# US 12,216,229 B2 · Parametric CNN for radar processing

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 12,216,229 B2
- Document ID: US12216229B2
- Status: Granted
- Kind: B2
- Application number: 18/239,521
- Filed: 2023-08-29
- Published: 2023-12-21
- Granted: 2025-02-04
- Earliest claimed: 2020-06-18
- Assignees: Infineon Technologies AG
- Applicants: Infineon Technologies AG
- Inventors: Avik Santra; Thomas Reinhold Stadelmayer
- Web page: https://patentgrep.com/patent/US12216229B2

## CPC

- G01S 7/417: {involving the use of neural networks}
- G01S 7/412: {based on a comparison between measured values and known or stored values}
- G01S 7/415: {Identification of targets based on measurements of movement associated with the target}
- G01S 13/89: for mapping or imaging
- G01S 7/285: Receivers
- G01S 7/352: {Receivers}
- G01S 13/06: Systems determining position data of a target
- G01S 13/32: using transmission of continuous waves, whether amplitude-, frequency-, or phase-modulated, or unmodulated
- G01S 13/34: using transmission of continuous, frequency-modulated waves while heterodyning the received signal, or a signal derived therefrom, with a locally-generated signal related to the contemporaneously transmitted signal
- G01S 2013/0245: {Radar with phased array antenna}
- G06N 3/045: Combinations of networks
- G06N 3/08: Learning methods
- G01S: RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

In an embodiment, a method includes: transmitting a plurality of radar signals using a millimeter-wave radar sensor towards a target; receiving a plurality of reflected radar signals that correspond to the plurality of transmitted radar signals using the millimeter-wave radar; mixing a replica of the plurality of transmitted radar signals with the plurality of received reflected radar signals to generate an intermediate frequency signal; generating raw digital data based on the intermediate frequency signal using an analog-to-digital converter; processing the raw digital data using a constrained L dimensional convolutional layer of a neural network to generate intermediate digital data, where L is a positive integer greater than or equal to 2, and where the neural network includes a plurality of additional layers; and processing the intermediate digital data using the plurality of additional layers to generate information about the target.

## Claims

### Claim 1

A method comprising:
transmitting a plurality of sensor signals towards a target;
receiving a plurality of reflected sensor signals that correspond to the plurality of transmitted sensor signals;
producing raw digital data representing the reflected sensor signals;
processing the raw digital data using a constrained L dimensional convolutional layer of a neural network to generate intermediate digital data, wherein L is a positive integer greater than or equal to 2, and wherein the neural network comprises a plurality of additional layers, and wherein processing the raw digital data comprises using a time-domain two-dimensional (2D) filter to filter the raw digital data in a fast time dimension and in a slow time dimension; and
processing the intermediate digital data using the plurality of additional layers to generate information about the target.

### Claim 2 (depends on 1)

The method of claim 1, wherein generating information about the target comprises classifying the target based on a set of classes.

### Claim 3 (depends on 1)

The method of claim 1, wherein generating information about the target comprises providing a location of the target using a range-Doppler image (RAI), a range-angle image (RAI) or a Doppler-angle image (DAI).

### Claim 4 (depends on 1)

The method of claim 1, wherein processing the raw digital data using the constrained L dimensional convolutional layer comprises processing the raw digital data using a 2D sinc filter of the constrained L dimensional convolutional layer.

### Claim 5 (depends on 1)

The method of claim 1, wherein processing the raw digital data using the constrained L dimensional convolutional layer comprises processing the raw digital data using a 2D Morlet wavelet filter of the constrained L dimensional convolutional layer.

### Claim 6 (depends on 1)

The method of claim 1, further comprising training the neural network by:
initializing the neural network; and
after initializing the neural network, feeding training data to the constrained L dimensional convolutional layer, wherein trainable weights of the constrained L dimensional convolutional layer comprise at least one of a slow-time cutoff frequency, a slow-time bandwidth, a fast-time cutoff frequency, and a fast-time bandwidth.

### Claim 7 (depends on 1)

The method of claim 1, further comprising training the neural network by:
initializing the neural network; and
after initializing the neural network, feeding a training data to the constrained L dimensional convolutional layer, wherein trainable weights of the constrained L dimensional convolutional layer comprise a center frequency and a standard deviation for each dimension of the constrained L dimensional convolutional layer.

### Claim 8 (depends on 1)

The method of claim 1, wherein the plurality of additional layers comprises a first maxpool layer followed by an unconstrained 2D convolutional layer followed by a second maxpool layer, followed by a dense layer, and followed by a softmax layer.

### Claim 9 (depends on 1)

The method of claim 1, wherein the target is a human target.

### Claim 10 (depends on 9)

The method of claim 9, wherein generating information about the target comprises classifying the target based on a set of classes, and wherein the set of classes comprises classes indicative of human activities.

### Claim 11 (depends on 1)

The method of claim 1, further comprising tracking the target based on the generated information about the target.

### Claim 12 (depends on 1)

The method of claim 1, wherein the plurality of sensor signals are a plurality of chirps.

### Claim 13 (depends on 1)

The method of claim 1, wherein the plurality of additional layers comprises a first additional constrained convolutional layer.

### Claim 14 (depends on 1)

The method of claim 1, wherein generating information about the target comprises generating a range-Doppler image indicative of a location of the target, the method further comprising processing the intermediate digital data using a second plurality of additional layers to generate a range-angle image indicative of the location of the target.

### Claim 15

A system comprising:
a sensor comprising:
a transmitter configured to transmit a plurality of sensor signals towards a target;
a receiver configured to receive a plurality of reflected sensor signals; and
an artificial intelligence (AI) accelerator having an input coupled to an output of the sensor, and configured to:
receive raw digital data representing the reflected sensor signal, and
process the raw digital data using a constrained L dimensional convolutional layer of a neural network to generate intermediate digital data, wherein L is a positive integer greater than or equal to 2, and wherein the neural network comprises a plurality of additional layers, and wherein the AI accelerator is configured to process the raw digital data using a time-domain two-dimensional filter to filter the raw digital data in a fast time dimension and in a slow time dimension; and
process the intermediate digital data using the plurality of additional layers to generate, at an output of the AI accelerator, data associated with the target.

### Claim 16 (depends on 15)

The system of claim 15, further comprising a digital signal processor (DSP) having an input coupled to the output of the AI accelerator, wherein the AI accelerator is directly connected to the output of the sensor.

### Claim 17 (depends on 16)

The system of claim 16, wherein the DSP is configured to track the target based on the output of the AI accelerator.

### Claim 18 (depends on 16)

The system of claim 16, wherein the AI accelerator and the DSP are integrated in a same integrated circuit.

### Claim 19

A system comprising:
a sensor configured to transmit a plurality of chirps towards a target, receive a plurality of reflected chirps; and
a processor having an input coupled to an output of the sensor, and configured to:
receive raw digital data representing the received reflected chirps, and
process the raw digital data using a neural network having a first constrained two dimensional convolutional layer followed by a plurality of additional layers to generate, at an output of the plurality of additional layers, data associated with the target, wherein the processor is configured to process the raw digital data using a time-domain two-dimensional filter to filter the raw digital data in a fast time dimension and in a slow time dimension.

### Claim 20 (depends on 19)

The system of claim 19, wherein the processor is an artificial intelligence (AI) accelerator.