Ford is Developing a Lane-Splitting Detection System

Dennis Chung
by Dennis Chung
Ford has filed a patent application for a system to detect lane-splitting motorcycles. The system uses various sensors on a car, such as cameras, RADAR, LIDAR and SONAR, to determine if a lane-splitting motorcycle is approaching from the rear. If a filtering bike is detected, the car would include that information in its collision avoidance system. Though mainly conceived for use in self-driving cars, the principles behind the design can also work for non-autonomous vehicles.

Ford’s patent application was filed on Nov. 23, 2016 but only published May 24, 2018, by the U.S. Patent and Trademark Office. Coincidentally, application was published just a week before General Motors settled a lawsuit with a motorcyclist involved in an accident with one of its autonomous cars.

The patent application describes how the system would focus specifically on the space between lanes and recognize when something is approaching through that region. A controller converts the imaging data to grayscale and then broken down into gradient levels, before comparing consecutive frames for any differences. A lane-splitting motorcycle would appear as cluster of pixels that would grow larger as it comes closer.

Figures 6A and 6B represent the images from two different frames reduced to grayscale. Figure 6C shows the difference between the two frames, with the white areas indicating an object (i.e. a lane-splitting motorcycle) appearing in the imaging area. This example shows what the system can detect when reducing the images to just three levels of grayscale gradients. The difference in contrast makes it easier for the controller to recognize a lane-splitting vehicle.

What happens after a lane-splitting motorcycle is detected depends on the situation and what other autonomous or safety systems are in use in the car. This may mean preventing a lane change until the motorcycle has passed, reducing the speed and performing lane changes slower or activating turn signals earlier before performing a lane change.

Michigan-based Delphi Technologies has its own patent application that would use data from a lane-splitting detector like Ford’s to autonomously move a car further away from the lane markers to give motorcycles more room while filtering.

In a separate application filed in 2016, Delphi Technologies’ patent would autonomously move cars over slightly in their lanes to give lane-splitting motorcycles more room.

As with all patent applications, we can’t predict if the ideas will ever wind up in real-life use. As the automotive industry delves deeper into autonomous vehicles, we hope automakers will follow Ford and Delphi’s lead and be mindful of motorcycles and lane-splitting.

Dennis Chung
Dennis Chung

Dennis has been a part of the Motorcycle.com team since 2008, and through his tenure, has developed a firm grasp of industry trends, and a solid sense of what's to come. A bloodhound when it comes to tracking information on new motorcycles, if there's a new model on the horizon, you'll probably hear about it from him first.

More by Dennis Chung

Comments
Join the conversation
 4 comments
Next