PTAB

IPR2023-01055

Tesla Inc v. Autonomous Devices LLC

Key Events

Petition

petition Intelligence

Title: Autonomous Device Operation System
Brief Description: The ’344 patent discloses systems and methods for enabling autonomous device operation. The technology involves a device learning from user operation by creating correlations between "circumstance representations" (e.g., sensor data, images) and instruction sets, storing these correlations in a knowledgebase, and using them to autonomously perform operations when new, similar circumstances are detected.

Prior Art Relied Upon: Grotmol (Patent 9,604,359).
Core Argument for this Ground:
- Prior Art Mapping: Petitioner argued that Grotmol discloses every element of the challenged claims. Grotmol teaches a robotic apparatus that learns to follow a trajectory through user-led training. During this "learning process," the robot stores correlations between camera images ("circumstance representations") and corresponding motor commands ("instruction sets") in a memory buffer ("knowledgebase"). For autonomous operation, the robot acquires a new image ("second circumstance representation"), compares it to the stored images to find the "best matching stored image" (a partial match), and then executes the motor command associated with that best match. Petitioner asserted this directly maps to the limitations of independent claims 1, 19, and 20. The arguments for dependent claims were based on Grotmol's additional disclosures, such as sharing trained configurations with other robots (meeting "second device" limitations) and using object recognition within images.

Prior Art Relied Upon: Grotmol (Patent 9,604,359) and the general knowledge of a Person of Ordinary Skill in the Art (POSITA).
Core Argument for this Ground:
- Prior Art Mapping: This ground addressed claims requiring learning from an "additional learning process" involving "another user" (claim 10) or "another (third) device" (claim 11). Grotmol taught that multiple learning processes can be used to train a robot on both "simple tasks" and "complex tasks."
- Motivation to Combine (Implicit): Petitioner contended that a POSITA would have found it obvious to use different trainers for different tasks to leverage their varying skills, thus satisfying the "another user" limitation. Furthermore, because Grotmol explicitly taught that a trained configuration from one robot could be loaded onto another, a POSITA would have naturally understood that additional learning processes could involve different devices.
- Expectation of Success: The petition implied a high expectation of success, as applying different users or devices to a known multi-stage training process was presented as a predictable and logical extension of Grotmol’s teachings.

Prior Art Relied Upon: Grotmol (Patent 9,604,359) and Hickman (Patent 8,639,644).
Core Argument for this Ground:
- Prior Art Mapping: This ground focused on claims requiring the use of a server to share learned data between devices. Grotmol taught storing a robot’s "trained configuration" and loading it onto "one or more other robots" to share learned behaviors. Hickman taught a "shared robot knowledge base" implemented in a cloud computing system (i.e., a server). In Hickman's system, a robot uploads its learned experiences, which are then stored and made available to other robots, allowing them to benefit from the collective knowledge.
- Motivation to Combine: A POSITA would combine Grotmol's system of generating transferrable learned behaviors with Hickman's cloud-based system for sharing them. Grotmol created the need to share complex learned data, and Hickman provided an advantageous and well-understood method for doing so at scale. Petitioner noted that Hickman was cited on the face of Grotmol, indicating the technologies were in the same field and known to those skilled in the art.
- Expectation of Success: The combination was argued to be predictable. Integrating Grotmol’s "trained configurations" into Hickman’s shared knowledge base was a straightforward application of a cloud storage solution to enhance a known local data-sharing capability, with no technical hurdles that would undermine a reasonable expectation of success.

"means for processing" (claim 20): Petitioner argued this term, governed by 35 U.S.C. §112(f), requires construction of its corresponding structure. Based on the specification and the parallel language in claim 1, Petitioner asserted the structure corresponding to the claimed functions (generating, anticipating, executing) is a "processor circuit or processor." This construction was critical to mapping claim 20 onto the processor-based system disclosed in Grotmol.

Petitioner requested the institution of an inter partes review and the cancellation of claims 1-20 of the ’344 patent as unpatentable.