Radar Signal Classification Using Integral Image Data Structures
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current radar systems for autonomous and semi-autonomous vehicles are limited in the number of detections they can process due to the high number of calculations required to classify received signals as noise or detections, leading to potential missed detections.
Innovation Solution
A system and method that uses an integral image data structure to efficiently classify received signals from a radar system by generating an initial image, computing an integral image, and determining energy sums within neighborhoods and guards to differentiate between noise and object detections, reducing the computational burden.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional signal classification methods are used to determine whether each cell contains noise or a detection, then detection accuracy can be maintained, but the number of computations required is excessively high, limiting the number of detections that can be processed
Solution Approach 1:
The patent segments the signal processing task by dividing the radar data into range bins and applying different processing strategies to different regions. The integral image technique segments the computational workload by pre-calculating cumulative sums, allowing rapid determination of energy distribution across multiple cells without processing each cell individually.
Solution Approach 2:
The patent performs preliminary actions by pre-calculating and storing integral images that contain cumulative energy information. This preliminary computation allows subsequent detection decisions to be made rapidly by simply querying pre-computed values rather than summing energy across multiple cells at detection time, significantly reducing real-time computational requirements.
2Measurement precision
If more computations are performed to accurately classify signals, then detection accuracy improves, but processing time increases and may lead to missed detections
Solution Approach 1:
The integral image data structure serves itself by containing all necessary cumulative energy information within its cells. Each cell in the integral image stores the sum of energies from previous cells, allowing the structure to provide its own classification information without requiring external computation, thus enabling rapid accurate classification.
3Reliability
If the radar system processes all received signals in detail, then detection accuracy is maximized, but the system becomes limited in the number of detections it can handle
Solution Approach 1:
The patent transitions from processing individual cells in isolation to a two-dimensional integral image space where cumulative energy information is organized across range and detection dimensions. This dimensional transformation allows the system to evaluate multiple potential detections simultaneously by examining patterns in the integral image, thereby maintaining reliability while increasing throughput.
Data Source
AI summary
A system for classifying received signals from a radar system into noise or a detection includes a source of a radar energy map and a memory that stores an integral image data structure for computing an integral image. The system includes a processor in communication with the source and the memory programmed to: generate an initial image including initial cells each having an energy value based on the radar energy map; compute the integral image based on the initial image; determine a coordinate location of an initial cell; determine coordinate locations of indices associated with corners of a neighborhood surrounding the initial cell; determine an energy sum of the neighborhood based on the indices and a value of respective cells from the integral image; determine an estimated noise associated with the initial cell based on the energy sum; and determine whether the initial cell indicates the detection of an object.