Laser Range Finder Floor Obstacle Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods using 2D laser range finders struggle to detect small floor obstacles on a mobile robot's driving path, as they rely on heuristic thresholds and require extensive training data, making it difficult to accurately classify drivable areas and detect small obstacles like boxes or protruded tiles.
Innovation Solution
A method involving generating normal floor characteristic data using a pre-registered one-class classification method, such as Chi-square test or SVDD, to determine if sensing values from a laser range finder indicate a normal driving surface or a floor obstacle, with correction for bias errors to enhance detection precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a 2D laser range finder is used for terrain detection, then data processing becomes easier and detection precision improves, but the detection range is limited and small floor obstacles cannot be effectively detected
Solution Approach 1:
The patent transforms the 2D laser range finder's limited detection capability into effective small obstacle detection by inclining the sensor and projecting terrain data into a 2D elevation map with height information. This dimensional transformation allows the system to detect small floor obstacles that would be invisible to a standard 2D scanner, converting a limitation into an advantage for precision detection.
2Area of stationary object
If a 3D laser range finder is used to detect wide range terrain, then detection range improves, but it becomes difficult to detect terrain near the mobile robot and system cost increases
Solution Approach 1:
The patent applies local quality by inclining the 2D laser range finder specifically to optimize detection of near-field terrain and small obstacles. Instead of using a costly 3D scanner for all scenarios, the system tailors the 2D scanner's orientation to excel at close-range detection, achieving localized optimization where precision is most needed.
3Ease of manufacture
If heuristic threshold methods are used for terrain classification, then implementation simplicity improves, but detection reliability deteriorates due to environment-specific parameter redefinition
Solution Approach 1:
The patent transforms terrain classification from heuristic threshold methods to a statistical parameter-based approach using Mahalanobis distance. By changing the classification parameters from fixed thresholds to dynamic statistical measures that account for data distribution and correlation, the system achieves environment-independent reliable detection while maintaining computational efficiency.
4Measurement precision
If grid map resolution is increased to detect smaller obstacles, then obstacle detection capability improves, but data processing complexity and computational load increase
Solution Approach 1:
The patent extracts only the essential height information from laser scan data to create a 2D elevation map, rather than processing complete 3D point clouds. This extraction approach maintains the ability to detect small obstacles by preserving vertical dimension data while significantly reducing computational complexity compared to full 3D processing.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach effectively detects small floor obstacles, improving the stability of mobile robot navigation by accurately identifying drivable areas and correcting for laser range finder biases, thereby preventing driving failures from small obstacles.
Implementation Method 1
a laser range finder for detecting a distance to an object
Data Source
AI summary
A method detecting a floor obstacle using a laser range finder according to the present invention includes the following steps: (a) generating normal floor characteristic data with regard to a flat normal driving surface having no floor obstacle; (b) registering the normal floor characteristic data on a pre-registered one-class classification method; (c) obtaining sensing value of the laser range finder according to the driving of a mobile robot; (d) generating sensing value-floor characteristic data with respect to the sensing value; and (e) determining whether the sensing value indicates a normal driving surface or a floor obstacle by applying the sensing value-floor characteristic data to the one-class classification method. Therefore, an obstacle including a relatively small floor obstacle existing on the driving path of a mobile robot can be detected more effectively using a laser range finder, thereby providing more stably an area where the mobile robot can travel.