Radar Object Detection via IQ Data and Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Radar systems are unreliable for identifying ground-based objects due to noise and ground clutter, limiting their effectiveness for aircraft landing and taxiing operations, especially in adverse weather conditions.
Innovation Solution
A radar-based method and apparatus that utilizes in-phase quadrature (IQ) data to generate range-doppler maps, combined with machine learning models like range-doppler bin estimation and azimuth angle regression neural networks, to accurately identify objects relative to a vehicle, reducing the impact of noise and ground clutter.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional radar systems are used to identify ground-based objects, then the system structure remains simple, but the detection reliability deteriorates due to noise and ground clutter
Solution Approach 1:
The patent segments the radar signal processing into distinct stages: range-doppler mapping, machine learning classification, and model generation. By dividing the complex detection task into manageable segments, each handled by specialized components (IQ data processor, range-doppler mapper, machine learning model), the system achieves high detection reliability while maintaining manageable processing complexity through modular architecture.
Solution Approach 2:
The patent introduces machine learning models as intermediary components between raw radar data and final object identification. These models act as mediators that process range-doppler maps and generate accurate object models, bridging the gap between simple radar hardware and complex detection requirements, thereby improving reliability without requiring entirely new hardware systems.
2Measurement precision
If radar systems are used to identify ground-based objects, then the system remains relatively simple, but the measurement precision deteriorates due to ground clutter noise
Solution Approach 1:
The patent converts the harmful ground clutter noise into beneficial information by using machine learning models to distinguish between clutter and actual objects. The noise patterns in the range-doppler maps are analyzed and used to train models that can reliably differentiate between ground clutter and real objects, transforming the previously harmful noise factor into a training resource for improved detection precision.
Solution Approach 2:
The patent changes the parameter representation from raw radar signals to range-doppler maps, which transform the data into a different parameter space where objects and clutter have distinct characteristics. This parameter transformation enables machine learning models to more effectively separate objects from ground clutter, improving measurement precision by working in a transformed parameter domain rather than the original signal domain.
3Measurement precision
If visual identification techniques are used to identify runway, then the accuracy improves, but the system becomes limited by weather conditions such as fog
Solution Approach 1:
The patent replaces visual identification mechanisms with radar-based electromagnetic detection. Instead of relying on optical systems that are blocked by fog and adverse weather, the system uses radar waves that can penetrate through these conditions. The machine learning-enhanced radar system provides runway identification capability that is independent of weather conditions, substituting the mechanical/optical vision system with an electromagnetic sensing system that has superior weather adaptability.
4Adaptability or versatility
If LIDAR systems are used to identify ground-based objects, then the system is not limited by weather conditions, but the range and resolution are predefined and limited
Solution Approach 1:
The patent creates a universal radar-based detection system that can perform multiple functions: identifying ground-based objects, detecting airborne objects, and providing weather-independent operation. By making the radar system multi-functional through machine learning enhancement, it can adapt to various detection scenarios and object types without requiring separate specialized systems for each function, thereby achieving weather independence with flexible detection capabilities.
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 reliable identification of ground-based objects and aircraft landing surfaces, even in adverse weather, by processing radar signals to improve detection accuracy and confidence, facilitating safe landing and taxiing operations.
Implementation Method 1
Radar is utilized for a variety of purposes including the identification of objects in a scene
Implementation Method 2
converting the IQ data to one or more initial range-doppler maps
Implementation Method 3
evaluating the one or more initial range-doppler maps with a machine learning model to generate the model
Data Source
AI summary
A method, apparatus and computer program product are provided to generate a model of one or more objects relative to a vehicle. In the context of a method, radar information is received in the form of in-phase quadrature (IQ) data and the IQ data is converted to one or more first range-doppler maps. The method further includes evaluating the one or more first range-doppler maps with a machine learning model to generate the model that captures the detection of the one or more objects relative to the vehicle. A corresponding apparatus and computer program product are also provided.