Bio-Inspired Hardware/Software Motion Sensor System

This work describes a combined hardware and software system that has been developed for imitating the vision of biological specimens. The hardware portion is a CMOS (Complementary Metal Oxide Semiconductor) foveated sensor system. The software portion attempts to mimic a part of the decision-making process using input from the hardware.;In the first chapter, we introduce the basic theory and background information. In the second chapter, we discuss the CMOS sensor system. Conventional flat 2D sensors suffer from a narrow viewing angle and thus are not well-suited for imitating the red eye tree frog's wide field of view (FOV). Thus, a novel method to make a pseudo-hemispherical surface with conventional sensor chips has been introduced. The CMOS image sensor is located at the center and the CMOS motion sensors are located above, below, left and right of the center chip to form a cross. They are tilted toward the focal point to make a pseudohemispherical surface. This structure can provide a wide view angle and the ability to track objects, and measure the size and speed of an object. The CMOS motion sensor had been fabricated through MOSIS (www.mosis.com) in a 0.5 micron technology and tested. The output of the chip is a set of pulse streams in which the altered duty cycle indicates the location of the object. During testing, the object was controlled by Newport step motor controller and manually. The motion sensor is able to detect both motions and returns altered duty cycle.;In the third chapter, we present the development of a novel sigmoid function for use in the decision making process. Since the red eye tree frog must decide to either deceive or survive in the same situation, the conventional sigmoid function can't be used for this project; it produces only one output. So I developed the "4-council men theory" that produces 4 opinions at a situation. It imitates human society . 4 council men never give up their opinions but can agree with one different topic. The hidden 4 opinions are collected and used for producing survival instinct, deception, escape and idle case of the red eye tree frog.;In Chapter 4, a novel software program developed using Visual Studio 2010 C;The final chapter summarizes the results and presents options for future work and future use of the system. This red eye tree frog artificial intelligence (A.I.) is able to produce a deceptive opinion when the predator approaches and activates an opinion of escaping when more than one predator shows up. This 4-council men theory can be applied to future collections of autonomous, robotic groups in order to help a robot to survive if it can make another robot collect incorrect information by deception. (Abstract shortened by UMI.)...