John F. Gardner

Kinematics of redundantly actuated closed chains

The instantaneous kinematics of a hybrid manipulation system, which combines the traditional serial chain geometry with parallelism in actuation, and the problem of coordination is discussed. The indeterminacy and singularities in the inverse kinematics and statics equations and measures of kinematic performance are analyzed. Finally, coordination algorithms that maintain an optimal force distribution between the actuators while avoiding or exploiting singularities are presented. >

Efficient computation of force distributions for walking machines on rough terrain

The force distribution problem for legged vehicles on rough terrain is considered. A general formulation of the force distribution problem in which the feet contact the ground at arbitrary inclinations, is presented. Three techniques are used to solve the force distribution problem for three representative tasks. The Moore-Penrose pseudo-inverse, a numerical optimization scheme and an approximation to the optimal solution are described. The optimal scheme computes the forces which minimize the maximum ratio of tangential foot reaction force to foot normal force. The approximation is used to achieve certain desirable characteristics of the optimal scheme with considerably less computational resources.

Dynamic models of friction wedge dampers

Friction wedges play a central role in the vertical dynamics of railroad freight cars. They also play a role in lateral dynamics and stability. While friction wedges are mechanically very simple, their inherent nonlinearities lead to complications in modeling and dynamics. In this paper we present two simplified models of the fundamental physics of friction wedge dampers in the geometry typical of a three-piece truck. We also review the friction wedge model used in NUCARS and compare the three.

Force Distribution in Walking Machines Over Rough Terrain

This paper presents a computationally efficient technique for the solution of the force distribution problem in walking machines. It differs from previous techniques in two important respects. First, the formulation of the problem allows for arbitrarily oriented surface normals at the point of contact between the feet and the ground. This is an important extension since the primary purpose of legged vehicles is locomotion on rough terrain. Second, the solution technique allows for the introduction of nonlinear constraints which can be tailored to achieve secondary goals in system performance. An example is presented which is based on the geometry of the Ohio State University Adaptive Suspension Vehicle which indicates that the technique performs favorably when compared to pseudo-inverse and computationally intensive optimization methods.

Two projects for undergraduate mechatronics class: success and failure

The vast majority of mechatronics courses currently offered in Mechanical Engineering programs across the country have a substantial project experiences as a central part of the course. The proper design of the project experience is crucial for the success of the educational experience. In this paper, two particular mechatronics projects that were actually used in a senior-level elective mechatronics course at Penn State will be examined. While they are very similar in basic design, they were extremely different in practice: one being a huge success, the other a dismal failure. The features that contributed to the success and failure of the projects will be addressed as will the overall structure of Penn States mechatronics program so that the paper can be given the proper perspective.

Applications of neural networks for coordinate transformations in robotics

The use of artificial neural networks is investigated for application to trajectory control problems in robotics. The relative merits of position versus velocity control is considered and a control scheme is proposed in which neural networks are used as static maps (trained off-line) to compute the inverse of the manipulator Jacobian matrix. A proof of the stability of this approach is offered, assuming bounded errors in the static map. A representative two-link robot is investigated using an artificial neural network which has been trained to compute the components of the inverse of the Jacobian matrix. The controller is implemented in the laboratory and its performance compared to a similar controller with the analytical inverse Jacobian matrix.

A Solution for the Force Distribution Problem in Redundantly Actuated Closed Kinematic Chains

Proper control of robotic systems which incorporate closed kinematic chains is important in many applications. Among these are the multi-robot work cell and legged vehicles. In these no unique solution exists for the force distribution corresponding to a specified trajectory. A framework within which additional constraint equations may be written is presented here, and the force distribution solved in closed form for a walking machine application. These constraints are related to system performance goals of interest, such as improved traction and/or load sharing among the legs. The proposed technique is shown to be computationally simpler than other alternative solutions to the same problem.

Characteristics and approximations of optimal force distributions in walking machines on rough terrain

The force distribution problem for legged vehicles on rough terrain is considered. A general formulation of the force distribution problem, In which the feet contact the ground at arbitrary inclinations is presented. Three techniques are used to solve the force distribution problem for two representative tasks. The Moore-Penrose pseudo-inverse, a general numerical optimization scheme and an approximation to the optimal solution are described. The optimal scheme computes the forces which minimize the maximum ratio of tangential foot reaction force to foot normal force. The approximation is used to achieve certain desirable characteristics of the optimal scheme with considerably less computational complexity.<<ETX>>

Mechatronics II: advanced mechatronics for mechanical engineering students

Increasingly, the field of mechatronics is seen as an essential background of a practicing mechanical engineer. A course in advanced mechatronics is described which is offered as a joint graduate-undergraduate course and is intended to be a second course in mechatronics. As opposed to many mechatronics courses, this course focuses more on intelligent product design than on DSP or controls applications. Another unique feature of the course is a class-wide project experience in which 10 separate teams work together on various subsystems of a complex robotic rover vehicle. The class was offered for the first time in the Spring of 1999 and 20 students, both graduate and undergraduate participated.

Experimental Results for Force Distribution in Cooperating Manipulator Systems Using Local Joint Control

Local control schemes using only position and rate errors to generate control forces are widely used for control of open- chain, serial-link robotic mechanisms, When two or more such open chains interact, closed kinematic chain, redundantly actu ated mechanisms are formed. Recent work has shown that the vector of joint forces produced using a local proportional-plus- derivative feedback scheme for the control of a cooperating manipulator system results in a vector of joint torques with a minimum weighted Euclidean norm. The current work presents the derivation of this force distribution and experimental evi dence to corroborate the analytic results. The data presented are obtained from an experimental cooperating manipulator system developed specifically for use in the application of theo retical control approaches in cooperating hardware systems.