Roku Inc v. Universal Electronics Inc

1. Case Identification

• Case #: IPR2022-01289
• Patent #: 9,847,083
• Filed: July 20, 2022
• Petitioner(s): Roku, Inc.
• Patent Owner(s): Universal Electronics Inc.
• Challenged Claims: 1-19

2. Patent Overview

• Title: Voice Command Augmented Codeset Record
• Brief Description: The ’083 patent describes a system for configuring a controlling device (e.g., a universal remote) to command an appliance using voice. The core invention is a "codeset record" or data structure that contains not only command data and transmission protocols but also phonetic representations of function names and their synonyms.

3. Grounds for Unpatentability

Ground 1: Claims 1-6, 8-10, and 12-18 are obvious over [Aginsky](https://ai-lab.exparte.com/case/ptab/IPR2022-01289/doc/1004) in view of [Stammler](https://ai-lab.exparte.com/case/ptab/IPR2022-01289/doc/1005).

• Prior Art Relied Upon: Aginsky (Patent 8,995,981) and Stammler (Patent 6,839,670).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Aginsky disclosed the foundational system of the ’083 patent: a controllable device (e.g., TV) that transfers a data structure to a remote device to enable control. Aginsky’s data structure, shown in tables, included individual command function entries (rows) containing function names, command signals, and configuration data. Petitioner asserted that Aginsky’s "SIGNAL" columns, which could be a coded IR signal or Bluetooth signal, inherently disclosed the claimed "protocol and formatting information" and contained the underlying "command data." While Aginsky broadly taught that voice commands could be used, it did not detail their implementation. Stammler allegedly supplied the missing elements, teaching a speech dialog system that uses "phonetic word subunits" (e.g., phonemes) and "synonymous words" to create an extensive and user-friendly vocabulary for voice control.
  • Motivation to Combine: Petitioner contended that a person of ordinary skill in the art (POSITA) would combine Stammler's teachings with Aginsky's system to implement the voice control functionality that Aginsky suggested but left undeveloped. The motivation was to improve Aginsky's flexible universal remote programming method by incorporating Stammler's established techniques for robust voice recognition, which predictably increased user satisfaction by allowing for a wider range of voice commands.
  • Expectation of Success: A POSITA would have had a reasonable expectation of success because Aginsky’s data structure was designed to be flexible and extensible. Applying Stammler's known methods for representing voice commands (phonetics and synonyms) to this data structure was a predictable integration of known technologies to achieve an improved result.

Ground 2: Claims 1-6, 8-10, and 12-18 are obvious over Aginsky in view of [Hanson](https://ai-lab.exparte.com/case/ptab/IPR2022-01289/doc/1006).

• Prior Art Relied Upon: Aginsky (Patent 8,995,981) and Hanson (Application # US 2012/0271639).

• Core Argument for this Ground:

  • Prior Art Mapping: As in Ground 1, Petitioner relied on Aginsky for the basic system architecture and data structure. Hanson was presented as teaching a system for "automated speech command discovery" that used a "Command Mapping" data set. Petitioner argued this data set was analogous to the claimed codeset record, as it explicitly linked function identifiers to commands, synonyms, and alternative commands ("Alts"). Hanson further taught processing speech commands using a speech engine with phonetic algorithms. This combination, Petitioner argued, disclosed a codeset record containing individual command function entries, each with command data, a standard function identifier, a phonetic representation, and synonyms.
  • Motivation to Combine: A POSITA would have been motivated to integrate Hanson's teachings into Aginsky's system to provide a more user-friendly and effective voice control interface. Hanson explicitly addressed the problem of users employing synonymous terms to execute commands and provided a clear data structure to solve it. Applying this solution to Aginsky's system would be a direct application of a known technique to improve a known system.
  • Expectation of Success: Petitioner asserted a high expectation of success, as combining Hanson's detailed command mapping structure with Aginsky’s flexible data structure framework represented a straightforward application of a known voice-control solution to a system designed to be customizable.

• Additional Grounds: Petitioner asserted that claims 7, 11, and 19 were obvious over the combinations in Ground 1 and Ground 2 with the further addition of Dresti (Application # US 2003/0103088). Dresti was cited for its teaching of using location data to select region-specific settings, such as language, for a universal remote. Petitioner also asserted a fourth ground that claims 7, 11, and 19 were obvious over Aginsky, Hanson, and Dresti for the same reasons.

4. Arguments Regarding Discretionary Denial

• Petitioner argued that discretionary denial under 35 U.S.C. §325(d) would be inappropriate. For Grounds 1 and 2, the primary references (Aginsky, Stammler, and Dresti) were never before the Examiner during the prosecution of the ’083 patent, meaning the art and arguments were new.
• For Grounds 3 and 4, while the Examiner did consider Hanson, Petitioner contended that the Examiner made a material error by adopting the patent applicant's mischaracterization of Hanson's teachings. The Examiner was allegedly persuaded that Hanson’s command mapping was only for "discovering/learning" a command, not for executing it. Petitioner argued that Hanson clearly teaches using its command mapping to automatically trigger the execution of functions, and the Examiner's failure to appreciate this was a critical error that led to allowance.

5. Relief Requested

• Petitioner requested the institution of an inter partes review and the cancellation of claims 1-19 of the ’083 patent as unpatentable.