Flow Motion Research Development Ltd v. Peloton Interactive Inc

I. Executive Summary and Procedural Information

• Parties & Counsel:
  • Plaintiff: Flow Motion Research and Development Ltd. (Israel)
  • Defendant: Peloton Interactive, Inc. (Delaware)
  • Plaintiff's Counsel: Davis Firm PC; Cherian Harkins Dunham
• Case Identification: 2:26-cv-00610, E.D. Tex., 07/24/2026
• Venue Allegations: Venue is based on Defendant Peloton's alleged acts of infringement within the Eastern District of Texas and its maintenance of a regular and established place of business in the district, including a corporate campus and a retail store in Plano, Texas.
• Core Dispute: Plaintiff alleges that Defendant's Peloton Row products equipped with the "Form Assist" feature infringe three patents related to systems for monitoring a user's exercise form and providing real-time feedback.
• Technical Context: The technology involves using motion and position sensors to generate a skeletal or machine representation of a user, comparing it to an optimal form, and providing corrective feedback, a key feature in the competitive connected fitness market.
• Key Procedural History: The three asserted patents are part of the same patent family and claim priority to a provisional application filed in 2011. The complaint does not mention any prior litigation or administrative proceedings involving these patents.

Case Timeline

Date	Event
2011-01-26	Earliest Priority Date for '520, '055, and '035 Patents
2018-06-05	U.S. Patent No. 9,987,520 Issued
2022-09-24	Date of YouTube review video for Peloton Row cited in complaint
2023-09-05	U.S. Patent No. 11,745,055 Issued
2024-12-03	U.S. Patent No. 12,157,035 Issued
2026-07-24	Complaint Filed

II. Technology and Patent(s)-in-Suit Analysis

U.S. Patent No. 9,987,520 - "Method and System for Monitoring and Feed-Backing on Execution of Physical Exercise Routines"

• Patent Identification: U.S. Patent No. 9,987,520, "Method and System for Monitoring and Feed-Backing on Execution of Physical Exercise Routines," issued June 5, 2018.

The Invention Explained

• Problem Addressed: The patent's background section notes that while motion sensors were used in game consoles to track a player's movements, they lacked the ability to identify a specific exercise routine being performed or to provide feedback on the correctness of that routine's execution '987,520 Patent, col. 1:44-59
• The Patented Solution: The invention describes a system that uses motion and position sensors to generate a "skeletal model" of a user performing an exercise '987,520 Patent, col. 1:28-34 A training module then compares this generated model to routine information stored in a database, which represents an "optimal execution" of the exercise, in order to detect "dissimilarities" that indicate incorrect form '987,520 Patent, abstract A key aspect is the use of the "appearance of the exercise device" or a "machine representation model" of the device to assist in processing the sensor data, for purposes such as calibration or aiding the generation of the user's skeletal model '987,520 Patent, col. 17:25-34
• Technical Importance: This technology aims to transform general-purpose motion tracking into a specialized, automated coaching tool capable of providing detailed, real-time performance analysis for complex exercises.

Key Claims at a Glance

• The complaint asserts at least independent Claim 1 Compl. ¶35
• The essential elements of Claim 1 include:
  • A computing device, a plurality of motion and position sensors, a database with routine information, and a training module.
  • The training module is configured to track user movement on an exercise device.
  • The module uses the "appearance of the exercise device, appearance of a feature of it, or machine representation model of thereof."
  • This device appearance/model is used for at least one of: calibrating or adjusting sensor information, or assisting in the generation of a skeletal model of the user.
  • The module generates a skeletal model of the user from the sensory information.
  • The module compares the user's skeletal model to the stored routine information to detect dissimilarities indicating incorrect execution.
• The complaint reserves the right to assert additional claims Compl. ¶29

U.S. Patent No. 11,745,055 - "Method and System for Monitoring and Feed-Backing on Execution of Physical Exercise Routines"

• Patent Identification: U.S. Patent No. 11,745,055, "Method and System for Monitoring and Feed-Backing on Execution of Physical Exercise Routines," issued September 5, 2023.

The Invention Explained

• Problem Addressed: The patent addresses the same problem as the '520 Patent: the inadequacy of existing motion tracking systems for providing detailed, corrective feedback on physical exercise form '055 Patent, col. 1:44-59
• The Patented Solution: The '055 Patent also discloses a system using sensors to track a user's movements on an exercise device. Its training module is configured to use information about the exercise device itself (its appearance, a feature, or a machine representation model) to process the raw sensor data '055 Patent, abstract This processing can involve separating the user from the device, calibrating the sensor information, or assisting in the generation of a "machine representation model" of the user '055 Patent, col. 17:23-32
• Technical Importance: The invention provides a method for improving the accuracy of sensor-based form analysis by leveraging known information about the fixed exercise equipment the user is interacting with.

Key Claims at a Glance

• The complaint asserts at least independent Claim 1 Compl. ¶56
• The essential elements of Claim 1 include:
  • A computing device, a plurality of motion and position sensors, a database with routine information, and a training module.
  • The training module is configured to track user movement on an exercise device.
  • The module uses the "appearance of the exercise device, appearance of a feature of the said exercise device, or machine representation model of thereof."
  • This device appearance/model is used for at least one of: separating the device's appearance from the sensory information, calibrating or adjusting the sensor information, or assisting in the generation of a machine representation of the user.
  • The module generates at least a "machine representation model of the user" from the sensory information.
• The complaint reserves the right to assert additional claims Compl. ¶29

U.S. Patent No. 12,157,035 - "Method and System for Monitoring and Feed-Backing on Execution of Physical Exercise Routines"

• Patent Identification: U.S. Patent No. 12,157,035, "Method and System for Monitoring and Feed-Backing on Execution of Physical Exercise Routines," issued December 3, 2024.
• Technology Synopsis: The patent describes a system using motion and/or position sensors to monitor a user's exercise routine. A central feature is the use of information derived from the exercise device itself-such as its appearance or a machine model-to process the sensor data, for functions including separating the device from the sensor data, calibrating the information, or assisting in generating a machine representation of the user '035 Patent, abstract Compl. ¶72
• Asserted Claims: At least Claim 1 Compl. ¶76
• Accused Features: The "Form Assist" functionality in Peloton's rowing products, which is alleged to use sensors and a representation of the rower to track and provide feedback on a user's rowing form Compl. ¶¶78-83

III. The Accused Instrumentality

• Product Identification: The accused products are the "Peloton Row," "Cross Training Row+," "Peloton Row+ Pro," and subsequent Peloton rowing products that implement the "Form Assist" feature Compl. ¶¶3 Compl. ¶26
• Functionality and Market Context:
  • The "Form Assist" feature is described as using sensors located in the handlebar and seat of the rowing machine to detect the user's body position and movement Compl. ¶38 Compl. ¶43 The system does not use camera technology Compl. ¶12
  • The software allegedly calibrates to each user's body, analyzes their rowing stroke in real-time, and provides feedback on their form Compl. ¶25 This feedback is delivered via a "machine representation of the user in relation to a rower" displayed on the product's screen, which highlights areas of the user's form that need correction Compl. ¶25 Compl. ¶44 The complaint includes a screenshot from a Peloton support article showing an on-screen user silhouette on a rower representation, with an error such as "Over-compression" being highlighted Compl. ¶43
  • Peloton is positioned as a "global fitness and technology company" that provides "interactive exercise equipment" with software and content delivering "real-time instruction, performance metrics, and feedback" Compl. ¶¶8-9

IV. Analysis of Infringement Allegations

'520 Patent Infringement Allegations

Claim Element (from Independent Claim 1)	Alleged Infringing Functionality	Complaint Citation	Patent Citation
a computing device;	The Accused Products include a computing device, such as the integrated screen and processor on the Peloton Row.	¶37	col. 17:10-12
a plurality of motion and position sensors configured to generate sensory information...	The Accused Products use sensors in the handlebars and seat to generate sensory information about the user's movement.	¶38	col. 17:13-17
a database containing routine information representing at least one frame of exercise routine;	The Accused Products contain a database with information representing the proper form for a rowing exercise routine.	¶39	col. 17:18-20
a training module configured to: track at least one movement of a user... on an exercise device;	The "Form Assist" software tracks the user's movements during a rowing exercise.	¶40	col. 17:21-24
use at least one of: appearance of the exercise device, appearance of a feature of it, or machine representation model of thereof;	The "Form Assist" feature displays a machine representation of the rower and its seat. The complaint includes a visual from a Peloton support page showing this representation Compl. ¶41	¶41	col. 17:25-27
for at least one of: ... calibration and/or adjustment of the said generated sensory information, assistance in the generation of skeletal model...	On information and belief, the training module uses the device's appearance or model to calibrate sensor information related to the user's movement or to assist in generating the skeletal model.	¶42	col. 17:28-34
generate at least a skeletal model of the user from the sensory information;	The "Form Assist" feature generates a visual representation, or skeletal model, of the user based on the sensor data.	¶43	col. 17:35-36
compare the generated skeletal model... to routine information... to detect at least dissimilarities...	The system compares the user's form to an optimal form, and if a dissimilarity is detected (e.g., "bend your knees too early"), the corresponding body part lights up in red. A screenshot from a PCMag article shows various form warnings, illustrating this comparison Compl. ¶44	¶44	col. 17:37-43

'055 Patent Infringement Allegations

Claim Element (from Independent Claim 1)	Alleged Infringing Functionality	Complaint Citation	Patent Citation
a computing device;	The Accused Products include a computing device, such as the integrated screen and processor.	¶58	col. 17:3-5
a plurality of motion and position sensors configured to generate sensory information...	The Accused Products use sensors in the handlebars and seat to generate sensory information.	¶59	col. 17:6-10
a database containing routine information representing at least one frame of exercise routine;	The Accused Products contain a database with information representing the proper form for a rowing exercise.	¶60	col. 17:11-13
a training module configured to: track at least one movement of a user... on an exercise device;	The "Form Assist" software tracks the user's movements on the rower.	¶61	col. 17:14-17
use at least one of: appearance of the exercise device, appearance of a feature of the said exercise device, or machine representation model of thereof;	The "Form Assist" feature displays a machine representation of the rower and its seat.	¶62	col. 17:18-22
for at least one of: separate from the sensory information..., calibration and/or adjustment..., assistance in the generation of machine representation of the user...	On information and belief, the training module uses the device's appearance or model to calibrate sensor information or assist in generating the user's machine representation.	¶63	col. 17:23-29
generate at least a machine representation model of the user from the sensory information.	The "Form Assist" feature generates a visual representation of the user based on the sensor data. A screenshot from a Peloton blog shows this representation being generated from handlebar and seat sensors Compl. ¶64	¶64	col. 17:30-32

Identified Points of Contention:

• Scope Questions: The infringement theory may depend on how broadly the court construes "skeletal model" ('520 Patent) and "machine representation model" ('055 Patent). The patents primarily depict traditional stick-figure models '520 Patent, FIG. 1, while the complaint alleges infringement by a more stylized user silhouette Compl. ¶15 A central question will be whether the accused visualization falls within the scope of these terms.
• Technical Questions: The claims require using the device's appearance or model for specific technical purposes like "calibration," "adjustment," or "assistance in the generation" of the user model. The complaint makes these allegations "on information and belief" Compl. ¶42 Compl. ¶63 A key factual dispute will likely be what evidence demonstrates that the accused software uses information about the rower's geometry to actively process the user's sensor data in the manner claimed, as opposed to simply displaying a static background image.
• Technical Questions: For the '520 Patent, the analysis will require evidence that the accused system performs a "comparison" between the user's generated model and stored "routine information" to "detect dissimilarities." The mechanism by which Peloton's "Form Assist" identifies form errors will be a critical point of inquiry for the court.

V. Key Claim Terms for Construction

• The Term: "skeletal model" ('520 Patent) / "machine representation model" ('055 Patent)

  • Context and Importance: These terms are foundational to the infringement allegations, as they define the output of the sensor processing that is then analyzed for feedback. The visual and technical nature of Peloton's on-screen user avatar will be compared against the definition of these terms.
  • Intrinsic Evidence for a Broader Interpretation: The specification describes a skeletal model as "a collection of joints and lines representing the body bones connected" '520 Patent, col. 1:29-31, which could be argued to cover any abstract representation of the human form's key points and segments. The term "machine representation model" is facially broader and may be argued to cover any computer-generated depiction of the user.
  • Intrinsic Evidence for a Narrower Interpretation: The primary embodiment shown in the patents is a classic stick-figure model '520 Patent, FIG. 1 '055 Patent, FIG. 1 A defendant may argue that the term "skeletal model" is limited to this explicit joint-and-bone data structure, and that "machine representation model" should be read in the same context, potentially excluding the filled-in silhouette shown in the accused products Compl. ¶15

• The Term: "use... appearance of the exercise device... for... calibration and/or adjustment of the said generated sensory information"

  • Context and Importance: This limitation appears to be a key point of novelty, linking knowledge of the static exercise machine to the dynamic data from the user. Proving that the accused system performs this function is critical for the plaintiff. Practitioners may focus on this term because it requires a specific technical interaction between device information and user sensor data.
  • Intrinsic Evidence for a Broader Interpretation: A plaintiff could argue this covers any process where known geometric properties of the device are used to filter, validate, or interpret sensor data. For example, using the known fixed position of the seat track to help interpret the movement of the seat sensor could be argued as "adjustment."
  • Intrinsic Evidence for a Narrower Interpretation: The specification provides specific examples, such as a "calibration stage" where the device's location is identified relative to the sensors to allow for its "separation" from the user's image '055 Patent, col. 13:13-34 A defendant might argue the claim is limited to such explicit calibration or separation steps, which may differ from how an integrated system with embedded sensors operates.

VI. Other Allegations

• Indirect Infringement: The complaint alleges inducement of infringement, stating that Peloton actively encourages and instructs its customers to use the accused "Form Assist" feature through user manuals, online support materials, and product demonstrations Compl. ¶46 Compl. ¶66 Compl. ¶85
• Willful Infringement: The complaint alleges willful infringement based on Defendant having knowledge of the asserted patents "at least due to the filing and service of this Complaint" Compl. ¶50 Compl. ¶70 Compl. ¶89 This frames the willfulness claim as being based on post-suit conduct.

VII. Analyst's Conclusion: Key Questions for the Case

1. A core issue will be one of definitional scope: Can the term "skeletal model", which is rooted in the patent's depiction of a stick-figure, be construed to cover the stylized, solid-color silhouette used in Peloton's "Form Assist" feature, or is there a fundamental mismatch in form?
2. A second central question will be evidentiary and technical: What evidence will be presented to show that the accused system actively "uses" a "machine representation model" of the rower for the specific purpose of "calibrating" or "adjusting" user sensor data, as claimed, versus simply displaying the rower as a passive graphical element?
3. For the '520 patent specifically, the case will likely examine the mechanism of error detection: Does the accused "Form Assist" feature operate by performing a direct "comparison" between a generated user model and stored "routine information" to find "dissimilarities", as required by Claim 1, or does it identify form errors through an alternative technical method?