# US 12,601,780 B2 · Methods and systems for detecting compromised sensors using dynamic watermarking > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,601,780 B2 - Document ID: US12601780B2 - Status: Granted - Kind: B2 - Application number: 18/209,835 - Filed: 2023-06-14 - Published: 2023-12-14 - Granted: 2026-04-14 - Earliest claimed: 2022-06-14 - Assignees: The Texas A&M University System - Applicants: The Texas A&M University System - Inventors: Prasad N. Enjeti; Le Xie; Panganamala R. Kumar - Web page: https://patentgrep.com/patent/US12601780B2 ## CPC - G01R 31/2829: {Testing of circuits in sensor or actuator systems (testing of apparatus for measuring electric or magnetic variables G01R35/00; testing of indicating or recording apparatus G01D; in airbag systems B60R21/0173; checking gas analysers G01N33/007; monitoring or fail-safe circuits for electromagnets H01F7/1844)} - G01R: MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES (indicating correct tuning of resonant circuits H03J3/12) ## Abstract Systems and methods for detecting compromised sensors using dynamic watermarking. In some examples, a method includes injecting a dynamic random signal into an input of a power distribution system. The power distribution system includes at least one sensor and at least one power electronic controller configured to use the at least one sensor. The method includes monitoring a sensor signal from the at least one sensor. The method includes determining whether the at least one sensor is compromised based on a comparison between the dynamic random signal and the sensor signal. ## Claims ### Claim 1 A method comprising: injecting a dynamic random watermarking signal into an input of a power distribution system, the power distribution system comprising at least one sensor and at least one power electronic controller configured to use the at least one sensor; monitoring sensor signal from the at least one sensor; and determining whether the at least one sensor is compromised by sampling real time measurements from the sensor signal, sampling real time measurements from a system model that includes the dynamic random watermarking signal, computing a statistical variance between the real time measurements from the model that includes the dynamic random watermarking signal and the real time measurements from the sensor signal, detecting an increase in the statistical variance, and, in response to detecting the increase in the statistical variance, indicating the presence of a possible attack on the at least one sensor. ### Claim 2 (depends on 1) The method of claim 1 wherein the power distribution system is a grid-tied photovoltaic system and the at least one power electronic controller comprises a direct current/alternating current (DC-AC) inverter. ### Claim 3 (depends on 2) The method of claim 2 wherein injecting the dynamic random watermarking signal into the input comprises injecting the dynamic random watermarking signal into a direct current (DC) input of the inverter. ### Claim 4 (depends on 2) The method of claim 2 wherein the input is an electrical node between a photovoltaic power source and the inverter. ### Claim 5 (depends on 1) The method of claim 1 wherein injecting the dynamic random watermarking signal into the input comprises injecting the dynamic random watermarking signal into a control input of a power conversion device. ### Claim 6 (depends on 1) The method of claim 1 wherein the dynamic random watermarking signal has a Gaussian distribution with zero mean and an amplitude range sufficiently small such that the dynamic random watermarking signal propagates though the power distribution system without disturbing the performance of one or more power conversion stages. ### Claim 7 (depends on 1) The method of claim 1 wherein determining whether the at least one sensor is compromised comprises performing two or more variance tests. ### Claim 8 (depends on 6) The method of claim 6 wherein performing two or more variance tests comprises using moving average sampling. ### Claim 9 (depends on 1) The method of claim 1 wherein determining whether the at least one sensor is compromised comprises determining that a sequence of measurements from the sensor signal fails at least one variance test by an error amount exceeding a threshold error. ### Claim 10 (depends on 1) The method of claim 1 wherein determining whether the at least one sensor is compromised comprises detecting if the at least one sensor is malfunctioning and/or the at least one sensor is processing an incorrect signal. ### Claim 11 A system comprising: a watermarking device comprising: an input for receiving an electrical signal; a watermark injector circuit configured for injecting a dynamic random watermarking signal into the electrical signal; and an output for outputting the electrical signal with the dynamic random signal superimposed on the electrical signal to a power distribution system, the power distribution system comprising at least one sensor and at least one power electronic controller configured to use the at least one sensor; and a monitoring system, implemented on at least one processor, configured for monitoring a sensor signal from the at least one sensor and determining whether the at least one sensor is compromised by sampling real time measurements from the sensor signal, sampling real time measurements from a system model that includes the dynamic random watermarking signal, computing a statistical variance between the real time measurements from the model that includes the dynamic random watermarking signal and the real time measurements from the sensor signal, detecting an increase in the statistical variance, and, in response to detecting the increase in the statistical variance, indicating the presence of a possible attack on the at least one sensor. ### Claim 12 (depends on 11) The system of claim 11, wherein the power distribution system is a grid-tied photovoltaic system and the at least one power electronic controller comprises a direct current/alternating current (DC-AC) inverter. ### Claim 13 (depends on 12) The system of claim 12, wherein injecting the dynamic random signal into the input comprises injecting the dynamic random signal into a direct current (DC) input of the inverter. ### Claim 14 (depends on 12) The system of claim 12, wherein the input is an electrical node between a photovoltaic power source and the inverter. ### Claim 15 (depends on 11) The system of claim 11, wherein injecting the dynamic random watermarking signal into an input comprises injecting the dynamic random watermarking signal into a control input of a power conversion device. ### Claim 16 (depends on 11) The system of claim 11, wherein the dynamic random watermarking signal has a Gaussian distribution with zero mean and an amplitude range sufficiently small such that the dynamic random watermarking signal propagates though the power distribution system without disturbing the performance of one or more power conversion stages. ### Claim 17 (depends on 11) The system of claim 11, wherein determining whether the at least one sensor is compromised comprises performing two or more variance tests. ### Claim 18 (depends on 16) The system of claim 16, wherein performing two or more variance tests comprises using moving average sampling. ### Claim 19 (depends on 11) The system of claim 11, wherein determining whether the at least one sensor is compromised comprises determining that a sequence of measurements from the sensor signal fails at least one variance test by an error amount exceeding a threshold error. ### Claim 20 (depends on 11) The system of claim 11, wherein determining whether the at least one sensor is compromised comprises detecting if the at least one sensor is malfunctioning and/or the at least one sensor is processing an incorrect signal.