The middle years of college-level engineering have been identified as an area that needs increased research attention. To date, little has been done to understand how students' learn conceptual knowledge in engineering science courses. This study examines how students use conceptual knowledge during the learning of control systems engineering. This work in progress will discuss the initial findings of a study examining junior-level mechanical engineering students solving controls homework and laboratory problems, exploring topics such as PID control, state feedback, and pole placement.

To accomplish successfully the complex future mission in civilian and military scenarios, robotic helicopters need to have controllers that exploit their full dynamic capabilities. The absence of high-fidelity simulation models has prevented the use of well established multivariable control techniques for the design of high-bandwidth full-flight-envelope control systems. Existing model-based controllers are of low bandwidth and cover only small portions of the vehicle's flight envelope. In this paper we present the results of the synergistic use of high-fidelity integrated modeling strategies, robust multivariable control techniques, and classical gain scheduling for the rapid and reliable design of high-bandwidth full-flight envelope controllers for robotic helicopters. We implemented and flight tested a gain-scheduled H/sub /spl infin// loop-shaping controller on the Carnegie Mellon University (CMU) Yamaha R-50 robotic helicopter. During the flight tests, the CMU R-50 flew several high-speed maneuvers. We believe that our modeling/control approach quickly delivers controllers that exploit the full dynamic capabilities of the airframe and thus are ready to be used by higher level navigation systems for complex autonomous missions.

Actuator arrays are planar arrangements of two degree-of-freedom actuators that cooperatively translate and orient objects for efficient manipulation. This paper derives the equations of motion for an object with the friction interface switching between slip and no-slip contact at each actuator. A controller derived from kinematics is presented and its stability analyzed with the multivariable circle criterion. Stability results are verified through simulation.

The modular distributed manipulator system is an array of actuators that is capable of manipulating objects in the plane. Each actuator in the array consists of two closely spaced, orthogonally mounted motorized roller wheels whose combined action approximates a programmable vector force. The paper first derives the equations of motion for transported objects by explicitly considering the traction force from each wheel of the actuator. This contrasts with prior work, which assumed an ideal single point of contact for each actuator. A translational velocity field is developed based on these dynamics and is compared with a discrete elliptic field derived. Errors in translational dynamics under an elliptic velocity field are discussed. The paper then presents simulation results that demonstrate some of the difficulties encountered when using non-ideal actuators.

An actuator array performs distributed manipulation where an object being transported and manipulated rests on a large number of stationary supporting actuators. The authors have developed a macroscopic actuator array consisting of many motorized wheels. As opposed to a MEMS array, the analysis requires the explicit modeling of the discreteness in the system, including the set of supports, distribution of weight, and generation of traction forces. Using an open-loop wheel velocity field, discreteness causes undesirable behavior such as unstable rotational equilibria, suggesting the use of object feedback. Discrete distributed control algorithms are derived by inverting the dynamics of manipulation. These algorithms reduce the many-input-three-output control problem to a three-input-three-output control problem.

Provides closed-form equations parameterizing the C/sup 2/ smooth path that globally minimizes the Euclidean norm of a robot's peak base reaction force while avoiding obstacles during 3D maneuvers in a gravity-free environment. Also, describes a computationally efficient technique that leads to a path typically having a peak force within 5% of the optimal path. The equations used to define the robot's motion are formulated after mapping the initial configuration, final (or goal) Cartesian location, and obstacles into a new space, the center of mass (CM) space. This is a Cartesian-like space that allows direct application of many existing control techniques, such as resolved rate control. In the CM space, a series of path segments guide the robot around the obstacles. Solving a system of equations based on these segments for boundary condition dependent constants determines the path. Currently, closed-form equations are unavailable for the boundary dependent constants, preventing exact determination of the globally optimal path. This paper introduces a procedure for locating the optimal path. Its final step uses sequential quadratic programming to locate boundary dependent constants. The equation formulations assume that the initial configuration of the robot is known and that the robot mass and obstacle positions are constant during the maneuver. The method developed has direct applicability to redundant and nonredundant robots. A detailed example, based on a nonredundant robot avoiding a single obstacle, illustrates the concepts presented.

The modular distributed manipulator system (MDMS) is a macroscopic actuator array which can manipulate objects in the plane. The piecewise-constant dynamics of manipulation on the MDMS are developed based on an exact discrete representation of the system. The resulting dynamics are inverted enabling the calculation of an open-loop vectorfield which provides arbitrary uniform object dynamics. The vector field positions, and under certain assumptions, orients objects.

An actuator array is a form of distributed manipulation where an object being transported and manipulated rests on a large number of supporting actuators. On a discrete array employing an open-loop field, some size and shape objects may have unstable equilibria due to this discreteness. The functional relationship between object dimensions relative to array spacing and rotational stability is examined, and a map of this function is generated. The geometry behind this relationship is also examined, and analytical expressions for the boundaries between stable and unstable regions in the map are derived.

We are developing a materials handling system where many small simple actuators cooperate to transport and to manipulate large objects in the plane. A discrete set of cells, each comprising two actuators, are fixed in a planar array. By coordinating the actuators in the cells on which an object rests, an object can be transported and manipulated. In essence, this system is an improvement over traditional conveyor systems in that objects can be re-oriented, as well as conveyed. Such an array provides flexible materials handling in which many objects independently can be manipulated and transported at the same time. The array is coordinated in a distributed manner where each cell has its own controller and each controller communicates with its neighbors. Towards the goal of motion planning, in this paper we consider the dynamics of parcel transport and manipulation. The parcel dynamics are based on an exact discrete representation of the system, unlike other methods where a continuity assumption is made. Two types of contact models are considered.

A method for nonlinear function identification and its application to learning control are presented. The control objective is to identify and compensate for a nonlinear disturbance function. The nonlinear disturbance function is represented as an integral of a predefined kernel function multiplied by an unknown influence function. Sufficient conditions for the existence of such a representation are provided. The learning rule indirectly estimates the unknown function by updating an influence function estimate. It is shown that the controller achieves the disturbance cancellation asymptotically. The method is extended to the repetitive control of robot manipulators. Simulation and actual real-time implementation results using the Berkeley/NSK robot arm show that the proposed learning is more robust and converges at a faster rate than conventional repetitive controllers.>