Tesla Inc v. Granite Vehicle Ventures LLC

1. Case Identification

• Case #: IPR2025-00944
• Patent #: 11,738,765
• Filed: May 2, 2025
• Petitioner(s): Tesla, Inc.
• Patent Owner(s): Granite Vehicle Ventures LLC
• Challenged Claims: 1-8, 10-20

2. Patent Overview

• Title: Fault Management System for Autonomous Vehicles
• Brief Description: The ’765 patent describes a computer-implemented method for managing the operation of a self-driving vehicle (SDV) by assessing its operational state. The system determines if a fault has occurred, evaluates whether the fault exceeds a danger threshold by calculating processor and/or human driver competence levels, and implements a corrective action using a fault-remediation table.

3. Grounds for Unpatentability

Ground 1: Obviousness over Attard, Frazer, and McNew - Claims 1-2 and 17 are obvious over Attard, Frazer, and McNew.

• Prior Art Relied Upon: Attard (Patent 9,406,177), Frazer (Patent 9,494,926), and McNew (Patent 10,377,303).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Attard, the primary reference, disclosed the core system of an SDV that uses sensor data to describe an operational state and determines faults by comparing "confidence assessments" against thresholds. When a confidence assessment (φ) falls below a threshold (φmin), a fault is determined to have occurred. For limitations not expressly taught by Attard, Petitioner asserted that Frazer taught the use of a fault-remediation table (its "event table 112") to organize and execute pre-defined corrective actions based on specific detected conditions. To address the corrective action of transferring control to a human driver, Petitioner asserted McNew taught disengaging a semi-autonomous system and transitioning the vehicle to manual control in response to events like driver inattention.
  • Motivation to Combine: A POSITA would combine Attard with Frazer to improve the efficiency and reduce the latency of Attard's fault-handling system. Frazer's use of an event table was a well-known technique for organizing if-then logic to quickly determine appropriate responses. A POSITA would further incorporate McNew's teachings to add a critical safety feature: transitioning to manual control when the autonomous system is insufficiently competent, an action already contemplated by Attard.
  • Expectation of Success: Petitioner contended a POSITA would have a high expectation of success, as the combination involved applying a known data organization technique (Frazer's table) to a known fault-handling method (Attard's system) and implementing a known safety response (McNew's manual transition), all using conventional processing and memory components disclosed in the references.

Ground 2: Obviousness over Hampiholi and Attard - Claim 7 is obvious over Hampiholi and Attard.

• Prior Art Relied Upon: Hampiholi (Application # 2016/0267335) and Attard (Patent 9,406,177).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued this combination taught a system capable of determining both a human driver competence level (HDCL) and a processor competence level (CPCL). Hampiholi was asserted to teach determining an HDCL by measuring a driver's distraction level (a "severity rank R") based on sensor inputs like driver-facing cameras. Attard was asserted to teach determining a CPCL (its overall confidence assessment "Φ") based on sensor data related to the vehicle's ability to operate autonomously. The combined system would thus use both competence levels to determine and implement corrective actions.
  • Motivation to Combine: A POSITA would combine Hampiholi and Attard to create a more comprehensive and safer SDV. Such a system could account for faults arising from both driver incompetence (e.g., distraction, as taught by Hampiholi) and system/environmental issues (e.g., sensor failure or bad weather, as taught by Attard). This would further Hampiholi's stated goal of increasing safety by allowing the vehicle to respond to a wider range of potential failure modes.
  • Expectation of Success: Success was predictable because both references described vehicle control systems using sensors and processors. Integrating Attard's processor competence calculation into Hampiholi's driver-focused system would involve known programming techniques to create a more robust decision-making logic for triggering corrective actions.

Ground 3: Obviousness over Attard, Frazer, McNew, An, Kang, and Schunder - Claims 3-4 and 18-19 are obvious over this combination.

• Prior Art Relied Upon: Attard (Patent 9,406,177), Frazer (Patent 9,494,926), McNew (Patent 10,377,303), An (Patent 9,063,543), Kang (Patent 9,688,145), and Schunder (Patent 9,451,030).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground built upon the base combination of Attard, Frazer, and McNew to add the claimed "weighted voting system." Petitioner argued that An taught using a weighted voting system to compute an "autonomous driving state risk index," which is analogous to Attard's confidence assessment. To meet limitations requiring weights based on active learning data, Petitioner introduced Kang, which taught weighting inputs based on weather condition data (e.g., rain, snow). Finally, Schunder was asserted to teach sourcing this weather data from a "cohort of other vehicles" via a server that collects and distributes crowd-sourced reports.
  • Motivation to Combine: A POSITA would be motivated to modify the base combination with An's weighted voting to improve the accuracy of the processor competence calculation. Further, a POSITA would incorporate Kang's weather-based weights and Schunder's crowd-sourced data to make the system more adaptable and robust, allowing it to give greater weight to sensor inputs from more hazardous conditions (e.g., icy roads) and to use more accurate, real-time weather data from nearby vehicles.
  • Expectation of Success: The combination represented the use of known techniques to improve a known system. Integrating weighted inputs was a common method for refining control system logic, and using crowd-sourced data for environmental conditions was an established concept. A POSITA would have found it a matter of straightforward programming to implement this multi-layered weighted voting system.
• Additional Grounds: Petitioner asserted eight additional obviousness challenges (Grounds 4-11) against the remaining claims. These grounds relied on various combinations of the primary references cited above, with the addition of DeRuyck (Patent 9,714,037) for determining human driver competence using weighted inputs and Gunderson (Application # 2007/0268158) for basing those weights on data from a cohort of other drivers.

4. Relief Requested

• Petitioner requests institution of an inter partes review and cancellation of claims 1-8 and 10-20 of the '765 patent as unpatentable.