FMCW Radar Interference Suppression Using Convolutional Neural Network
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Radar sensors in automotive applications face interference issues due to overlapping radar signals from nearby vehicles, which can impair their operation and accuracy in detecting objects and maintaining safe distances.
Innovation Solution
A method using a convolutional neural network to filter out interfering signals from received radar signals by generating datasets from frequency-modulated chirps and processing them to reduce interference, enhancing the accuracy of radar systems in automotive applications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If radar sensors are deployed in multiple vehicles operating in close proximity, then the coverage and functionality of radar systems are improved, but interference signals between radar sensors increase
Solution Approach 1:
The patent applies neural network-based signal processing to convert the harmful interference signals from other radar sensors into usable information. The neural network is trained to recognize and filter out interference patterns while preserving valid radar echoes, effectively transforming the harmful multi-sensor interference into a benefit by enabling reliable operation in dense radar environments through intelligent signal discrimination
Solution Approach 2:
The patent changes the processing parameters of radar signals by applying neural network transformations to the received signal data. The neural network modifies signal characteristics such as amplitude, phase, and frequency components to distinguish between interference and valid targets, enabling the radar system to maintain performance despite the presence of multiple operating sensors
2Reliability
If neural network filtering is applied to reduce interfering signals, then the reliability of radar operation is improved, but the device complexity increases
Solution Approach 1:
The patent replaces traditional mechanical or algorithmic signal filtering methods with a neural network-based processing system. Instead of using fixed filter parameters or complex mathematical transformations, the system employs a trained neural network that automatically adapts to different interference scenarios, reducing the need for manual tuning and simplifying the overall processing architecture while improving reliability
Solution Approach 2:
The neural network is trained offline to automatically learn and adapt to various interference patterns and signal characteristics. Once trained, the network performs self-service by autonomously distinguishing between interference and valid radar returns without requiring real-time human intervention or complex adaptive algorithms, thereby improving reliability while keeping the operational complexity manageable
Data Source
AI summary
A method for a radar device is described below. According to an example implementation, the method comprises transmitting an RF transmission signal that comprises a plurality of frequency-modulated chirps, and receiving an RF radar signal and generating a dataset containing in each case a particular number of digital values based on the received RF radar signal. A dataset may in this case be associated with a chirp or a sequence of successive chirps. The method furthermore comprises filtering the dataset by way of a neural network to which the dataset is fed in order to reduce an interfering signal contained therein. A convolutional neural network is used as the neural network.