robot Re-Localization by Movement Through Similar Environments

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing robot localization technologies face challenges in achieving accurate pose determination due to uncontrollable drift errors during SLAM processes, especially when robots are moved, suspended, or dragged, leading to inaccurate re-localization.

Innovation Solution

A method where the robot moves from its current position to a new position during the localization process, acquiring more environment information and comparing it with the stored environment map to identify its pose, thereby distinguishing similar regional environments and improving localization accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the robot stays at the original position for localization, then the localization process is simple and quick, but the localization accuracy is insufficient due to similar regional environments causing multiple possible poses

Engineering Contradiction:
Improvelocalization accuracyVSAvoidlocalization process complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent transitions from static localization (single position) to dynamic localization (multiple positions along a trajectory). By adding the temporal and spatial dimension of movement, the system collects environment information from multiple viewpoints, which resolves the ambiguity of similar regional environments and improves localization accuracy.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The system performs preliminary movement along a planned trajectory before final pose determination. This preliminary action of moving to multiple positions and collecting environment information in advance allows the robot to distinguish similar regions and accurately identify its pose in the global map.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If the robot moves to a new position during localization, then more environment information is acquired improving accuracy, but the localization time and movement complexity increase

Engineering Contradiction:
Improvelocalization accuracyVSAvoidlocalization time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent implements continuous environment information acquisition during the movement process. Instead of discrete sampling, the robot continuously collects data along the trajectory, which improves localization accuracy while minimizing unnecessary停顿 and reducing overall localization time.

Inventive Principle:
Principle #20Continuity of useful action

Solution Approach 2:

The system employs dynamic localization by moving the robot along a trajectory rather than keeping it static. This dynamic approach allows the robot to gather diverse environment information from multiple positions, improving the ability to distinguish similar regions and achieve accurate pose identification.

Inventive Principle:
Principle #15Dynamics

3Ease of operation

If SLAM technology is used for autonomous navigation, then the robot can navigate autonomously, but drift errors occur when the robot is moved, suspended, or dragged requiring re-localization

Engineering Contradiction:
Improveautonomous navigation capabilityVSAvoidlocalization reliability
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The patent implements a feedback mechanism where the robot compares environment information collected during movement with the pre-built global map. This feedback loop allows the system to detect drift errors and accurately re-localize by matching current observations with map data, thereby improving localization reliability after unauthorized movements.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3808512B1Robot localizing method, robot, and storage medium
Publication Date: 2024.09.25 ECOVACS ROBOTICS CO LTD
  • EP3808512B1 patent drawingFigure 1a~1c
  • EP3808512B1 patent drawingFigure 1d~1e
  • EP3808512B1 patent drawingFigure 1f~1g

AI summary

Provided is a method for localizing a robot. The robot may move from a current position to a new position during the localizing process, more environment information may be acquired during the new movement, and then the acquired environment information is compared with an environment map stored in the robot, which facilitates successfully localizing a pose of the robot in the stored environment map. In addition, during the movement and localization of the robot, environment information at different positions is generally different, so that similar regional environments may be distinguished, and the problem that an accurate pose cannot be obtained because there may be a plurality of similar regional environments when the robot stays at the original position for localizing may be overcome.