PTAB

IPR2025-00194

Digital Global Systems Inc v. DeepSig Inc

Log In
Key Events
Petition
petition Intelligence

1. Case Identification

2. Patent Overview

  • Title: Techniques for Mitigating Signal Distortion
  • Brief Description: The ’540 patent discloses techniques for mitigating signal distortion in wireless communications by using a nonlinear pre-distortion machine learning model (NPDMLM), such as a neural network, to generate a pre-distorted signal that counteracts distortion introduced by transmitter components like power amplifiers.

3. Grounds for Unpatentability

Ground 1: Obviousness over Jüschke - Claims 1-3 and 7-10 are obvious over Jüschke.

  • Prior Art Relied Upon: Jüschke (EP 2,538,553 A1).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Jüschke taught every element of the challenged claims. Jüschke disclosed a "more flexible radio transmitter" that implements a "pre-distortion neural network processor" to mitigate transmitter impairments. This processor, a type of nonlinear machine learning model, receives an input digital baseband signal, processes it using a neural network (NN) with nonlinear functions to correct for distortion, generates a pre-distorted signal, and sends it to an analog transmitter for transmission to a receiver. Petitioner contended that Jüschke's disclosure of implementing its flexible transmitter in various wireless systems, including base stations, rendered obvious the claim limitation of obtaining the signal at a "wireless device." The NN's weight and bias parameters, optimized for specific deployment scenarios like different cellular standards (GSM, LTE), were argued to meet the limitations regarding model parameters for different deployment scenarios.
    • Motivation to Combine (for §103 grounds): Not applicable (single reference ground).
    • Expectation of Success (for §103 grounds): Not applicable.

Ground 2: Obviousness over Jüschke in view of Holt - Claims 1-10, 12-14, and 16-22 are obvious over Jüschke in view of Holt.

  • Prior Art Relied Upon: Jüschke (EP 2,538,553 A1) and Holt (Patent 10,552,738).
  • Core Argument for this Ground:
    • Prior Art Mapping: This ground asserted that Jüschke taught most claim limitations, and Holt supplied any missing elements, particularly those related to model training. Holt disclosed techniques for adaptively training machine learning models (including NNs) in communication systems using a "loss function" that describes the difference between the initial input data and the final output data received from a decoder model after transmission. This addresses claim limitations requiring the determination of a "distance metric" between the transmit signal and a received radio signal to update model parameters. Petitioner argued that Jüschke's training method, which uses a feedback loop from the transmitter's output, would be improved by incorporating Holt's method of using the signal as actually received by a remote receiver. This combination was argued to render obvious claims requiring updating model parameters based on a distance metric computed between the transmitted and received signals (e.g., claims 4 and 12). Holt also taught that its encoder/decoder models function like an autoencoder, rendering claim 16 obvious.
    • Motivation to Combine (for §103 grounds): A POSITA would combine Jüschke and Holt to improve the accuracy and robustness of Jüschke's NN training. Using Holt's technique of training based on signals actually recovered by a receiver would provide more accurate training data that accounts for real-world transmission and channel effects, which is a known method for improving similar systems.
    • Expectation of Success (for §103 grounds): A POSITA would have a reasonable expectation of success because combining the references involved the application of a known training technique (Holt's loss function) to improve a known system (Jüschke's pre-distorter), both of which operate in the same technical field of wireless communications using NNs.

Ground 3: Obviousness over Jüschke in view of Holt and Dzierwa - Claims 2-6 are obvious over Jüschke in view of Holt and Dzierwa.

  • Prior Art Relied Upon: Jüschke (EP 2,538,553 A1), Holt (Patent 10,552,738), and Dzierwa (Patent 10,122,479).
  • Core Argument for this Ground:
    • Prior Art Mapping: This ground built upon Ground 2, adding Dzierwa to address claim 2's limitation of generating a pre-distorted signal "for the particular deployment scenario." Dzierwa taught using machine learning to "observe and learn an RF environment" and generate a "knowledge map" based on learning data. It further disclosed an "optimization module" that uses environmental parameters (e.g., noise, hardware parameters) to optimize a signal of interest. Petitioner argued that Dzierwa's teachings on learning and adapting to specific RF environmental conditions would have rendered it obvious to modify the Jüschke/Holt system to generate a pre-distorted signal specifically tailored to the particular, learned deployment scenario.
    • Motivation to Combine (for §103 grounds): A POSITA would combine Dzierwa with the teachings of Jüschke and Holt to predictably improve the performance of the pre-distortion system. Incorporating Dzierwa’s environmental learning and optimization would allow Jüschke’s pre-distorter to adapt its parameters not just to the transmitter hardware, but also to the specific, dynamic RF environment, leading to more accurate distortion correction.
    • Expectation of Success (for §103 grounds): The combination was presented as a predictable integration of complementary technologies. A POSITA would expect success in applying Dzierwa’s environmental optimization techniques to Jüschke’s pre-distortion model, as this represents combining familiar elements to achieve a known improvement in signal processing performance.

4. Arguments Regarding Discretionary Denial

  • Petitioner argued the Board should not exercise discretionary denial under Fintiv.
  • The petition asserted that Factors 1-5 of the Fintiv analysis favor institution because the ’540 patent was not subject to any co-pending district court litigation.
  • Regarding Factor 6, Petitioner contended it presented strong, meritorious grounds for invalidity.
  • The petition also argued that under the Advanced Bionics factors, institution is favored because none of the asserted prior art references were previously applied during prosecution, and they are not cumulative to the art of record.

5. Relief Requested

  • Petitioner requests institution of an inter partes review and cancellation of claims 1-10, 12-14, and 16-22 of the ’540 patent as unpatentable.