LiDAR Object Box Merging for Stable Vehicle Path Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle control systems using LiDAR sensors often incorrectly merge virtual boxes, leading to inaccurate identification of external objects, which can result in unexpected changes in the driving route of a host vehicle.
Innovation Solution
A vehicle control apparatus and method that determines whether to merge virtual boxes generated by LiDAR sensors based on distance and object separation from road boundaries, allowing for accurate identification of pedestrians and reducing processor load by optimizing merge operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If virtual boxes are merged based on LiDAR data, then object detection coverage is improved, but object identification accuracy deteriorates due to incorrect merging
Solution Approach 1:
The patent segments the merging process into multiple stages: initial virtual box generation from LiDAR data, followed by selective merging based on specific criteria (distance thresholds, object type matching, spatial relationships). This segmentation allows the system to maintain detection coverage while improving identification accuracy by only merging boxes that meet predefined criteria, preventing incorrect merges between unrelated objects.
2Reliability
If merging operations are performed on all virtual boxes, then detection completeness is improved, but processor load increases
Solution Approach 1:
The patent applies preliminary filtering actions before the main merging operation. Virtual boxes are pre-screened based on distance thresholds, object type compatibility, and spatial relationship criteria. Only boxes that pass these preliminary checks are candidates for merging, significantly reducing the number of combinations that require intensive processing while maintaining detection completeness.
Solution Approach 2:
The patent implements local quality control by applying different merging criteria to different regions and object types. For example, stricter merging criteria are applied to pedestrians near road boundaries, while different criteria apply to vehicles in open areas. This localized approach optimizes processor usage by avoiding unnecessary merging evaluations while ensuring accurate detection in critical zones.
3Productivity
If merging criteria are relaxed to reduce processing complexity, then processor load is reduced, but merging accuracy deteriorates
Solution Approach 1:
The patent uses parameter changes to balance processing efficiency and merging accuracy. Multiple distance thresholds (e.g., first threshold value, second threshold value) are defined to create hierarchical merging criteria. Object type parameters are also used to modulate merging behavior - for instance, pedestrians have different merging parameters than vehicles. These parameter changes enable the system to maintain high merging accuracy while avoiding computationally expensive operations on all possible box combinations.
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
The solution effectively reduces the risk of incorrect object identification, improves driving route stability, and decreases processor load by implementing intelligent merge and unmerge operations of virtual boxes.
Implementation Method 1
the external object may be identified by using a LiDAR
Data Source
AI summary
The present disclosure may relate to a vehicle control apparatus and a method thereof. A vehicle control apparatus may include a sensor and a processor. The processor may convert a first virtual box, associated with a first object and obtained via the sensor at a first time, into a second virtual box associated with a second time that, determine a plurality of third virtual boxes, associated with the second virtual box, based on a plurality of data points obtained at the second time via the sensor, merge at least part of the plurality of third virtual boxes into a merge box, maintain or cancel the merging of the at least part of the plurality of third virtual boxes, and output a signal indicating a result of the maintaining or canceling of the merging.