Abstract: A surface grading system includes a traction unit and a scraper implement attached thereto via a three-point hitch system. A controller automatically controls an actuator of the three-point hitch system to automatically adjust a position of the scraper implement relative to the traction unit to automatically profile the ground surface to achieve a desired profile. The controller includes an Artificial Neural Network that receives a pitch angle of the traction unit and a leveling variation as input values, and outputs a hitch control signal. The Artificial Neural Network includes a first layer using a logarithmic sigmoid function as an activation function, and an output layer using a linear function as an activation function. The hitch control signal raises or lowers the three-point hitch system as the traction unit moves across the ground surface.

Abstract: A surface grading system includes a traction unit and a scraper implement attached thereto via a three-point hitch system. A controller automatically controls an actuator of the three-point hitch system to automatically adjust a position of the scraper implement relative to the traction unit to automatically profile the ground surface to achieve a desired profile. The controller includes an Artificial Neural Network that receives a pitch angle of the traction unit and a leveling variation as input values, and outputs a hitch control signal. The Artificial Neural Network includes a first layer using a logarithmic sigmoid function as an activation function, and an output layer using a linear function as an activation function. The hitch control signal raises or lowers the three-point hitch system as the traction unit moves across the ground surface.