Multi-FFT Radar Signal Detection with Dynamic Size Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing radar signal processing methods are inefficient for real-time detection due to their inability to adapt to the variety of signals received, leading to suboptimal detection of targets or objects.
Innovation Solution
A method involving multi-FFT processing with simultaneous computation of different sizes and automatic size selection, combined with real-time signal selection using a selection algorithm, and the use of a Chebyshev window for signal breakdown, allowing for the detection of radar signals across varying parameters like pulse duration and frequency modulation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a single fixed-size FFT computation is used, then the processing is simple and fast, but it cannot adapt to the large variety of radar signal parameters leading to poor detection efficiency
Solution Approach 1:
The patent implements dynamic FFT size selection by maintaining multiple FFT computation buffers with different sizes (e.g., 64, 128, 256, 512, 1024 points) and dynamically selecting the appropriate buffer size based on the detected signal characteristics. This allows the system to adapt to varying signal parameters while maintaining efficient real-time processing through pre-configured buffers rather than computing FFT sizes on-the-fly.
Solution Approach 2:
The system changes the parameter of FFT size by providing multiple pre-configured FFT buffers with different transformation sizes. The selection algorithm automatically chooses the appropriate FFT size parameter based on signal duration and frequency characteristics, enabling the system to handle diverse radar signals (short/long pulses, frequency modulated signals) without manual reconfiguration.
2Measurement precision
If multiple FFT sizes are computed simultaneously, then detection accuracy improves through better signal matching, but processing time and computational load increase
Solution Approach 1:
The patent performs preliminary action by pre-configuring multiple FFT buffers with different sizes before signal processing begins. These buffers are prepared in advance with appropriate window functions and parameters, so when a signal arrives, the selection algorithm can quickly match it to the most appropriate pre-prepared buffer without requiring complex real-time computation or buffer reconfiguration.
Solution Approach 2:
The system segments the FFT processing into multiple parallel computation paths, each with a dedicated buffer of a specific size. Instead of sequentially trying different FFT sizes, the patent divides the processing into concurrent segments that can operate simultaneously, with a selection mechanism that chooses the results from the most appropriate segment based on signal characteristics.
3Reliability
If signal detection uses fixed thresholds, then the processing is straightforward, but it fails to account for varying signal durations and frequencies leading to missed detections
Solution Approach 1:
The patent applies local quality by associating different detection thresholds with different FFT buffer sizes and signal characteristics. Instead of using a single global threshold, the system configures local thresholds specific to each FFT buffer size and signal type (e.g., shorter thresholds for brief pulses, longer thresholds for extended signals), allowing optimized detection for each local signal condition.
Solution Approach 2:
The selection algorithm incorporates feedback by continuously monitoring detected signal characteristics (duration, frequency content, energy level) and using this information to select both the appropriate FFT buffer size and the corresponding detection threshold. This feedback loop enables the system to adapt to varying signal conditions while maintaining reliable detection through criterion-based selection.
Data Source
AI summary
The present invention concerns a method for processing radar signals in a detection system comprising at least one computing architecture comprising at least one memory to store at least one set of signal-related programs and/or data, at least one processor executing said programs to implement said method comprising at least a first step (E1) to digitize radar signals, a second step to generate spectra from the digitized signals obtained at the first step (E1), via FFT computation, said method being characterized in that FFT computation at the spectra generation step (E2) comprises at least one multi-FFT processing comprising at least: simultaneous FFT computation (E20) of different sizes with automatic selection of said sizes and real-time selection of signals (E21) in all the spectra of the different FFT computations, via a selection algorithm.