Road Friction Data Mapping for Environment Perception Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems lack high-quality training data for environment detection sensors to accurately estimate friction conditions on roadways, essential for both autonomous and non-autonomous vehicles, as current data is either unavailable or of low quality.
Innovation Solution
A method that associates environment perception data with data indicative of friction conditions by receiving and transforming position data and friction condition data from a second sensor, allowing for precise and reliable mapping and association, enabling the generation of high-quality training data for sensors like optical cameras and lidar units.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If environment perception sensors are trained with available training data, then the sensors can provide environment perception data, but the training data quality is low or unavailable leading to inaccurate friction condition detection
Solution Approach 1:
The patent uses a coordinate transformation system as an intermediary to bridge friction condition data from a second sensor (in vehicle coordinates) with environment perception data from a first sensor (in sensor coordinates). This intermediary transformation enables accurate association between the two data types, creating high-quality training data that improves both measurement precision and reliability
Solution Approach 2:
The patent creates a transformed copy of friction condition data by applying coordinate transformations to map data from the second sensor's coordinate system to the first sensor's coordinate system. This copied and transformed data can then be directly associated with environment perception data, providing high-quality training samples without requiring additional physical sensors
2Loss of information
If a second sensor is used to detect friction conditions, then friction data can be obtained, but the system complexity increases and data association with environment perception data becomes difficult
Solution Approach 1:
The patent makes the second sensor's data universal by transforming it into the first sensor's coordinate system. This allows friction condition data to be associated with any environment perception data from the first sensor, making the system more versatile and reducing the need for multiple dedicated sensors while maintaining information availability
Solution Approach 2:
The patent replaces complex mechanical data association methods with mathematical coordinate transformations. Instead of physically aligning sensors or using complex calibration mechanisms, the system uses coordinate geometry to automatically associate data from multiple sensors, reducing mechanical complexity while maintaining data integrity
3Manufacturing precision
If coordinate transformation is applied to associate friction data with environment perception data, then high-quality training data can be generated, but the processing complexity increases
Solution Approach 1:
The patent performs preliminary coordinate transformations on friction condition data before association with environment perception data. By pre-transforming the data into the correct coordinate system, the system ensures high association precision while simplifying the overall processing pipeline, as the transformation is done once rather than repeatedly during association
Data Source
AI summary
The disclosure relates to associating environment perception data of a first vehicle with data indicative of a friction condition. The environment perception data, position data, and data indicative of a friction condition can be received from a second sensor associated with the first vehicle. A first position can be transformed in a coordinate system centered on the first sensor using coordinate transformation based on the position data. Moreover, the first position can be mapped onto at least one corresponding element of the environment perception data, and the at least one element can be associated with the data indicative of a friction condition.