PTAB

IPR2024-01465

Google LLC v. Cerence Operating Co

Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Voice Command Detection with Minimized Power Consumption
  • Brief Description: The ’750 patent relates to techniques for mobile devices to minimize power consumption by using a hierarchical approach to detect and process voice commands. The device applies low-power processing to received acoustic information first and engages higher-power processing only as needed to evaluate a potential voice command.

3. Grounds for Unpatentability

Ground 1: Obviousness over Basye and Shenhav - Claims 1-8, 11, 13-15, and 19 are obvious over Basye in view of Shenhav.

  • Prior Art Relied Upon: Basye (Application # 2014/0163978) and Shenhav (Application # 2014/0006825).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Basye teaches a device with distinct lower and higher power modes, where a low-power processor performs initial voice activity detection and keyword spotting (e.g., a "wakeword") before activating a higher-power application processor. Shenhav was argued to teach explicitly requesting that a remote server perform further recognition on a potential wakeword while the device’s main processor remains in a low-power standby mode. The combination allegedly teaches all limitations of the independent claims, including detecting a specific phrase in low-power mode, requesting server recognition, receiving an indication from the server that a voice command is present, and then selecting to transition to a higher power mode to respond.
    • Motivation to Combine: A POSITA would combine Shenhav with Basye to further Basye’s stated goal of reducing power consumption. Shenhav's method of keeping the main processor inactive until a remote server confirms a wakeword provides a clear power-saving benefit that improves upon Basye’s hierarchical system.
    • Expectation of Success: Petitioner asserted success was expected, as the combination involved applying well-known remote speech recognition techniques to battery-powered mobile devices to solve the known problem of high power consumption.

Ground 2: Obviousness over Basye, Shenhav, and Parker - Claims 9 and 16 are obvious over Basye in view of Shenhav and Parker.

  • Prior Art Relied Upon: Basye (Application # 2014/0163978), Shenhav (Application # 2014/0006825), and Parker (Application # 2015/0106089).
  • Core Argument for this Ground:
    • Prior Art Mapping: This ground builds on the Basye and Shenhav combination by adding Parker, which teaches providing an immediate indication to the user (e.g., an audio acknowledgement) to confirm that an acoustic command has been received. Petitioner argued this indication is provided while the device is still in low-power mode, prior to activating the second (higher-power) processor or receiving a full response from the server.
    • Motivation to Combine: A POSITA would combine Parker with the Basye/Shenhav system to improve user experience by addressing the known time lag and user uncertainty inherent in server-side speech processing. Providing immediate feedback assures the user their command was heard without waiting for the full, power-intensive processing cycle to complete.
    • Expectation of Success: The acknowledgement technique taught by Parker was described as compatible with any client-server speech recognition architecture, such as that proposed in the Basye/Shenhav combination.

Ground 3: Obviousness over Basye, Shenhav, and LeBeau - Claims 10 and 18 are obvious over Basye in view of Shenhav and LeBeau.

  • Prior Art Relied Upon: Basye (Application # 2014/0163978), Shenhav (Application # 2014/0006825), and LeBeau (Application # 2012/0034904).

  • Core Argument for this Ground:

    • Prior Art Mapping: This ground adds LeBeau to the primary combination to teach the evaluation of contextual cues—such as the device’s location, time of day, and battery status—during a processing stage.
    • Motivation to Combine: A POSITA would incorporate LeBeau’s context-aware functionality to make the power management of the Basye/Shenhav system more intelligent. Using cues like battery level to inform processing decisions is a logical and desirable improvement for a battery-powered mobile device focused on conserving energy.
    • Expectation of Success: Basye already considered some contextual cues (e.g., power level), making the addition of further well-understood cues from LeBeau a predictable enhancement.
  • Additional Grounds: Petitioner asserted an additional obviousness challenge (Claims 12, 17) based on the combination of Basye, Shenhav, and Conroy (Application # 2011/0314185). This ground argued it was obvious that certain tasks initiated by a voice command, such as playing audio as taught by Conroy, could be performed while the device selects to remain in a low-power mode.

4. Arguments Regarding Discretionary Denial

  • Petitioner argued against discretionary denial under both 35 U.S.C. §325(d) and the Fintiv factors.
  • Under §325(d): Petitioner asserted that the core prior art references and the proposed combinations were not substantively considered by the Examiner during prosecution, constituting a material error.
  • Under Fintiv: Petitioner argued the factors strongly weigh against denial because co-petitioner Google is not a party to the parallel district court litigation, the litigation is in a very early stage with minimal activity or investment, and the merits of the invalidity grounds presented in the petition are compelling.

5. Relief Requested

  • Petitioner requests institution of an inter partes review and cancellation of claims 1-19 of Patent 11,087,750 as unpatentable.