Vehicle Object Recognition Using Map-Based Ghost Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle systems often misrecognize static objects like buildings or street trees as dynamic objects due to sensor noise, leading to unnecessary deceleration and degraded ride quality during autonomous driving.
Innovation Solution
An apparatus and method that utilize a sensor module and processor to set a first region based on vehicle and map data, identifying road boundaries and removing noise corresponding to moving objects within that region, thereby preventing misrecognition of static objects as dynamic.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor fusion system is used to detect moving objects, then detection capability is improved, but false recognition of static objects as dynamic objects increases
Solution Approach 1:
The patent introduces map data as an intermediary to verify the authenticity of detected objects. The processor cross-references sensor-detected objects with map information (buildings, road boundaries, street trees) to determine whether detected objects are genuine moving objects or false ghosts, thereby resolving the contradiction between detection capability and false recognition rate
Solution Approach 2:
The system implements feedback by continuously comparing sensor detection results with map data and using this comparison to filter false positives. The processor uses map information as a reference framework to validate sensor readings, providing feedback that eliminates false ghost detections while preserving genuine moving object detection
2Reliability
If ghost objects are removed based on map data, then false recognition is reduced, but system complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-loading map data (buildings, road boundaries, street trees) into the system before operation. This pre-established reference framework allows the processor to quickly verify detected objects against known static locations, reducing false recognition without requiring complex real-time analysis of each detected object
Solution Approach 2:
The system changes the parameter of object verification from sensor-based detection alone to a composite verification method that incorporates map data parameters. By adding the map data reference layer, the system transforms the verification process from purely sensor-dependent to a multi-parameter comparison, improving reliability while maintaining manageable complexity through efficient data structures
3Ease of operation
If unnecessary deceleration is prevented by removing ghosts, then ride quality is improved, but detection accuracy requirements increase
Solution Approach 1:
The patent segments the detection space into zones based on map data (areas where buildings, road boundaries, or street trees are located). By dividing the environment into segmented regions with known static object locations, the system can efficiently determine whether detected objects are ghosts or real without requiring uniformly high detection accuracy across all areas, thus improving ride quality while managing detection requirements
Data Source
AI summary
An apparatus configured to prevent misrecognition of an object in a vehicle and a method therefor are provided. The apparatus may comprise a sensor module configured to collect vehicle data and at least one processor electrically connected to the sensor module. The at least one processor may be configured to set a first region through which a moving object is unable to pass, based on the vehicle data collected by means of the sensor module and map data and removes noise corresponding to at least one moving object, when the at least one moving object is detected in the first region. In addition, various embodiments recognized through the specification are possible.