Sonos Inc v. Google LLC

1. Case Identification

• Case #: IPR2023-01288
• Patent #: 9,632,748
• Filed: August 8, 2023
• Petitioner(s): Sonos, Inc.
• Patent Owner(s): Google LLC
• Challenged Claims: 1-4, 7, 9-12, and 14-15

2. Patent Overview

• Title: Arbitration Among Multiple Voice-Enabled Devices
• Brief Description: The ’748 patent discloses techniques for a group of computing devices to manage voice commands. The system designates a single "best candidate" device to monitor for and act on spoken audio input containing a command, a process the patent refers to as "arbitration."

3. Grounds for Unpatentability

Ground 1: Anticipation of Claims 1-4, 7, 9-12, and 14-15 under 35 U.S.C. §102 over Rosenberger

• Prior Art Relied Upon: Rosenberger (Patent 8,340,975).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Rosenberger discloses every limitation of the challenged independent and dependent claims. Rosenberger teaches a system of "speech recognition control devices" that solves the "duplicate response" problem when multiple devices in proximity hear the same voice command. Petitioner mapped Rosenberger’s "coordinating controller" to the claimed "first computing device." This device receives a spoken audio input, determines which device should handle the command by comparing a locally calculated "weighted signal" (WS) with WS values received from other devices, and transmits a "stop" message to the losing devices. The winning device then proceeds to interact with the user, which includes performing speech recognition on subsequent utterances and executing an associated action. This process, Petitioner asserted, directly maps to the method steps of claims 1, 7, and 11. Petitioner further argued Rosenberger's disclosure of using context (e.g., a "wait timer" and microphone signal strength) in its WS calculation anticipates the dependent claims.

Ground 2: Obviousness of Claims 2-3, 9-10, and 12 under 35 U.S.C. §103 over Rosenberger in view of Jang

• Prior Art Relied Upon: Rosenberger (Patent 8,340,975) and Jang (Application # 2013/0073293).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground was presented as an alternative for dependent claims 2, 3, 9, 10, and 12, which add limitations requiring the arbitration decision to be based on specific device "contexts." Petitioner argued that to the extent Rosenberger does not explicitly disclose the specific contexts listed in claim 10 (e.g., remaining power, currently executing applications), Jang provides these missing elements. Jang addresses the same duplicate response problem as Rosenberger and explicitly teaches using various "voice recognition results" for arbitration, including remaining power and the relevancy of a currently executing application to a voice command.
  • Motivation to Combine: Petitioner contended a person of ordinary skill in the art (POSITA) would combine Rosenberger with Jang because both references are in the same field and address the identical problem of arbitrating among multiple voice-enabled devices. A POSITA reviewing Rosenberger's system, which relies on a generic "weighted signal," would be motivated to consult references like Jang to find specific, effective factors to incorporate into that signal. Jang expressly touts the benefits of using its contextual factors—such as selecting a device with more battery power or one already running a relevant application—to make arbitration more efficient.
  • Expectation of Success: A POSITA would have a reasonable expectation of success in this combination. It involved applying a known technique (using specific contextual data for arbitration from Jang) to a similar existing system (Rosenberger's) to achieve the predictable result of a more robust and efficient device selection process.

4. Key Claim Construction Positions

• For the purposes of the petition, Petitioner adopted the Patent Owner's broad constructions for key terms from a parallel litigation.
• A critical term, "a command to temporarily refrain from performing speech recognition," was construed broadly based on the Patent Owner's litigation positions. Petitioner argued that under this broad construction, a command could be satisfied by a device refraining for a period of "practically zero" seconds or by refraining from processing the current audio input while remaining ready to process the next one. This strategic adoption of a broad construction simplifies the mapping of prior art, such as Rosenberger's "stop" message, to the claim limitations.

5. Arguments Regarding Discretionary Denial

• §314(a) (Fintiv): Petitioner argued that discretionary denial under Fintiv would be inappropriate. A parallel district court case was stayed pending the outcome of an ITC Investigation, meaning minimal resources had been invested and there was no risk of duplicative efforts or conflicting decisions. Furthermore, Petitioner noted that under current PTab guidance, parallel ITC proceedings do not trigger Fintiv-based denials.
• §325(d): Petitioner asserted that denial under §325(d) would be improper because the petition raises new arguments based on prior art not previously considered by the USPTO. The primary references, Rosenberger and Jang, were argued to be materially different from the art of record during prosecution, as they directly address the core concept of an arbitration process among multiple voice-enabled devices to select a single respondent, a topic the prosecution history art allegedly failed to address.

6. Relief Requested

• Petitioner requests institution of an inter partes review (IPR) and cancellation of claims 1-4, 7, 9-12, and 14-15 of the ’748 patent as unpatentable.