Autonomous robot: An image and signal processing approach

In robotics, the most challenging tasks are how the machine perceives and interacts with its environment. Our research uses a wireless robot, that provides the mobility, video, audio, and sensor capability needed to control it. Images are captured and the colors are differentiated to locate objects, which drastically increases the response time and comprehension of the robot. Scanning entire images for patterns requires substantial computations, which is not an optimal solution. Human eyes process information in great detail only from the cone of vision and the rest is a perpetually blurry area in the peripheral vision. Building a robot's vision similar to an eye is more realistic with less overhead than the traditional methods. In this paper, we demonstrate a method which locates an object first by color and then narrows the results by comparing other properties. The robot will then maneuver towards it until a set depth is reached. This along with many other criteria are integrated together to achieve a type of low-level "artificial-intelligence." This novel approach is used to navigate an autonomous robot that will not collide with any other obstacles. Further applications can now efficiently be built based on the design already applied to the robot.