Method for diagnosing localization status and autonomous mobile robot performing the same

Localization Status Diagnosis for Autonomous Mobile Robots Using Machine Learning-Based Distance and Heading Error Metrics

This technology is a localization status diagnosis method that determines whether an autonomous mobile robot has successfully localized itself by calculating distance error metrics based on range sensors and heading error metrics based on odometry, then inputting these into a supervised binary classification algorithm for self-diagnosis.

Existing localization diagnosis methods have faced challenges with high dependency on specific algorithms or fluctuating sensor data reliability depending on environmental conditions, making it difficult to achieve universal and robust diagnosis.

This technology proposes a method that defines distance error metrics using the average error of highly reliable range measurements and heading error metrics based on tolerance ranges, utilizing a trained binary classification model to determine success or failure in real time. It can be applied to indoor service robots and logistics robots, significantly enhancing operational stability by enabling the robot to detect when it has lost its position and initiate recovery procedures.

‍

‍

Key Features:

• A training phase where distance error metrics, heading error metrics, and localization success status are registered as training data.
• A sensing phase where multiple distance measurements are collected via range sensors at the robot's current position.
• A calculation phase where the robot's estimated heading at its current position is derived from odometry information.
• A diagnostic configuration that applies metrics calculated from measurements to a machine learning model to determine the success of localization.

‍

This invention was developed with support from the Ministry of Science and ICT for the "Intelligent Growth Autonomous Driving System for Unmanned Vehicles Operating Safely in Congested Living Road Environments" project, and the Ministry of Agriculture, Food and Rural Affairs for the "Agricultural Production Unmanned Automation Workforce Training and Research Support" project.