RADAR Sensor Processing Using External Sensor Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In autonomous vehicles, RADAR systems face challenges with processing time-consuming data, leading to stale detections in time-sensitive applications due to multipath and multibounce effects, which can result in false detections and inefficient processing.
Innovation Solution
Implementing a system that uses data from external sensors like imaging, LIDAR, or ultrasonic sensors to predict the future position of objects, allowing RADAR sensors to focus processing on prioritized locations, thereby enhancing processing efficiency by using machine learning models to integrate and filter data from these external sources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If RADAR systems process all data from the entire field of view, then detection coverage is comprehensive, but processing time increases and detections become stale
Solution Approach 1:
The system performs preliminary actions by using external sensor data to predict future object positions before RADAR data collection. This allows the RADAR system to pre-identify regions of interest and prioritize data collection in those areas, rather than processing the entire field of view uniformly. The prediction step prepares the system in advance, enabling faster targeted processing.
Solution Approach 2:
The system applies local quality by differentiating processing priorities across different spatial regions. Instead of uniform processing, it identifies specific subspaces (regions of interest) where objects are predicted to be located and allocates enhanced processing resources to those local areas. This selective approach maintains detection accuracy in critical regions while reducing overall processing time.
2Reliability
If RADAR systems collect data from the entire field of view, then detection coverage is complete, but processing efficiency decreases
Solution Approach 1:
External sensors perform preliminary detection and prediction of object positions, allowing the RADAR system to pre-identify which regions require detailed scanning. This preliminary action filters out areas unlikely to contain objects, enabling the RADAR to focus computational resources efficiently on promising regions while maintaining overall detection completeness.
Solution Approach 2:
The system segments the field of view into multiple subspaces based on predicted object positions. Instead of processing the entire field uniformly, it divides the detection space into priority regions (where objects are predicted) and non-priority regions, applying different processing intensities to each segment. This segmentation maintains detection completeness while improving processing efficiency through selective focus.
3Measurement precision
If RADAR systems process multipath and multibounce effects, then detection accuracy improves, but processing complexity increases
Solution Approach 1:
The system performs preliminary filtering using external sensor data to identify true object positions before RADAR processing. By predicting object locations in advance, it can pre-distinguish between genuine targets and multipath/multibounce artifacts, reducing the complexity of subsequent processing while maintaining detection accuracy through focused analysis of predicted regions.
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
This approach enables faster and more accurate RADAR detections by prioritizing data collection based on predicted object positions, reducing processing time and improving the reliability of perception tasks in autonomous vehicles.
Implementation Method 1
a RADAR sensor can be used to identify objects and measure range (e.g., distance) to an object as well as to determine movement of an object
Implementation Method 2
a light detection and ranging (LiDAR) sensor can be used to determine ranges (variable distance) of one or more targets by directing a laser to a surface of an entity and measuring the time for light reflected from the surface to return to the LiDAR
Implementation Method 3
measuring the time for light reflected from the surface to return to the LiDAR
Data Source
AI summary
Systems and techniques are provided for performing RADAR sensing based on external sensor feedback. An example method includes receiving position data associated with at least one object, wherein the position data is based on sensor data obtained by at least one sensor; identifying, based on the position data, at least one subspace from a field of view of a radar sensor, and collecting radar sensing data from the at least one subspace, wherein the radar sensing data corresponds to the at least one object.