Detecting, tracking, and transmitting to identified objects using models in a modular system

Patent metadata

Field	Value
Publication number	US-12455364-B2
Application number	US-202217978807-A
Country	US
Kind code	B2
Filing date	Nov 1, 2022
Priority date	May 20, 2022
Publication date	Oct 28, 2025
Grant date	Oct 28, 2025

Abstract

A modular, radio frequency (“RF”) system includes one or more directional antennas and is configured with both hardware and software components to enable the RF system to monitor (e.g., detect or track signals or objects) and/or interact with (e.g., track signals or objects, or transmit signals) objects in particular directions. The RF system includes one or more machine learning models to determine, based on received signals, one or more signals to transmit.

First claim

What is claimed is:

1. A computer-implemented method for applying a machine learning model to identify one or more RF signals, the computer-implemented method comprising, by one or more hardware processors executing program instructions:
receiving raw RF signal data through two or more directional antennas corresponding to a first RF system, wherein the two or more directional antennas are configured to be selectably activated, wherein the first RF system comprises a processing module, and wherein the processing module comprises a machine learning component;
sampling, by the processing module of the first RF system, the raw RF signal data to generate sampled RF signal data;
transmitting the sampled RF signal data to the machine learning component for input into a machine learning model;
receiving, from the machine learning component, an output from the machine learning model; and
based on RF signal attributes and the output, identifying a type of an object, wherein the RF signal attributes comprise bandwidth, channel, and signal rate that are each associated with the raw RF signal data.

2. The computer-implemented method of claim 1 , further comprising, by the one or more hardware processors executing program instructions:
selecting for activation one or both of the two or more directional antennas for receiving the raw RF signal data, wherein the two or more directional antennas comprise broad-bandwidth directional antennas.

3. The computer-implemented method of claim 2 , further comprising, by the one or more hardware processors executing program instructions:
determining, based on positioning of the two or more directional antennas, a location of the object.

4. The computer-implemented method of claim 1 , further comprising, by the one or more hardware processors executing program instructions:
filtering the raw RF signal data.

5. The computer-implemented method of claim 4 , wherein the filtering includes a complete or partial suppression of one or more aspects of the raw RF signal data, the sampled raw signal data, or a subset of RF signal data.

6. The computer-implemented method of claim 4 , wherein the filtering comprises removing one or more RF signals from the raw RF signal data that correspond to: (1) RF signals associated with friendly equipment, (2) RF signals associated with equipment that has been manually or automatically flagged as friendly, or (3) a preconfigured whitelist or blacklist.

7. The computer-implemented method of claim 1 , wherein the sampling of the raw RF signal data comprises sampling the raw RF signal data at a preconfigured timestep.

8. The computer-implemented method of claim 7 , wherein the preconfigured timestep is between Oms and 15 ms.

9. The computer-implemented method of claim 7 , wherein the preconfigured timestep is further based at least in part on hardware components associated with an RF system implementing the method.

10. The computer-implemented method of claim 1 , wherein the output includes predicted classes and probabilities corresponding to a subset of RF signals from the sampled RF signal data.

11. The computer-implemented method of claim 10 , wherein identification of the type of object is further based on bandwidth, channel, and signal rate that are each associated with the subset of RF signals.

12. The computer-implemented method of claim 1 , wherein an output of the machine learning model includes predicted classes and probabilities corresponding to one or more RF signals that are identified by the machine learning model.

13. The computer-implemented method of claim 1 further comprising training the machine learning model, wherein training the machine learning model comprises:
based at least in part on annotations corresponding to a signal of interest included in a spectrogram, generating a first subset of RF signal data corresponding to the signal of interest, wherein the spectrogram is generated based on raw RF signal training data; and
inputting, into the machine learning model, the first subset of RF signal data for training the machine learning model to identify the signal of interest.

14. The computer-implemented method of claim 13 , wherein the spectrogram comprises the raw RF signal training data as a function of frequency, time, and/or intensity.

15. The computer-implemented method of claim 13 , wherein the annotations further correspond to a period of time in which the signal of interest is present on the spectrogram.

16. The computer-implemented method of claim 13 , wherein the annotations further correspond to one or more frequencies or frequency bands associated with the signal of interest.

17. The computer-implemented method of claim 13 , wherein the generating of the first subset of RF signal data includes removal of at least a portion of the raw RF signal training data that is not part of the signal of interest.

18. A system comprising:
a computer readable storage medium having program instructions embodied therewith; and
one or more processors configured to execute the program instructions to cause the system to perform the computer-implemented method of claim 1 .

19. A computer program product comprising a non-transitory computer-readable storage medium having program instructions embodied therewith, the program instructions executable by one or more processors to cause the one or more processors to perform the computer-implemented method of claim 1 .

20. The computer-implemented method of claim 1 , wherein the raw RF signal data is sampled at a rate that is based on at least one of: (1) a temperature of the one or more hardware processors, or (2) a quantity of raw RF signal data.

Assignees

• Anduril Industries Inc

Inventors

• El-Akkad Sam
• Fischer Christopher
• Whitaker Travis
• Pearson Bryden
• Berk Todd
• Pham Thao
• Hsu Jon
• Dart Cameron

Classifications

• G01S13/0209 Primary

  Systems with very large relative bandwidth, i.e. larger than 10 %, e.g. baseband, pulse, carrier-free, ultrawideband · CPC title

• G01S13/524

  based upon the phase or frequency shift resulting from movement of objects, with reference to the transmitted signals, e.g. coherent MTi · CPC title

• G01S7/006

  using shared front-end circuitry, e.g. antennas (G01S13/765, G01S13/825 take precedence) · CPC title

• G01S13/765

  with exchange of information between interrogator and responder · CPC title

• G01S13/582 Primary

  adapted for simultaneous range and velocity measurements · CPC title