Axon Enterprises Inc v. Airspace Systems Inc

1. Case Identification

• Case #: IPR2025-00629
• Patent #: 10,514,711
• Filed: February 24, 2025
• Petitioner(s): Axon Enterprise, Inc., Dedrone Holdings, Inc., and Skydio, Inc.
• Patent Owner(s): Airspace Systems, Inc.
• Challenged Claims: 1-20

2. Patent Overview

• Title: Flight Control Using Computer Vision
• Brief Description: The ’711 patent relates to systems and methods for the flight control of an aerial vehicle. The technology involves receiving an image from an onboard sensor, detecting a target within the image using a machine learning model, determining a three-dimensional relative location of the target, and performing a flight control operation based on that location.

3. Grounds for Unpatentability

Ground 1: Claims 1-3, 5, 6, 9, 10, 12, 14-16, and 18-20 are obvious over Beard in view of Laufer.

• Prior Art Relied Upon: Beard (Patent 9,769,387) and Laufer (Application # 2017/0059692).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Beard taught a UAV flight control system that captures images of a target, uses image processing to determine a desired orientation, and tracks the target by adjusting its position and speed. Beard’s system determined a 3D relative location based on image parameters like pixel size, elevation, and bearing, and adjusted the UAV’s speed based on deviation between the target’s direction and the UAV’s current flight direction. Petitioner contended Laufer disclosed a surveillance system that used Doppler radar to guide an interceptor drone to a target's vicinity and then used machine learning on camera images to identify and track drone threats.
  • Motivation to Combine: A POSITA would combine Laufer with Beard to enhance Beard’s object detection and target acquisition capabilities. Laufer’s use of machine learning would provide a known, automated technique for implementing Beard’s described object detection. Furthermore, Laufer’s method of using radar to establish initial visual contact offered a more robust and flexible way to guide the UAV to a target, especially when out of visual range, compared to Beard’s reliance on data from a target's smartphone.
  • Expectation of Success: The combination involved applying known and predictable techniques (machine learning for object detection, radar for initial guidance) to a similar system (drone flight control). A POSITA would have expected success in integrating these elements to create a more effective target tracking system.

Ground 7/8: Claims 1, 3, 5-9, 15, 19, and 20 are anticipated by Pierce (Ground 7) or obvious over Pierce in view of Kumar (Ground 8).

• Prior Art Relied Upon: Pierce (Application # 2017/0160751) and Kumar (Application # 2017/0161911).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner asserted Pierce described a method for controlling drone movement for object tracking using an attached camera and a neural network to generate a bounding box around a target. For calculating the target’s relative position, Pierce explicitly incorporated the teachings of Kumar by reference. Kumar taught a system for determining a target's 3D relative position using parameters derived from a bounding box, such as its area in pixels and the known physical size of the object type. Petitioner argued this combination taught every limitation of the independent claims, including performing flight control based on the 3D location and determining a speed adjustment factor based on the difference between the target’s yaw angle and the drone’s yaw angle.
  • Motivation to Combine: The anticipation ground relied on Pierce's express incorporation of Kumar, making Kumar's disclosure part of Pierce's own. For the obviousness ground, a POSITA would have been motivated to look to Kumar because Pierce’s system required a target's relative position to function but did not specify how to calculate it, instead providing an express suggestion to use Kumar’s position estimation system for that purpose.
  • Expectation of Success: A POSITA would have had a high expectation of success, as the combination merely involved implementing the specific calculation method from Kumar that Pierce itself directed the artisan to use.

Ground 9/10: Claims 2, 12, and 18 are obvious over Pierce and Laufer (Ground 9) or Pierce, Kumar, and Laufer (Ground 10).

• Prior Art Relied Upon: Pierce (Application # 2017/0160751), Laufer (Application # 2017/0059692), and Kumar (Application # 2017/0161911).

• Core Argument for this Ground:

  • Prior Art Mapping: This combination augmented the Pierce/Kumar system with Laufer's teachings. Petitioner argued that Laufer’s disclosure of using ground-based Doppler radar to generate an "interception location" taught the limitation in claim 12 of receiving a geographical location from a ground-based sensor and autonomously navigating toward it. Laufer’s explicit teaching of using quadcopter or multirotor drones was argued to satisfy the multirotor UAV limitation of claim 2.
  • Motivation to Combine: A POSITA would combine Laufer with the Pierce/Kumar system to overcome a key deficiency: the inability to track a target not visible to the drone’s camera. Laufer provided a known solution by using radar to guide the drone to the target's general location, at which point the more precise image-based tracking system of Pierce/Kumar could take over. This would improve the overall efficiency and flexibility of the system.
  • Expectation of Success: The integration was predictable, as it involved combining a well-known radar guidance system with an image-based tracking system to solve the common and foreseeable problem of initial target acquisition when the target is out of visual range.

• Additional Grounds: Petitioner asserted numerous additional obviousness challenges based on combinations of Beard and Pierce with other references, including Khan (for model-type classification), Malecki (for 3D vector conversion), Sahasrabudhe (for limiting speed adjustment factors), and Hogasten (for radar-based distance measurement), to address various dependent claim limitations.

4. Arguments Regarding Discretionary Denial

• Petitioner argued that discretionary denial under Fintiv would be improper because the parallel district court litigation was in its nascent stages, with no trial date set and minimal investment by the parties.
• Petitioner also contended that denial under 35 U.S.C. §325(d) was unwarranted because the prior art references and combinations presented in the petition were not substantively considered or applied against the claims during the original prosecution history.

5. Relief Requested

• Petitioner requests institution of an inter partes review (IPR) and cancellation of claims 1-20 of Patent 10,514,711 as unpatentable.