Human Perceptron Method to Design Robot Shapes and Develop Architecture for 3D Camera Device
Abstract:
In this paper, we study light weight material plastics. We study different shapes. They include connector, frame, ring, triangle shape, star and wheel. The shapes when assembled form the manual controlled robot. We de velop human perceptron method to collect data of the dimensions, color and shape features of the robot. We include the method to develop 3D computer aided design (CAD) model. The structural similarity index (SSIM) is 0.42 giving good accuracy to actual object. We develop correlation model between the dimension and pixel. The pixel conversion factor is 19.2 pixels/mm. The calculated pixels for overall connector length of 50 mm are 960 pixels. We understand the relation between meter scale shape component of the robot to the pixel for the f irst time. We design an integrated 3D camera system. We use the neural network hardware rules to obtain the weight of individual connector pair segment to be used for training the segment movement. The segment neural network moves individually in x, y and z directions. The laser distance meter is used to record and store in excel sheet the x, y, z movement. The distance meter has Bluetooth enabled. The integrated system has laptop with python to visualize the given object. We visualize 3D image. We obtain high resolution camera image compa rable to actual object. The weight for the connector pair segment in the neural network hardware predicts the new x, y and z, respectively. The advantages of lightweight materials are they can be frequently assembling or disassembled. There is no need to use physical wiring concept. It is easy to scale the shape and complete robot from millimetre to meter.
