Method for evaluating the reliability of estimated distance types for laser distance sensor measurements and method for estimating the position of a mobile robot using the same

Localization technology evaluating sensor reliability through ray-tracing-based distance type estimation

This technology is a reliability evaluation method for mobile robot localization that estimates the distance type of laser range sensor measurements and calculates reliability weights for those types through sample pose-based reference distance calculation and distance error analysis.

Existing laser range sensor-based localization suffers from degraded scan-matching performance when measurements are distorted by dynamic obstacles, environmental changes, or optical anomalies such as glass and mirrors.

This technology proposes a method that extracts preliminary samples from an estimated pose to generate a ray-tracing-based reference distance set, estimates the distance type based on the error from the measured distance, and then calibrates the observation model's reliability by combining the frequency and deviation of the types. It can be applied to service robots in commercial facilities with many glass walls, fundamentally reducing localization failures caused by reflection and transmission.

‍

‍

Key Features:

• A step of extracting a plurality of preliminary samples based on the estimated pose of the current location
• A step of calculating a reference distance set by applying each preliminary sample to a reference distance calculation algorithm
• A step of estimating the distance type through the error between the reference distance set and the measured distance
• A configuration that calculates reliability weights by combining the frequency of the estimated type with the deviation from the reference distance

‍

This invention was developed through the following projects: the Intelligent Growing Autonomous Driving System for Unmanned Vehicles Operating Safely in Congested Residential Road Environments (Ministry of Science, ICT and Future Planning); the Development of Commercial-Grade Autonomous Driving Controllers for Unmanned Transport Robots in Diverse Environments (Ministry of Science, ICT and Future Planning); the Development of Learning-Based Robot Mobility Intelligence for Robust Indoor/Outdoor Integrated Autonomous Driving (Ministry of Trade, Industry and Energy); and the Agricultural Production Unmanned Automation Workforce Training and Research Support (Ministry of Agriculture, Food and Rural Affairs).