Obviousness

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

To establish obviousness under 35 U.S.C. § 103, it must be shown that the claimed invention as a whole would have been obvious to a person having ordinary skill in the art (POSA) at the time of the invention. This requires identifying:

1. Prior art references that teach or suggest elements of the claimed invention.
2. A motivation to combine these references to achieve the claimed invention.
3. A reasonable expectation of success in combining the references.

Prior Art References for US10624575

The patent US10624575 itself lists prior art keywords as "microactivity," "user," "threshold," "acceleration data," and "feature value." The Cooperative Patent Classification (CPC) codes assigned to US10624575 are particularly relevant for identifying pertinent prior art: A61B5/4806 (Sleep evaluation), A61B5/4815 (Sleep quality), A61B5/11 (Measuring movement of the entire body or parts thereof), A61B5/1118 (Determining activity level), and A61B5/72 (Signal processing specially adapted for physiological signals or for diagnostic purposes).

Given these classifications, a POSA would be aware of various technologies related to:

• Wearable devices for physiological monitoring: The patent explicitly mentions wearable devices such as wristbands, watches, rings, and necklaces, which are common in health monitoring.
• Accelerometer data for activity sensing: The use of multi-axial accelerometers to collect acceleration data for determining activity is a well-established technique in the field of physiological monitoring and activity tracking. CPC A61B5/11 directly covers "Measuring movement of the entire body or parts thereof, e.g. head or hand tremor or mobility of a limb," and A61B5/1118 covers "Determining activity level."
• Sleep monitoring: The core of US10624575 is sleep monitoring, which falls under CPC A61B5/4806 ("Sleep evaluation") and A61B5/4815 ("Sleep quality").
• Signal processing of physiological data: The patent details various signal processing steps, including determining mean values, difference vectors, and weighted sums, which are standard signal processing techniques. CPC A61B5/72 is for "Signal processing specially adapted for physiological signals or for diagnostic purposes," with sub-groups like A61B5/7235 for "Details of waveform analysis" and A61B5/7264 for "Classification of physiological signals or data."

Obviousness Combinations

Independent Claim 1 of US10624575 outlines a method for monitoring sleep. We will analyze its obviousness by combining known prior art elements.

Combination 1: Prior Art generally disclosing wearable devices with accelerometers for activity tracking + Prior Art disclosing signal processing techniques for activity data + Prior Art disclosing using activity data for sleep monitoring.

• Teaching/Suggestion:

  • Wearable devices with accelerometers for activity tracking: The use of wearable devices with accelerometers to measure a user's movement for activity tracking is fundamental and widely known in the art. Companies like Philips, Samsung, Apple, and Fitbit are prominent assignees in the IPC class A61B5/11, which covers measuring body movement. Such devices are routinely used to collect multi-axial acceleration data.
  • Signal processing techniques for activity data: The specific steps in Claim 1 for determining activity amount (obtaining distance vectors, rotating elements, extracting elements for reduced-dimension vectors, summing, squaring, and dividing) represent common mathematical and signal processing operations for quantifying movement from raw accelerometer data. The CPC A61B5/72 and its sub-classifications such as A61B5/7235 and A61B5/7264 confirm that signal processing, waveform analysis, and classification of physiological signals are standard practices in diagnostic and monitoring applications. For instance, calculating Euclidean distances from acceleration data is a basic way to represent magnitude of movement. Operations like translation (rotation of elements) and extracting subsets of data are standard data manipulation techniques for feature extraction in time series analysis.
  • Using activity data for sleep monitoring: Prior art clearly indicates that activity data from wearable devices has been used to determine sleep states and evaluate sleep quality. The CPC A61B5/4806 specifically addresses "Sleep evaluation" and A61B5/4809 addresses "Sleep detection, i.e. determining whether a subject is asleep or not," indicating that determining sleep from activity is a known goal. The patent itself acknowledges that "A wearable device can be used to record activity amount of a user to determine a sleep state of the user, and evaluate sleep quality of the user using the sleep state."

• Motivation to Combine: A POSA would be motivated to combine these elements to improve the accuracy and robustness of sleep monitoring systems. Existing systems might be inaccurate due to "microactivities" that occur when a user is awake but still (e.g., reading, using a phone) or when the device is not being worn. The motivation would be to refine the "activity amount" calculation and introduce a "microactivity feature value" to better distinguish between these states and actual sleep. This is explicitly stated as a problem to be solved in US10624575's background: "Some microactivities of the user before or after sleep (e.g., reading, using a cell phone, etc.) are similar to those in the sleep state. Although an activity amount can be used to determine the sleep state, the microactivities of the user often cannot be accurately recognized, which can affect monitoring the sleep quality of the user." A POSA would seek to improve this by applying more sophisticated signal processing to discern subtle differences in movement patterns. The detailed steps in Claim 1 for calculating the activity amount and then the microactivity feature value are directed precisely at this goal.

• Reasonable Expectation of Success: Given the widespread use of accelerometers in wearables and the established techniques for signal processing and sleep detection, a POSA would have a reasonable expectation of success in combining these known elements. The mathematical operations for transforming acceleration data into an activity amount, and then further processing that data into a microactivity feature value, are standard engineering practices. The identification of thresholds to differentiate between states (waking, sleeping, not wearing the device) is also a routine calibration and classification task in such systems.

Conclusion for Claim 1 Obviousness:

Claim 1, while detailing specific steps for calculating activity amount and a microactivity feature value, essentially describes applying known signal processing techniques to accelerometer data from a wearable device to more accurately determine a user's sleep state by distinguishing subtle movements from actual sleep or non-wearing states. The individual components of the claim—wearable accelerometers, activity detection, signal processing (including transformations like rotations, extractions, sums, and squares), and the use of thresholds for classification (e.g., distinguishing near-zero activity from actual zero activity)—were well-known in the prior art related to physiological monitoring and sleep analysis. A POSA, facing the known challenge of distinguishing low-activity states in sleep monitoring, would have been motivated to refine activity measurement and apply further signal processing to create a feature value that could differentiate these states, with a reasonable expectation of success.