Vehicles, Remotely piloted

The design and validation of a low-level backstepping controller for speed and
heading that is adaptive in speed for a twin-hulled underactuated unmanned surface
vessel is presented. Consideration is given to the autonomous launch and recovery of an
underwater vehicle in the decision to pursue an adaptive control approach. Basic system
identification is conducted and numerical simulation of the vessel is developed and
validated. A speed and heading controller derived using the backstepping method and a
model reference adaptive controller are developed and ultimately compared through
experimental testing against a previously developed control law. Experimental tests show
that the adaptive speed control law outperforms the non-adaptive alternatives by as much
as 98% in some cases; however heading control is slightly sacrificed when using the
adaptive speed approach. It is found that the adaptive control law is the best alternative
when drag and mass properties of the vessel are time-varying and uncertain.

This project addresses the simulation, control and optimization of underwater vehicle performance. An analytical model of underwater vehicle motion has been developed. This model is based on a set of six degree of freedom nonlinear differential equations of motion. These equations incorporate inertial, hydrodynamic, hydrostatic, gravity and thruster forces to define the vehicle's motion. The forces are calculated and the equations of motion solved using a finite difference method of integration. An automatic closed loop control strategy has been developed and integrated into the motion model. The controller determines control plane deflection and thruster output based on sensor provided input, maneuver request and control gain constants. The motion model simulates the effects of these controller requests on the vehicle motion. The controller effects are analyzed and an optimal set of control gains is determined. These optimal gains are determined based on a quantitative comparison of a pre-defined Performance Index (PI) function. The PI is a function of critical performance values, i.e., energy consumption, and user defined weighted constants. By employing an iteration technique the PI is minimized to provide an optimal set of control gains.

A "hybrid" telerobotic simulation system that is suitable for telemanipulation rehearsal, operator training, human factors study and operator performance evaluation has been developed. The simulator also has the capabilities for eventual upgrade for supervisory control. It is capable of operation in the conventional rate-control, master/slave control and a data driven preprogrammed mode of operation. It has teach/playback capability which allows an operator to generate joint commands for real time teleoperation. For high-level task execution, the operator selects a specific task from a set of menu options and the simulator automatically generates the required joint commands. The simulator was developed using a three dimensional graphic model of an increasingly popular manipulator, TITAN 7F. A closed-form solution for inverse kinematics of the manipulator was found. Degeneracies from inverse kinematics solutions were observed to exist for certain arm configurations, although the manipulator can physically attain such configurations. An approach based on known facts about the manipulator geometry and physical constraints coupled with heuristics was used to generate physically attainable joint solutions from the inverse kinematics. The conditions that cause solution degeneracy were demonstrated to be related to singularity conditions. A novel object interaction detection strategy was implemented for more realistic telemanipulation. The object detection technique was developed based on the use of superellipsoid, which has a convenient inside-outside function for interference testing. The manipulator, with its end-effector and payloads, if any, were modeled as superquadric ellipsoids. A systematic way of determining transformation matrices between the superquadric manipulator links was developed. The interaction detection technique treats both moving and stationary objects in a consistent manner and has proved to be easy to implement and optimize for real-time applications. The feature has been applied for the simulation of pick-and-place operations and collision detection. It is also used to provide visual feedback as a low-cost force reflection and can be interfaced with a bilateral controller for force reflection simulation.