Traffic Technology Today.com
Subscribe to Traffic Technology
Subscribe to Traffic Technology
   Sort by: relevance most recent
  


NEWS >>

University of Minnesota developing bicycle collision-prediction warning system

More than 48,000 bicycle-vehicle crashes occur each year in the USA, with a high proportion due to either the driver not seeing the cyclist or the cyclist behaving in an unpredictable manner. A project at the University of Minnesota’s (UoM) Roadway Safety Institute is developing a sensor system for a bicycle that can predict imminent bicycle-vehicle crashes and provide an audio warning of the bicycle’s presence to the motorist by sounding a loud horn.

The new bicycle safety system will also record a video camera image of the surroundings for a short duration after an imminent collision is predicted. Automakers have begun equipping many new vehicles with collision-prediction systems that warn motorists of possible collisions and help them take actions to avoid a crash, so the researchers are investigating the use of similar technology on bicycles to predict impending crashes and warn both riders and drivers.

The sensor system uses a novel sensor fusion algorithm and a combination of inexpensive sonar and laser sensors. The two types of sensing devices have complementary features that together can provide a robust reliable sensor system for imminent collision prediction. The warning system focuses on predicting two of the most common types of bicycle-vehicle crashes: rear-end collisions on urban roads in which the vehicle is approaching the cyclist from behind; and collisions at traffic intersections, where either the motorist is turning, or the cyclist is riding across the roadway.

The researchers have developed a collision-prediction algorithm and tested it in simulation studies. The team has also built an instrumented Schwinn bicycle, with sensors, electronics and a small computer for evaluation. Nine months into the project, researchers have tested the new warning system on the UoM campus, with preliminary experiments showing that the sensor suite on the bicycle can accurately estimate vehicle position and orientation for scenarios involving vehicles approaching cyclists from behind and vehicles turning right at intersections.

“This project is unique in many ways,” explained Rajesh Rajamani, professor of mechanical engineering at the UoM and lead investigator on the project. “Previous projects on vehicle tracking have been developed only for cars and have mostly focused on highway scenarios. This is perhaps the first research project where tracking of vehicles at a traffic intersection has been developed, and also the first project for a bicycle collision-warning system that goes beyond rear or straight-frontal collisions.

"If our final results demonstrate that the system works reliably and performs effectively, it could be commercialized into an inexpensive product sold to bicycle riders. In addition, the technology we’re developing has the potential to be extended to motorcycles and pedestrians in the future.”

June 22, 2015

Email


MAGAZINE >>

Read latest issueNEW DIGITAL EDITION:

The April/May 2017 issue of Traffic Technology International is now online.

Click here to read digital version
Click here to subscribe

Read now >>

VISION ZERO >>

Read latest issueNEW DIGITAL EDITION:

The January 2017 issue of Vision Zero International is now online.

Click here to read digital version
Click here to subscribe

Read now >>

INTERTRAFFIC WORLD >>

Read latest issueNEW DIGITAL EDITION:

Intertraffic World 2017 showcase is now online.


Click here to read digital version
Click here to subscribe

Read now >>


TOLLTRANS >>

Read latest issueNEW DIGITAL EDITION:

Tolltrans 2017 is now online.



Click here to read digital version
Click here to subscribe

Read now >>