Waymo, Google’s AV-focused arm created in 2009, is thinking seriously about bikes. Nathaniel Fairfield, Waymo’s principal software engineer, has been collecting and tracking data on cyclists, to help AVs predict their road movements. In addition to these results, Waymo programmed their cars to pass bikes in accordance with state laws. The predictability of cyclists and pedestrians is one of the bigger challenges that AV creators are facing.
According to a survey conducted by a cycling and pedestrian non-profit – ‘Bike Pittsburgh’, cyclists felt safer around AVs than around human drivers. However, according to Prof. Anthony Rowe, AVs still need some additional help to detect cyclists. Bikes are not as predictable as cars. They can act like cars on the side of the road, then change and act like pedestrians walking on the sidewalk.
Rowe and his team are developing bike instruments that provide cars with information to predict cyclist movement, to avoid collisions. The instruments will eventually be embedded in a mobile phone on the front of the bike, once the program is more developed. Rowe’s bike, pictured above, is outfitted with a range of sensors that are helping his team to learn more about cyclists behavior and movements.
The post was written by SCI Fellow Ramy Barhouche.