Jonathan Holmes and his fellow colleagues at GTRI developed an automated road repair system that can detect and fill in any cracks in the asphalt. The system could potentially replace a multi-man road crew with a single robot hitched to a truck, all while doing the job faster and more safely.
GTRI’s automated system first detects cracks in the road using a stereoscopic camera that takes two pictures of the road as it’s illuminated by an array of LEDs. The onboard computer analyzes the information to produce a “crack map” that depicts the size and location of all the cracks in the road.
According to the researchers, the computer compiles this data within 100 milliseconds after taking the first picture. Once the computer locates a crack, the computer signals a series of 12 nozzles to spray a sealant solution to fill in the crack.
In the road tests, the system proved to be effective enough to identify 83 percent of the cracks in the road. The system was also able to fill in lines of damage that were smaller than an eighth of an inch.
One downside to the system is that it operates at a maximum speed of three miles per hour (4.8 kilometers per hour), so while the system won’t be going anywhere fast, it could be less expensive and be deployed more often since it only requires a single human worker to drive the truck–although that too might change soon.
The researchers plan on improving their system by coding a more robust crack-detection algorithm that is not confused by oil stains, lane stripes, raised-pavement markers, tar, and other debris. The researchers also plan on building a full-scale prototype for use on a four-meter (13-foot) wide section of road that may incorporate a new 3D laser scanning system.
[Georgia Tech via Gizmag]
Like this? You might also enjoy…
- Celebrate Alan Turing’s Birthday and Check Out the Lego Turing Machine
- Sony and LL Cool J Rap It Up with Web-Based Music Collaboration Bundle
- Five Ways to Hack Your Video Game Controllers Into Something Awesome
Get more GeekTech: Twitter – Facebook – RSS | Tip us off | Follow Kevin Lee