Mobile Robot Pose Estimation Using Point Cloud and IMU Fusion

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing mobile robot attitude detection methods using inertial measurement units (IMUs) are inaccurate due to large error values, especially when encountering potholes or doorsills, affecting positional posture and obstacle sensing.

Innovation Solution

A posture estimation method combining point cloud data from a linear laser and IMU data to accurately determine the mobile robot's attitude, using least squares optimization to correct the robot's posture in the world coordinate system.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If inertial measurement unit is used for attitude detection, then the robot can perform planar tasks, but the measurement precision is low due to large error values

Engineering Contradiction:
Improveattitude measurement precisionVSAvoidrobot posture accuracy
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent combines point cloud data from linear laser with attitude data from inertial measurement unit to perform joint optimization. This merging of multiple data sources compensates for the limitations of each individual sensor, achieving high-precision posture estimation that overcomes the low measurement precision of the IMU alone.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent introduces a work plane model as an intermediary between the IMU and the final posture estimation. The point cloud data serves as intermediate information that helps establish the work plane, which then acts as a reference for correcting the IMU measurements and achieving accurate attitude detection.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Adaptability or versatility

If robot encounters potholes or doorsills, then the robot may bump and change attitude, but this causes incorrect obstacle marking and inaccurate positional posture

Engineering Contradiction:
Improverobot ability to handle complex terrainVSAvoidpositional posture accuracy
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent dynamically adjusts the work plane model based on real-time point cloud data. When the robot encounters potholes or doorsills causing attitude changes, the system updates the work plane representation to reflect the new terrain conditions, maintaining accurate posture estimation despite the dynamic environmental changes.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent implements a feedback mechanism where the point cloud data continuously validates and corrects the work plane model. When the robot encounters unexpected terrain features, the point cloud provides feedback that triggers recalculation of the work plane, ensuring the posture estimation remains accurate even after bumps or attitude changes.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP4644838A1Pose estimation method and apparatus of mobile robot, mobile robot and medium
Publication Date: 2025.11.05 BEIJING ROBOROCK INNOVATION TECH CO LTD
  • EP4644838A1 patent drawingFigure 1
  • EP4644838A1 patent drawingFigure 2-a~2-b
  • EP4644838A1 patent drawingFigure 3-a~3-b

AI summary

A pose estimation method and apparatus for a mobile robot, the mobile robot, and a storage medium. The method comprises: obtaining an environment three-dimensional point cloud when a mobile robot executes a planar task, and on the basis of a preset height threshold, determining a working plane point cloud according to the environment three-dimensional point cloud (S110); obtaining inertial measurement information of the mobile robot (S120); and according to the inertial measurement information and the working plane point cloud, determining the pose of the mobile robot in a world coordinate system (S130).