Bikes could soon use AI to think for themselves, Shimano patent suggests

New invention shows automatic seatpost, suspension and saddle adjustments made through machine learning

Patent imagery of a bike and AI machine learning devices
(Image credit: USPO / Shimano)

Our bikes could soon use machine learning to think and make decisions for themselves, a new Shimano patent suggests. 

Published earlier this week, the patent for a “learning model” details a product that draws on artificial intelligence (AI) to control dropper seatposts, front suspension and saddle tilt. 

The device, a small screen attached to the handlebar, is designed to learn from a rider’s preferences in different situations, in order to wirelessly automate changes itself mid-ride through “telescopic mechanisms”.

The rider then gives feedback on the changes through Facebook-esque “like” and “dislike” buttons, informing the device’s learning, and helping it perform better. 

The image attached to the patent suggests the product is for a gravel bike, but could also be used on a mountain bike. Some features, such as adjustable saddle tilt and dropper posts - the later famously used by Matej Mohorič at Milan-San Remo  - would be applicable on the road. 

According to Shimano, the automated device will help improve riding comfort and efficiency. 

The patent, filed under US 11866114 B2, reads: “Proper skill is required for the rider to appropriately control the seat post and the suspension during riding. It would be desirable to provide an optimum automatic control that is based on a traveling condition, a road surface condition, and/or a traveling purpose. 

“It is an object of the present invention to provide a component control device for optimally performing automatic control of a component having a telescopic mechanism, a method of creating a learning model, a learning model, and a computer-readable storage medium.”

Automatically controlled dropper seatposts and suspension products already exist on the market; however, the idea that a rider can give feedback and train the programming is novel. 

A patent drawing with an interface showing that courses have been learned by a device

(Image credit: USPO / Shimano)

Shimano’s proposed product also appears able to learn courses. A mock-up of the device interface gives the examples “Olympic Course” and “ABC Downhill”, presumably mountain bike courses, that have been learnt with the rider’s preferences. 

This suggests a rider could input their preferences on a recon lap, accounting for changes in track surface, then benefit from seamless adjustments in a race situation. That said, the technology may be too heavy to race with, as it features a large battery attached to the frame. 

As with all patents, it is uncertain how far Shimano have gone in developing the product, or if it will see the light of day at all. 

Cycling Weekly has previously been told that the Japanese parts manufacturer “never speculates on future product developments”, so we will have to wait and see. 

Thank you for reading 20 articles this month* Join now for unlimited access

Enjoy your first month for just £1 / $1 / €1

*Read 5 free articles per month without a subscription

Join now for unlimited access

Try first month for just £1 / $1 / €1