Test Bed for Demonstrating and Teaching Soft Sensor Concepts

Abstract

The smart sensor car is a test bed for demonstrating soft sensor concepts. The smart car follows the magnetic field coming from a wire track. Sensors on the smart car detect the magnetic field and generate signals. Those signals are conditioned and converted to digital numbers, which are used by the neural network as inputs. The neural network calculates the car position from these inputs. The car position is sent to a controller that calculates the car steering angle. The commands from the controller drive the smart sensor car around the track, where the sensors generate different signals resulting in different commands from the controller. The neural network is implemented on an Field Programmable Gate Array (FPGA) in a serial configuration.