Robotic speech recognition has made huge advances in recent years, allowing for easy voice interaction with robots. But robotic vision processing is still very rudimentary. Some robots "see," but they need powerful computers and are not very mobile. A team of researchers at the University of Arizona wants to change all this by mixing biology and electronics to create robotic vision.
The team has designed a visual navigation system by mimicking insect vision and demonstrated the concept by building a robot named Gimli. Instead of using standard microprocessors, the team devised electronic vision circuits based on a bunch of slower analog processors working in parallel.
The next step will be to develop a microchip-based vision system able to do specific tasks, such as following "a moving object like a soccer ball without getting confused by similarly shaped or colored objects." The team thinks the first such microchip might cost $30,000 to produce. But when the price goes down to $20, the market will be huge. This summary contains more details and pictures from Gimli.
The team has designed a visual navigation system by mimicking insect vision and demonstrated the concept by building a robot named Gimli. Instead of using standard microprocessors, the team devised electronic vision circuits based on a bunch of slower analog processors working in parallel.
The next step will be to develop a microchip-based vision system able to do specific tasks, such as following "a moving object like a soccer ball without getting confused by similarly shaped or colored objects." The team thinks the first such microchip might cost $30,000 to produce. But when the price goes down to $20, the market will be huge. This summary contains more details and pictures from Gimli.
| :: posted by Roland Piquepaille, 6:47 AM | Comments (0)
Links to this post |
0 Comments:
Post a Comment