AI in Focus – Autonomous Vehicles
One of the single largest applications of Artificial Intelligence (AI) is in the area of autonomous vehicles (AV). Many large companies are working in the AV space, including Uber, Waymo (owned by Alphabet, Google’s parent company), Uber ATG (owned by Uber), and Waterloo-based Blackberry QNX. There are also many startups including Kitchener-based Acerta. Large or small, the companies responsible for developing AVs have pursued a variety of strategies for intellectual property protection including patents, trade-secrets, copyright and trade-marks.
The “autonomous” feature of AVs is provided by many different individual systems, including:
- Sensing systems such as vision systems, radar, LiDAR, GPS;
- Control systems including lane keeping, distance keeping and braking, anti-collision systems, etc;
- Navigation systems including trip planning, route planning, mapping, etc; and
- Machine learning systems, such as object recognition, traffic sign and road marking recognition, etc.
LiDAR (or Light Detection and Ranging) is an example that assists us in understanding some of the challenges of intellectual property protection of inventions in the autonomous vehicles space. LiDAR is a surveying method used to measure distances to a target by shining pulsed laser light on a surface and measuring its reflection, and is one of the most interesting technologies associated with autonomous vehicles. Many of the players in the AV industry use LiDAR as a sensing technology to determine relevant objects in the vehicle’s environment and as data collection systems for other control and navigation systems. There are other sensor technologies used, including computer vision systems, and RADAR systems.
A major method of product protection of AI based inventions such as AVs is patent protection. In the case of LiDAR, there are many examples of granted patents covering the use of LiDAR by AVs including U.S. patents Nos. 9841763, 9254846,9097800, 8,836,922, 9,368,936, and 9,086,273.
The ‘763 U.S. patent is assigned to Uber, and discloses a predictive sensor array configuration system for an autonomous vehicle. The system can dynamically detect reflected anomalies affecting the sensor array as the AV travels along a route. The anomalies may include light conditions such as shadows or reflections, and based on these anomalies the configuration system may dynamically generate a set of configurations for the sensor array to compensate for the imminent lighting conditions. The system is predictive and can determine the configurations to preemptively compensate for anomalies that the AV encounters along its route.
The ‘846 U.S. patent is assigned to Waymo, and discloses a machine learning system for predictive reasoning method and system of controlling the speed of a vehicle. The ‘846 patent describes the use of sensor systems such as LiDAR, GPS or RADAR. The ‘846 patent includes the use of multiple LiDAR arrays that can scan a complete 360 horizon around the vehicle. The LiDAR array generates for the computing device controlling the AV a cloud of “point data representing obstacles or objects, which have been hit by the laser (in the LiDAR array) on the road and in the vicinity of the road”. The LiDAR array may provide intensity values of the reflected laser light.
The ‘800 U.S. patent is assigned to Waymo, and discloses a light detecting and ranging device associated with an autonomous vehicle. Based on a three-dimensional point map generated by the LiDAR system the method determines a three-dimensional point map of the scanning zone. The claimed ‘800 method further discloses further using RADAR sensors in the scanning zone corresponding to light reflective features and shows RADAR assisted LiDAR sensing.
Waymo v. Uber
Another avenue of product protection for AVs is trade-secret protection. LiDAR technology has been the subject of litigation between Uber and Waymo based on trade-secret protection. In Waymo v. Uber, Waymo alleged theft of trade secrets and patent infringement by Uber based on Uber’s acquisition of Uber ATG, a self-driving truck startup. Uber ATG’s LiDAR technology was widely touted as one of the keys for its acquisition by Uber.
Waymo alleged that Uber ATG, a company started by former Waymo employee Anthony Levandowski had taken trade secret information prior to quitting to start Uber ATG. Waymo alleged that Levandowski had downloaded more than 14,000 files prior to resigning to start Uber ATG. Waymo argued that this trade secret information was used by Levandowski and Uber ATG to build LiDAR sensor components. The trade secret information included aspects of the design of printed circuits used by the LiDAR, including the position and orientation of laser diodes and photodetectors on the circuit boards. Other trade secret information included in Waymo’s claim were the selection, materials, size, position and orientation of optical elements used in the laser transmitter/detector, the resolution profile used in determining the position and orientation of laser diodes, and the know-how associated with the resolution profile and object detection, and the laser transmission frequency used to accurately detect objects.
Under the US Defend Trade-Secrets Act, a trade secret is information including practices, processes, designs, etc where its proprietor takes reasonable measures to keep the information secret and where the information derives independent economic value (actual or potential) from not generally being known or ascertainable. While details of the operation of Waymo’s LiDAR system are included in patent publications, several specific details of its implementation in their AVs was kept as trade-secret information.
The patent infringement claims alleged by Waymo were based on the ‘922 patent, the ‘936 patent and the ‘273 patent. The patent infringement allegations for the ‘922 patent and the ‘273 patent related to an early LiDAR design no longer in use by Uber, and the claims were subsequently dropped. The ‘936 patent however, remained at issue.
Unexpectedly during the course of litigation however, a third party named Eric Swildens submitted an Ex parte re-examination request for the remaining ‘936 patent. Re-examination requests are a rarely used process whereby a third party can have a patent re-examined by an examiner at the USPTO to determine its patentability. Mr. Swildens submitted the re-examination request along with several prior art references.
The Waymo v Uber litigation settled after only five days at trial, but the Ex parte re-examination of the ‘936 patent continues. The AV space, including developing LiDAR technologies will be an interesting space to continue to watch as heavyweights and start-ups alike begin to commercialize AV products.