Search Results

The development of automated guidance for agricultural tractors has addressed several basic and applied issues of agricultural equipment automation. Basic analyses have included the dynamics of steering systems and posture sensors for guidance. Applied issues have evaluated the potential of several commercial sensing systems and a commercial mechanical guidance system. A research platform has been developed based on a Case 7220 Magnum1 2-wheel drive agricultural tractor. An electrohydraulic steering system was used and characterized in support of automated guidance control. Posture sensing methods were developed using GPS, geomagnetic direction sensors (GDS), inertial, and machine vision sensing systems. Sensor fusion of GPS-inertial-machine vision and GPS-GDS-machine vision provided the most flexible and accurate guidance and capable for operation under dynamically changing field conditions.

This paper presents a programmable E/H control valve consisting of five individually proportional flow control valves. With a hybrid control algorithm, this valve has programmable valve characteristics, such as adjustable valve deadband and flow control gain, and programmable valve functions, such as different center functions. System analyses and experimental evaluations indicate that this programmable valve is capable of replacing conventional E/H control valves in practical applications.

Machine operators rely on intuition and experience to evaluate vehicle performance. As we increasingly turn to automation, it is important to automatically evaluate sensor data and system performance. Fuzzy logic allows us to take advantage of domain knowledge to evaluate data and to describe a system linguistically. In this paper, two automated fuzzy evaluation systems are described. In the first, a fuzzy quality module evaluates output from a simulated noisy sensor. In the second system, a fuzzy quality module evaluates the output from a machine vision system. Results from both systems indicate that fuzzy logic was able to accurately categorize the output in support of machinery decision making for automated control.

This paper presents system identification methods of using both frequency and time domain analyses to estimate system parameters for a nonlinear electrohydraulic (E/H) tractor steering system. The frequency domain method identified parameters of the system using a linear model with a nonlinear gain function. The time domain method identified those parameters using a discrete time expression. A method of multiple models was used to represent the nonlinear system for both frequency and time domain analyses. Simulation and test results showed that the modified identification method could satisfactorily identify the parameters of the nonlinear E/H tractor steering system.