Dual 2D Laser Radar Obstacle Detection for Autonomous Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional obstacle detection methods for autonomous mobile vehicles, particularly on non-paved roads or fields, incorrectly identify inclined ground as obstacles due to abrupt changes in direction or speed, leading to unnecessary avoidance maneuvers.
Innovation Solution
The use of two 2D laser radars positioned at different orientations and a processing unit to calculate the actual inclination of detected objects, distinguishing between obstacles and ground by comparing this inclination with a reference value, and adjusting distance data to prevent false obstacle recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional obstacle detection methods are used with a single 2D laser radar, then the system is simple and fast, but the vehicle is erroneously detected as having obstacles when inclined on non-paved roads
Solution Approach 1:
The system divides the detection task by using multiple 2D laser radars positioned at different locations and orientations on the vehicle. Each radar captures distance data from a specific viewpoint, and the processing unit integrates these segmented measurements to calculate actual ground inclination, resolving the contradiction between simple detection and reliable obstacle identification on inclined terrain
Solution Approach 2:
The processing unit acts as an intermediary that receives distance data from multiple laser radars, calculates vehicle inclination angle, and compensates the distance measurements to determine whether detected objects are true obstacles or false positives caused by vehicle inclination. This intermediary computation layer enables reliable detection without requiring complex sensor hardware
2Reliability
If multiple sensors are used to generate a 3D World Model, then obstacle detection is more accurate, but the calculation time increases significantly
Solution Approach 1:
The system extracts only the essential information needed for obstacle detection by using multiple 2D laser radars to capture distance data, then selectively processing this data to calculate vehicle inclination angle and compensate measurements. This extraction approach avoids the computational burden of generating complete 3D World Models while maintaining sufficient detection accuracy for the application
Solution Approach 2:
Instead of performing complete 3D World Model generation with all available sensors, the system applies partial action by using a simplified processing approach that focuses only on calculating vehicle inclination and compensating distance measurements. This partial processing reduces calculation time while maintaining the reliability needed for obstacle detection on non-paved roads
3Speed
If the vehicle moves abruptly on non-paved roads, then the vehicle responds quickly to terrain changes, but the laser radar erroneously detects ground as obstacles due to vehicle inclination
Solution Approach 1:
The system performs preliminary action by continuously calculating the vehicle inclination angle from multiple laser radar measurements before making obstacle detection decisions. This preliminary inclination calculation compensates for the effects of abrupt vehicle movements on non-paved roads, enabling the system to maintain reliable ground detection accuracy even during rapid vehicle responses to terrain changes
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 reduces false obstacle detection, enhances autonomous driving performance on non-paved surfaces by accurately identifying road inclination and irregularities, and provides a cost-effective solution using only two laser radars and a processing unit.
Implementation Method 1
distance data obtained by the 2D laser radar
Implementation Method 2
distance data obtained by the 2D laser radar which is arranged in parallel with the horizontal plane
Data Source
AI summary
Disclosed is apparatus for distinguishing between ground and an obstacle for autonomous mobile vehicle, comprising an upper 2D laser radar 1, a lower 2D laser radar 2, and a processing unit 10, the processing unit 10 comprising a distance data receiving part 11, an inclination calculating part 12, a ground and obstacle determining part 13, and a transmitting part. Also disclosed is a method for distinguishing between ground and an obstacle for autonomous mobile vehicle by using the apparatus for distinguishing between ground and an obstacle for autonomous mobile vehicle of claim 1, in which the detected object is determined as an obstacle when the actual inclination (g) of the detected object is larger than the reference inclination, and as ground when the actual inclination (g) of the detected object is smaller than the reference inclination.