LIDAR Multipath Object Identification for Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current navigation systems using LIDAR data face challenges in accurately distinguishing between real objects and ghost objects, particularly those resulting from multipath reflections, which can lead to incorrect object detection and tracking, affecting navigation accuracy and safety.
Innovation Solution
A method and system for processing LIDAR measurement data that identifies potential aggressor objects, tracks their movement outside the LIDAR field of view, and characterizes them as multipath objects, allowing for the removal of ghost objects from the data set to improve navigation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If LIDAR data is used for navigation, then navigation coverage and real-time awareness are improved, but false detection of ghost objects occurs due to multipath reflections
Solution Approach 1:
The system performs preliminary identification of potential aggressor objects (reflective surfaces that cause multipath) and tracks their positions across multiple frames before they can generate ghost objects. This advance detection and tracking enables the system to preemptively mark and remove multipath artifacts when they appear in subsequent LIDAR scans, improving detection accuracy without requiring complex real-time processing during the actual navigation task.
2Measurement precision
If multipath objects are removed from LIDAR data, then navigation accuracy is improved, but computational complexity increases
Solution Approach 1:
By pre-identifying and tracking potential aggressor objects across multiple LIDAR frames, the system builds a database of multipath source positions before they generate ghost detections. This preliminary tracking reduces the computational burden during actual navigation, as the system only needs to check against the pre-computed aggressor list rather than analyzing all possible reflection paths in real-time, thus improving precision while limiting complexity increase.
Solution Approach 2:
The system extracts only the critical information needed for multipath identification - the positions and characteristics of potential aggressor objects - from the full LIDAR point cloud data. By separating and isolating only the relevant multipath source information for tracking and removal, the system avoids processing the entire complex scene in real-time, reducing computational complexity while maintaining high measurement precision for object location accuracy.
3Speed
If real-time object detection is performed, then navigation responsiveness is improved, but false positives from ghost objects occur
Solution Approach 1:
The system performs preliminary tracking of potential aggressor objects across multiple LIDAR frames to build a database of multipath sources before real-time navigation decisions are made. This advance preparation allows the system to maintain fast real-time detection by simply checking against the pre-computed aggressor list during navigation, rather than performing complex multipath analysis in real-time, thus preserving both speed and reliability.
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 enhances navigation by reducing false positives and negatives, improving the accuracy of object detection and tracking, thereby enhancing safety and efficiency in vehicle navigation.
Implementation Method 1
A lidar system includes a laser configured to emit light, where the emitted light is directed toward a region within the field of regard of the camera and a receiver configured to detect light returned from the emitted light
Implementation Method 2
characterizing one or more of the plurality of objects as one or more multi-path objects using tracked position of the one or more potential aggressor objects
Data Source
Figure 1A
Figure 2
Figure 3A~3D
AI summary
A method of processing of LIDAR measurement data including: receiving successive LIDAR 3D data sets over a time from a LIDAR system moving, during the time, through space, each LIDAR 3D data set corresponding to a measurement field of view (FOV) of the LIDAR system; identifying a plurality of objects in the LIDAR 3D data sets; designating at least one of the plurality of objects as at least one potential aggressor object; tracking position of the one or more potential aggressor objects relative to the LIDAR system as the one or more potential aggressor objects move outside of the measurement FOV of the LIDAR system; and characterizing one or more of the plurality of objects as one or more multi-path object using tracked position of the one or more potential aggressor objects.