Obviousness

Obviousness Analysis under 35 U.S.C. § 103

This analysis identifies combinations of prior art that would render the claims of US patent 6968001 obvious to a person having ordinary skill in the art (PHOSITA). The motivation to combine these references would stem from the desire to improve equalizer performance, adaptability, and efficiency in communication systems, especially in the context of time-varying channels and variable data rates.

A PHOSITA in the field of communication systems equalization would have a strong understanding of adaptive filters, various equalizer structures (e.g., FIR, DFE), performance metrics (e.g., MSE, C/I, SINR), and the challenges of optimizing equalizer performance in dynamic environments. They would also be aware of the trade-offs between equalizer length, adaptation speed, and steady-state performance.

Combination 1: US5666378A in view of US5268930A and general knowledge of adaptive equalizers.

This combination would render obvious the method and apparatus claims (Claims 1, 11, 12, 13, and 15) related to training an equalizer with different configurations and selecting based on performance, particularly for equalizers whose configurations involve filter coefficients.

• US5666378A (High performance modem using pilot symbols for equalization and frame synchronization): This patent discloses using pilot symbols for equalization and synchronization in a modem. This directly addresses the "training the equalizer on a received pilot signal" aspect found in claims 13 and 15, and the general concept of using a "known sequence of samples" for training in claim 2. It teaches that pilot signals can be used to perform channel state information (CSI) measurements, which are then used to precondition data for transmission.
• US5268930A (Decision feedback equalizer): This patent describes a decision feedback equalizer (DFE). A DFE is a type of nonlinear equalizer that reduces intersymbol interference (ISI) and consists of a tapped delay line with a forward and feedback filter. DFEs are known to adapt based on actual decisions and do not amplify noise, making them effective for high-speed links. A PHOSITA would understand that varying the lengths of the feedforward and feedback filters in a DFE constitutes different equalizer configurations. The concept of setting filter coefficients to zero to effectively change equalizer length is a well-known technique in adaptive filtering.
• Motivation for Combination: A PHOSITA would be motivated to combine these references to enhance the performance and adaptability of equalizers in communication systems. US5666378A provides the mechanism of using pilot signals for training and performance estimation. US5268930A (and the general knowledge of DFEs) offers a specific type of equalizer with adjustable filter lengths, which can be seen as different "configurations." The problem of optimizing equalizer performance for varying channel conditions (as described in US6968001) is a known challenge. Adaptive equalizers are commonly used to automatically adjust to time-varying channel properties. Therefore, it would be obvious to a PHOSITA to train a DFE, as described in US5268930A, using pilot symbols as disclosed in US5666378A, and to evaluate the performance of different DFE configurations (e.g., varying tap lengths) to select the optimal one. The choice of different filter coefficients directly relates to "determining a first/second set of values for the set of parameters" and "setting a second set of values...effectively changes the configuration of the equalizer" as stated in claims 1, 11, 12, 13, and 15. The estimation of a "performance parameter" (e.g., C/I or MSE) is also a standard practice for evaluating equalizer performance.

Combination 2: US20020196844A1 in view of general knowledge of adaptive equalizers and performance metrics.

This combination would target the method and apparatus claims, focusing on adaptive equalization and the general concept of trying different equalizer parameters and selecting the best one.

• US20020196844A1 (Adaptive equalizer system for short burst modems and link hopping radio networks): This reference explicitly discusses adaptive equalizer systems, which are designed to adapt to time-varying properties of the communication channel. It would be understood by a PHOSITA that adaptive equalizers inherently involve adjusting filter coefficients or other parameters to optimize performance. Adaptive equalization involves calculating parameters of the equalizing filter based on the received signal, often using training signals or statistical models.
• General Knowledge of Adaptive Equalizers and Performance Metrics: It is common knowledge in the field that adaptive equalizers constantly change their filter characteristics over time to track time-varying channels. Algorithms like Least Mean Square (LMS) and Recursive Least Squares (RLS) are widely used to determine tap coefficients and update them. Performance metrics such as Carrier-to-Interference (C/I) ratio, Signal-to-Noise Ratio (SNR), Mean Square Error (MSE), or Signal to Interference and Noise Ratio (SINR) are standard for evaluating equalizer effectiveness.
• Motivation for Combination: Given that adaptive equalizers (as in US20020196844A1) are designed to handle varying channel conditions, a PHOSITA would be motivated to systematically explore different operating "configurations" or "parameters" of such an equalizer to optimize its performance. This optimization would naturally involve defining different sets of parameters, training the equalizer with these sets, estimating a performance parameter for each, comparing these parameters, and selecting the best configuration for subsequent data processing. This directly addresses the core method of US6968001 claims. The patent itself notes that "optimizing the equalizer by selecting the number of taps is difficult, as this requires a balancing of competing goals" and that an "optimum number of taps balances such considerations and compromises between good steady-state performance and good transient performance." This highlights the known problem that US6968001 attempts to solve, and the proposed solution would be an obvious application of known adaptive equalization techniques.

Combination 3: US5787118A in view of US6735244B1 and knowledge of common equalizer parameters.

This combination focuses on the broader concept of varying equalizer parameters beyond just tap coefficients, such as phase offset or sector identifiers.

• US5787118A (Adaptive diversity equalizer): This patent describes an adaptive diversity equalizer, implying the handling of multiple received signals, possibly from different paths or antennas, to improve overall signal quality. Diversity techniques are used to compensate for fading channel impairments and often involve multiple receiving antennas. While the specific parameters it varies are not explicitly detailed in the abstract, the term "diversity" suggests the possibility of managing multiple input streams or channels.
• US6735244B1 (Data transmission system and receiver unit thereof): This patent broadly covers a data transmission system and receiver unit. While the abstract does not specify equalizer parameters, it represents general knowledge in receiver design.
• Motivation for Combination: A PHOSITA would understand that in complex communication environments, especially those involving diversity or multiple possible signal sources (e.g., soft handoff in cellular systems), an equalizer's optimal performance might depend on parameters beyond just filter tap coefficients. For instance, in a soft handoff situation, a receiver might be able to receive data samples from multiple transmitters, and a "sector identifier" could be a relevant parameter to optimize reception. US6968001 explicitly mentions that equalizer parameters can include "a DC offset value or a phase offset value for the input samples, or a sector identifier." US5787118A, dealing with diversity, would motivate a PHOSITA to consider how different signal paths or sources could be treated differently by an equalizer. It would be obvious to extend the concept of trying different equalizer configurations (as in Combination 1 and 2) to include other known communication system parameters such as phase offset (for accurate symbol detection) or sector identifiers (for multi-cell reception) to find the best reception quality in a system context described by US6735244B1.

Obviousness Considerations under 35 U.S.C. § 103

The legal standard for obviousness under 35 U.S.C. § 103 requires determining whether "the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains."

The claims of US patent 6968001 broadly cover a method and apparatus for optimizing an equalizer by iteratively testing different parameter sets, measuring performance, comparing, selecting the best, and configuring the equalizer. The problem of optimizing equalizer performance for time-varying channels is a known challenge in communication systems. Adaptive equalizers, by their nature, seek to automatically adjust to channel conditions.

The prior art demonstrates that:

• Adaptive equalizers were well-known and used to combat ISI in time-varying channels.
• Decision Feedback Equalizers (DFEs) were known as a type of adaptive equalizer with adjustable forward and feedback filter lengths.
• The use of pilot signals for training equalizers and estimating performance (e.g., C/I, MSE) was a standard practice.
• The concept of varying equalizer "parameters" or "configurations" to optimize performance was also implicitly or explicitly understood in the context of adaptive equalizers, where coefficients are continuously adjusted. The explicit mention in US6968001 of dynamically adjusting the effective length of an equalizer by setting tap coefficients to zero is a known technique.
• The use of performance metrics (like C/I, MSE, SINR) for comparing different equalizer settings was standard.

Therefore, a PHOSITA, seeking to improve the "universal" performance of a receiver over a variety of channel conditions and rates of channel variation (the stated goal of US6968001), would have been motivated to combine these known elements. The idea of "virtual parallel equalizers" in US6968001, where a single piece of hardware is reconfigured rapidly to emulate multiple equalizers, leverages the idle time of a fast equalizer. This is an efficient use of existing resources, a common engineering goal. Given that training and filtering take "only a fraction of the allowable time T", the idea of performing multiple trainings sequentially within a half-slot time would be an obvious implementation choice to maximize resource utilization and adapt to varying conditions.

The specific "equalizer parameters" beyond tap coefficients, such as DC offset, phase offset, or sector identifiers, are also well-known aspects of receiver design and signal processing. Applying the iterative optimization method to these parameters would be a straightforward extension of known adaptive equalization principles.

For these reasons, the claims of US patent 6968001, particularly those related to determining multiple equalizer configurations, training, estimating performance, comparing, selecting, and then configuring the equalizer based on that selection, would have been obvious to a PHOSITA at the time of the invention.