Obviousness

Based on the prior art references detailed previously, this analysis examines the potential obviousness of US patent 10,959,649 under 35 U.S.C. § 103.

Obviousness Standard and the Person of Ordinary Skill

Under 35 U.S.C. § 103, a patent claim is invalid if the differences between the claimed invention and the prior art are such that the invention as a whole would have been obvious at the time the invention was made to a person having ordinary skill in the art (PHOSITA).

For the technology in patent '649, a PHOSITA would likely be an engineer or computer scientist with a bachelor's degree and experience in wearable consumer electronics, sensor data processing (specifically from accelerometers and GPS), and the development of fitness-tracking algorithms.

Obviousness Combination of Prior Art

A strong argument for obviousness can be constructed by combining the teachings of US 2008/0133139 A1 (Vock et al.) and US 2014/0278211 A1 (Yuen et al.). The combination of these references teaches nearly all elements of the independent claims of patent '649, and the remaining gap could be considered an obvious step to a PHOSITA seeking to improve accuracy.

1. What the Prior Art Teaches:

• Vock et al. discloses the core system for stride length calculation. It teaches:

  • Using an accelerometer to determine a user's step rate (step frequency).
  • Maintaining a database that correlates step rates with corresponding stride lengths.
  • Using a GPS receiver to measure the actual distance traveled over a number of steps.
  • Calculating a "verified" average stride length for a specific activity by dividing the GPS-measured distance by the step count.

• Yuen et al. teaches personalizing this type of system from the outset. It discloses:

  • A wearable fitness monitor that determines stride length.
  • Using user-specific physical characteristics, such as height and gender, to calculate an initial stride length.
  • Using GPS to track activity and calibrate fitness metrics.

2. Motivation to Combine Vock et al. and Yuen et al.:

A PHOSITA would have been motivated to combine the teachings of Vock et al. and Yuen et al. to create a more accurate and user-friendly device. Vock’s system relies on a generic correlation between step rate and stride length, which could be inaccurate for a specific individual. Yuen teaches that initial accuracy can be significantly improved by using personal data like height and sex. A PHOSITA would see a clear benefit in using Yuen's personalization method to create the initial database of estimated stride lengths used in Vock's system. This would provide the user with a better "out-of-the-box" experience before any GPS calibration is performed. The combination is a predictable merging of two known methods to achieve a better result.

3. Analysis of the Combination Against the Claims:

When combined, Vock et al. and Yuen et al. teach the following elements of independent claim 1:

• Receiving user-specific information: Taught by Yuen et al.
• Determining a step rate: Taught by Vock et al.
• Determining estimated stride lengths for different step rates based on user information: This is the direct result of combining Yuen's personalization with Vock's step-rate-based table.
• Determining a verified stride length using location data (GPS): Taught by Vock et al.
• Storing the verified stride length: Implicitly taught by both for calibration purposes.

The one element not explicitly taught by this combination is the key limitation that allowed patent '649 to be granted:

...wherein the verified stride length associated with the first step rate or the first range of step rates is used to adjust the one or more estimated stride lengths associated with one or more second step rates or second ranges of step rates different from the first step rate or the first range of step rates.

4. Argument for Why the Final Step is Obvious:

Despite the USPTO's allowance, a strong argument can be made that bridging this final gap would have been obvious to a PHOSITA.

• Known Problem: The background of patent '649 itself notes the problem that a user's stride length varies at different speeds. A PHOSITA would be keenly aware that a calibration performed at a single speed (e.g., running) might not improve accuracy at other speeds (e.g., walking) if not applied correctly.
• Predictable Solution: The combined Vock and Yuen system provides a PHOSITA with a table of estimated stride lengths (the model) and a method to find a "ground truth" data point via GPS verification. If the verified stride length for a 120 steps/minute run is found to be 5% longer than the initial estimate, it is a common and logical step in data calibration to infer that the entire estimation model is off by a similar factor. Applying a proportional adjustment to the estimated stride lengths at other step rates is a straightforward application of data-fitting and error-correction principles.
• Finite Number of Identifiable, Predictable Solutions: To improve the overall accuracy of the stride length model across all step rates (the known problem), a PHOSITA would consider a few predictable solutions: (a) require the user to perform GPS calibration at every possible step rate, which is impractical (as noted by Oh et al.), or (b) use a single, verified data point to intelligently adjust the entire model. The latter approach, which involves inferring error and adjusting proportionally, is a standard technique in algorithm design and would be an obvious path to try.

Therefore, the final inventive step of using a verified stride length from one step rate to adjust estimated stride lengths at other rates represents the application of a well-known calibration technique to solve a known problem in the art. While the applicant successfully argued this step as non-obvious during prosecution, it remains vulnerable to an obviousness challenge by framing it as a predictable design choice for any engineer tasked with improving the accuracy of a fitness tracker.