PTAB

IPR2018-01259

Infinera Corporation v. Core Optical Technologies, LLC

Title: Cross Polarization Interface Canceler
Brief Description: The ’211 patent relates to systems and methods for canceling cross polarization interference (XPI) in optical fiber communications. The technology involves a receiver that separates an incoming optical signal into orthogonally polarized components and processes them through an XPI canceller (XPIC) to correct for signal distortions incurred during transmission.

Prior Art Relied Upon: Cusani 1992 (a June 1992 journal article), Cusani 1997 (an April 1997 journal article), and Kavehrad (a March 1984 technical journal article).
Core Argument for this Ground:
- Prior Art Mapping: Petitioner argued that Cusani 1992 discloses the core receiver architecture, including separating a polarized signal and routing it through a canceller using a Jones matrix. However, Cusani 1992’s "unitary" matrix could not correct for loss of orthogonality. Cusani 1997, an express elaboration on the 1992 paper, identified "Cross-Polarisation Interference (CPI)" as a "basic problem" and disclosed a "CPI canceller" to address it. To overcome the remaining limitation, Kavehrad taught an adaptive canceller in the analogous radio frequency art that used a non-unitary matrix to expressly correct for "loss of polarization orthogonality."
- Motivation to Combine: A POSITA would combine Cusani 1992 and Cusani 1997 as they address the same system. This combination still suffered from the known inability to correct for loss of orthogonality. A POSITA would have been motivated to look to analogous arts, like radio systems described in Kavehrad, to implement a more robust XPIC using a non-unitary matrix to solve this known problem.
- Expectation of Success: The combination involved applying established matrix-based cancellation techniques to solve a known problem in a predictable manner, yielding no unexpected results.

Prior Art Relied Upon: Cusani 1992, Cusani 1997, and Foschini (Patent 4,631,734).
Core Argument for this Ground:
- Prior Art Mapping: This ground substitutes Foschini for Kavehrad to address the limitations of the Cusani references. Foschini, like Kavehrad, taught a cross-polarization canceller in a radio system that could correct for loss of orthogonality. Crucially for dependent claims 25 and 33, Foschini also disclosed a method to mitigate broader forms of XPI by using complex matrix coefficients to apply both amplitude and phase shifting, thereby correcting for dispersive effects analogous to Polarization Mode Dispersion (PMD).
- Motivation to Combine: A POSITA would combine the base Cusani system with Foschini's teachings to remove the restrictions of a unitary matrix and handle broader, known forms of XPI. Given that PMD in optical systems is similar to multipath fading in radio systems, a POSITA would have found it obvious to apply Foschini's solution to the Cusani optical system.
- Expectation of Success: The combination would predictably improve the performance of the Cusani receiver by applying established signal processing techniques from a well-understood analogous art.

Prior Art Relied Upon: Cusani 1992, Cusani 1997, and Hsieh (a March 1991 journal article).
Core Argument for this Ground:
- Prior Art Mapping: This ground targets claims 35 and 37, which recite specific cancellation methods: a minimum mean square error (MMSE) canceller and a "diagonalizer" that is a "general inverse" of the transmission matrix. Hsieh explicitly compared the performance of diagonalizer and MMSE cancellers, teaching how to implement an MMSE "weight matrix" (Eq. 8) and a diagonalizer that were superior to other known designs.
- Motivation to Combine: A POSITA seeking to implement the specific, advanced cancellation techniques recited in claims 35 and 37 would have been motivated to integrate Hsieh's teachings into the base Cusani system. Hsieh directly provided the necessary engineering details for these exact methods.
- Expectation of Success: Hsieh provided an explicit mathematical framework and structural implementation for both MMSE and diagonalizer cancellers, which a POSITA could have predictably integrated into the Cusani receiver architecture to achieve the claimed functions.
Additional Grounds: Petitioner asserted that claims 16-18 are obvious over the combination of Cusani 1992, Cusani 1997, Foschini, and Tzeng (a 1987 journal article). Tzeng was introduced to teach the use of "double balanced optical receivers," a feature the ’211 patent itself admitted was part of the prior art.

Petitioner dedicated significant argument to the construction of "cross polarization interference" (XPI), which appears in all independent claims.
Petitioner argued for a broad construction to mean "crosstalk among signals... due to one or more of loss of orthogonality of the two optical signals, differential phase retardation, polarization dependent loss, rotation of the signals, etc."
This position was based on the assertion that the patent specification, its prosecution history, and its provisional application all indicate that XPI was understood to be a general "signal crosstalk penalty" caused by numerous factors, not just the "loss of orthogonality" to which it was limited in a prior proceeding. Petitioner argued the applicant explicitly stated during prosecution that its "XPIC network addresses all possible forms of XPI."

Petitioner argued that discretionary denial would be inappropriate despite a previously denied IPR against the same patent (IPR2016-01618, filed by Fujitsu).
Petitioner contended that the Board's prior denial was based on the failure of the previous petitioner to establish that its primary reference (Cusani 1992) disclosed the claimed XPI canceller.
This petition, by contrast, introduced the Cusani 1997 paper, which was not cited in the prior IPR. Petitioner asserted that Cusani 1997 expressly discloses a "CPI [cross polarization interference] canceller," thereby curing the specific deficiency that led to the previous denial and presenting a new, meritorious challenge.

Petitioner requests institution of an inter partes review and cancellation of claims 15-18, 23, 25, 30, 32, 33, 35, and 37 of Patent 6,782,211 as unpatentable under 35 U.S.C. §103.