Radar Ego-Velocity Estimation Using Odometric Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing radar systems face challenges in accurately and reliably classifying stationary and dynamic targets, particularly due to uncertainties in ego-velocity estimation and the complexity of algorithms, which limits their integration in low-performance hardware and affects the precision and robustness of driver assistance functionalities like Adaptive Cruise Control.
Innovation Solution
A method for operating a radar system that predicts ego-velocity and its variance, classifies targets using this information, and selects between two estimation methods based on the classification result, allowing for precise and robust classification of targets even with inaccurate ego-velocity information, and corrects odometric velocity using radar-based estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If odometric velocity estimation is used based on wheel encoders, then the velocity estimation is simple to implement, but the estimation accuracy deteriorates due to tire slip and unknown tire diameter
Solution Approach 1:
The patent introduces radar-based velocity estimation as an intermediary method to complement odometric estimation. The system uses radar to detect stationary targets and calculate velocity independently, then fuses this with odometric data to correct errors caused by tire slip and diameter uncertainties, thereby improving overall accuracy while maintaining implementation simplicity.
Solution Approach 2:
The patent changes the measurement parameters by switching from purely mechanical wheel encoder-based velocity estimation to a hybrid approach that incorporates radar-based velocity measurement. This parameter change allows the system to overcome the limitations of odometric methods by using electromagnetic wave-based velocity detection that is not affected by tire conditions.
2Reliability
If radar system is used for detecting stationary targets to estimate ego velocity, then the velocity estimation reliability is improved, but the system complexity and algorithm cost increase
Solution Approach 1:
The patent applies partial action by using radar velocity estimation only when necessary (when odometric estimation is unreliable or needs correction). The system selectively activates the more complex radar-based method based on predefined conditions, such as when stationary targets are detected or when velocity correction is needed, rather than continuously using the full radar classification algorithm.
Solution Approach 2:
The system performs self-service by automatically selecting between odometric and radar-based velocity estimation methods based on current operating conditions. The controller autonomously determines which method to use without requiring external intervention, switching between simple and complex methods based on the reliability of available data.
3Measurement precision
If complex classification algorithms are used to classify targets as stationary or dynamic, then the classification precision is improved, but the algorithm cost and hardware requirements increase
Solution Approach 1:
The patent segments the velocity estimation process into two distinct paths: a simple odometric estimation path for normal operation and a radar-based estimation path for correction or when higher accuracy is needed. This segmentation allows the system to maintain simple algorithms for most operations while having the capability to switch to more precise methods when necessary, avoiding the need to always run complex classification algorithms.
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
Enables precise and robust classification of stationary and dynamic targets, maintaining high classification quality even with inaccurate ego-velocity information, and provides reliable ego-velocity estimation for driver assistance systems.
Implementation Method 1
the radar system has at least one radar sensor for detecting at least one target outside the vehicle
Data Source
AI summary
A method is provided for operating a radar system of a vehicle. The radar system has at least one radar sensor for detecting at least one target outside the vehicle. A prediction of an ego-velocity (vEgo) of the vehicle is performed, so that a prediction result is determined. A classification for classifying the at least one detected target as a stationary target is then performed using the prediction result, so that a classification result is determined. One of at least two estimation methods is then selected for an estimation of the ego-velocity (vEgo), such that the selection is dependent on an evaluation of the classification result.