Announcing a very special event for the Computer Society!
Please join Dr. Zhen Zhu as he discusses his insights into using AI for navigation.
Location:
DeVry University
Room 8
1350 Alum Creek Dr
Columbus, OH 43209
Univeristy Directions
Google Directions
Refreshments will be provided, ample free parking available.
Please RSVP by clicking the link below.
Abstract:
Navigation and machine intelligence are two ideas that don't easily blend. Navigation is about precisely estimating the position, velocity and time in an efficient manner, which traditionally is not of major interest to the artificial intelligence research society. However, certain techniques that were originally developed for machine intelligence have been adopted to facilitate automatic navigation. For example, computer vision has been widely applied to guide robots and unattended vehicles. These machines, on the other hand, can serve as the host for embodied intelligence once they are equipped with the capabilities automatic navigation and guidance.
This work discusses the fusion of computer vision with navigation systems. Human vision is naturally one of the earliest ways to navigate in history. Most modern navigation techniques, such as absolute positioning using radio navigation including GPS, or dead reckoning with inertial navigation sensors, do not rely on vision. However, the unique advantages offered by the optical sensors urged the use of them in the navigation systems. One of the most used optical sensors is, not surprisingly, the camera. Image processing skills developed for computer vision has greatly contributed to the success of camera-based navigation techniques, although these techniques are often specialized for relative navigation in a local neighborhood. Meanwhile, despite the obvious advantage of GPS in providing absolute positioning, its availability is often limited when indoors or in urban canyons. Integration of GPS with cameras would lift the restriction on both.
Tight Optical Integration (TOI), a navigation algorithm that fuses GPS with computer vision, was recently proposed. When vehicles traveling in urban canyons do not have inadequate GPS satellites due to signal blockage or denial, TOI performs a tight integration of visual and GPS measurements, and maintains the capability of absolute positioning as a result. When the GPS constellation becomes completely blocked, for instance, in an indoor environment, TOI can automatically transit to operate with relative navigation in reference to the last known absolute position.
Speaker Bio:
Zhen Zhu received the Ph.D. degree in Electrical Engineering from Ohio University, Athens, Ohio in 2006. Currently he is a Senior Research Engineer with the Ohio University's Avionics Engineering Center. He is also an adjunct assistant professor at the School of Electrical Engineering and Computer Science of Ohio University. His research interests include a wide variety of topics in navigation and artificial intelligence. In the past eight years, he has worked on GPS signal processing and receiver design, the Local Area Augmentation System (LAAS) and the Wide Area Augmentation System (WAAS), software radio technology, urban navigation, optical signal processing, navigation sensor integration and embedded systems. He has also published on artificial neural networks, pattern recognition and self-organizing systems. Much of his recent research effort is devoted to the application of computer vision and machine intelligence in the navigation and guidance for autonomously operating vehicles.
