Systems comprised of many subsystems can benefit from reducing the number of inputs. Pin array shape displays are one example where this is particularly true. Many pin arrays use a row-column structure, as found on LCD screens, to reduce the number of inputs. However, reducing inputs in this manner makes simultaneous control of the pins challenging. Using semi-nonnegative matrix factorization (SNMF) to reduce the rank of the inputs solves this problem, but there has not yet been a method to analyze the stability of the resulting system. Additionally, the application of the SNMF System to pin arrays, particularly those using fluid power, presents many practical challenges. This paper proves that the SNMF rank-one approximation is strictly-output passive, leading to approaches to analyze the stability of the SNMF System. We also address practical challenges by showing that the SNMF System improves performance over the previous approach, line scanning, even in the presence of power limitations and by examining the effects of row and column dynamics and proposing approaches to limit these negative effects.

How do animals understand the physical world they live in? One answer, due to Gibson, is that their sensory systems are tuned to pick up relevant affordances for behavior, but how is it that the brain and the sensory apparatus become suitably adapted to perform this feat? To cast light on this question we have been investigating active touch sensing in mammals, including humans, and developing biomimetic robots that can help us understand these biological systems whilst also developing useful haptic technologies. An important focus has been on the vibrissal (whisker) system of rodents, and its emergence through evolution and development, which we have investigated through a combination of (i) ethological studies of behaving animals, (ii) computational neuroscience models of the neural circuits involved in vibrissal processing, and (iii) biomimetic robots embodying many of the characteristics of whiskered animals in their design and control. This work has resulted in a series of whiskered robots, the most recent of which, Shrewbot, is able to construct tactile maps of its environment and recognize and track moving objects. We are also studying humanoid touch, focusing on the development of Bayesian strategies for active tactile sensing with robot hands. Here our results have provided the first demonstration of hyperacuity in robot touch whilst also indicating that tactile perception is improved in unstructured environments by appropriate active control. The active sensing framework can also be applied to the development of haptic interfaces for human users that can augment our existing sensory capability. For instance, we are developing a head-mounted “remote touch” system that links distance sensors (ultrasound arrays) with vibrotactile displays. Here an interesting question is how the signals that are delivered through the displays should be modulated to take into account the intentional head and body movements of the user and in order to provide a meaningful and intuitive experience. The talk will present converging lines of evidence, from these different research strands, for the importance of active control in haptics. Our results will also be used to illustrate how experimental, computational, and robotic approaches can operate together to advance our understanding of sensorimotor cognition in behaving systems.

The need for knowing future manipulator motion arises in several robotics applications, including notification or avoidance of imminent collisions and real-time optimization of velocity commands. This paper presents a real-time, low overhead algorithm for identification of future manipulator motions, based on measurements of prior motions and the instantaneous sensed actuator velocity commanded by an operator. Experimental results with a human-controlled, two degree-of-freedom manipulator demonstrate the ability to quickly learn and accurately estimate future manipulator motions.

This research investigates and seeks to mitigate the undesirable effects of biodynamic feedthrough in backhoe operation. Biodynamic feedthrough occurs when motion of the controlled machine excites motion of the human operator, which is fed back into the control input device. This unwanted input can cause significant performance degradation, which can include limit cycles or even instability. Backhoe user interface designers indicate that this is a problem in many conventional machines, and it has also proved to degrade performance in this testbed. A particular backhoe control system, including the biodynamic feedthrough, is modeled and simulated. Cab vibration control is selected as a means to mitigate the biodynamic feedthrough effect. Two controller based methods are developed based on these models and presented, both of which use the working implement itself to reduce the cab motion. In this case, the backhoe arm has dual functionality, to perform excavation operations and to cancel cab vibration. Results show that significant reductions in cab motion can be obtained with minimal tracking performance degradation, without additional actuators.

This paper describes a novel auto-calibration state-trajectory-based control method and its application to electronic flow control for independent metering systems. In this paper, the independent metering architecture that is considered uses five Electro-Hydraulic Poppet Valves (EHPV's). The proposed control method is applied to four of these valves, arranged in a Wheatstone bridge configuration, to regulate the flow of hydraulic oil coming into and out of an actuator. For simplicity, the fifth valve is operated via open-loop to control the supply pressure. Experimental data presented herein demonstrate that the control method learns the valve's conductance characteristics (i.e. the inverse input-state dynamic map of the valve) while simultaneously controlling the motion of the hydraulic actuator.

Flexibility is often an unavoidable consequence of the desire for high speed and performance manipulators. This paper proposes a method that improves the performance of flexible manipulators through the employment of robust state estimation techniques. These techniques are based on discrete time Kalman filtering and sliding mode principles. A simple model for a single degree of freedom flexible manipulator is derived and a control scheme is chosen and implemented. The latter includes a robust non-linear estimator. Simulation and preliminary experimental results are presented that demonstrate the validity of the proposed control scheme.

In this paper, we present a control theoretic formulation for composing human control inputs with an automatic controller for shared control of a quadruped rescue robot. The formulation utilizes a model predictive controller to guide human controlled leg positions to satisfy state constraints that correspond to static stability for the robot. A hybrid control architecture that incorporates the model predictive controller is developed to implement a gait that guarantees stable foot-placements for the robot. The algorithm is applied to a simulation of a quadruped rescue robot with human input provided through haptic joysticks.

Hydraulic manipulators are commonly used for tasks that require large forces over a large workspace. Current hydraulic manipulators are often operated using individual joint control. Since many of these manipulators are long kinematic chains, such as excavators and concrete pump trucks, it can take years of training and experience to become a true expert operator, resulting in a large number of novice operators. Haptics and coordinated control have been shown to improve various operator performance measures such as task completion time and the amount of applied force. This paper outlines a novel coordinated control scheme and shows that coordinated position control is feasible for slow and/or large-workspace manipulators. It discusses the results of these two coordinated control modes with and without haptics on operator task completion time. The performance of a single expert operator is provided as a comparison. This paper also includes a fuel efficiency measurement absent in previous work. The results are benchmarked against current joystick control.

Haptic devices allow a human user to physically interact with the virtual or simulated world by generating real forces to represent interactions in the virtual world. By definition, haptic devices exist in the set of robotic systems whose workspaces must overlap with the workspace of the human. For that reason, safety of the human becomes a central concern when developing both the physical hardware and the control systems for new and novel interfaces. Passive haptic devices, ones that display forces to the user by using passive actuators such as brakes, clutches of continuously variable transmissions, address the concerns of stability and safety by using actuators that physically guarantee a bounded output and the safety of the human user. However, a device using passive actuators can only produce a subset of the forces generated by a similar device built around active actuators. The work presented here will explore the use of a passive haptic device specifically with respect to teleoperation and will present a new algorithm for determining feedback that improves the transparency, a dynamic compensating controller. This controller will be implemented in LabVIEW on passive hardware and will be compared to two simpler control schemes in two typical teleoperation tasks, shape identification and point to point motion around an obstacle. The results will show that the passive device provides feedback that allows the user to recognize large surface features. Furthermore, results will show that the dynamic compensating controller proves especially useful in dynamic situations represented by the motion task.

The haptic interaction with machines occupies orders of magnitude in size and power. Some of the most powerful machines are hydraulic and can benefit from the natural communication channel of haptics. Hydraulics can also provide interactions on a human scale with fine spatial resolution. In this talk, these extremes are addressed and the challenges and solutions provided by hydraulics and haptics are explored.

This paper investigates the effects of increasing the haptic device's physical dissipation on haptic two-port network (H2PN) performance. Two methods for increasing device dissipation are first explored through numerical H2PN models. Specifically, I/A (impedance/admittance) and A/I (admittance/impedance) H2PN's with low device dissipation, increased device physical dissipation, and electrical DC motor EMF dissipation are numerically compared. Second, I/A and A/I H2PN's with and without increased physical device dissipation are experimentally validated on HuRBiRT, a large scale haptic device with a hybrid active/passive actuated axis

This work investigates the modeling and control of a new dissipative passive haptic display that uses magnetorheological (MR) brakes as actuators. The device is energetically passive, meaning that it can only remove energy from the system. All motive force must be generated by the user, which guarantees stability of the system and safety of the human operator. A first order system approach is presented as a tool for modeling MR fluid behavior in a low-speed braking device. A simulation was developed that uses the brake model and the equations of motion to predict robot motion based on force input and actuator commands. The accuracy of these models was demonstrated with comparisons to experimental data. Two forms of path following velocity control were successfully implemented and shown to significantly reduce path error in preliminary experiments with human operators.

Experimentation and a survey of the literature clearly show that contact stability in a force reflecting teleoperation system requires high levels of damping on the master robot. However, excessive damping increases the energy required by an operator for commanding motion. The objective of this paper is to describe a new force reflecting teleoperation methodology that reduces operator energy requirements without sacrificing stability. We begin by describing a new approach to modeling and identifying the remote environment of the teleoperation system. We combine a conventional multi-input, multi-output recursive least squares (MIMO-RLS) system identification, identifying in real-time the remote environment impedance, with a discretized representation of the remote environment. This methodology generates a time-varying, position-dependent representation of the remote environment dynamics. Next, we adapt the target impedance of the master robot with respect to the dynamic model of the remote environment. The environment estimation and impedance adaptation are executed simultaneously and in real time. We demonstrate, through experimentation, that this approach significantly reduces the energy required by an operator to execute remote tasks while simultaneously providing sufficient damping to ensure contact stability.

The NSF Digital Clay project is focused on the design, prototyping, integration, and validation of a computer-controlled physical device capable of taking any of a wide range of possible shapes in response to changes in a digital 3D model or to changes in the pressure exercised upon it by human hands. Although it clearly is a natural and unavoidable evolution of 3D graphical user interfaces, its unprecedented capabilities constitute a major leap in technologies and paradigms for 3D display, for 3D input, and for collaborative 3D design. In this paper, we provide an overview of the Digital Clay project and discuss the challenges, design choices, and initial solutions for a new finger sculpting interface designed for the Digital Clay and prototyped using conventional 3D I/O hardware.

A laboratory robot with 3 m links is modeled using an assumed mode method. The model is compared to experiments and to several theoretical limiting cases. The most demanding comparison is the comparison of frequency response results for joint motion, link deflection and tip deflection. We conclude that with care the models of an arm can be effective in controller design in spite of hydraulic actuators and non-ideal kinematics and structure.

Since its introduction, the command shaping method has been applied to the control of many different types of flexible manipulators and the effectiveness in the vibration suppression has been verified. However designing an effective command shaper requires a priori knowledge about the system parameters. Previously, some efforts have been made to make the command shaper adaptive to the changes in the system parameters. In this paper we introduce the generalized time-delay command shaper and analyze its characteristics. Also we show that previously proposed command shapers are particular cases of the generalized command shaper. Then, direct adaptation of the generalized command shaper is studied. Finally, the experimental results of the adaptation are included to show the effectiveness of the direct approach using the generalized command shaper and to support the analysis result.

A multi degree of freedom manipulator can be commanded to generate base forces by utilization of the dynamic equations relating joint torques or joint accelerations to base forces. This can be extended to a full order case with three base moments and three base forces. In the experiments described two forces were sufficient to damp the fundamental modes. With a force command capability a variety of damping algorithms can be used to determine the desired force. A simple acceleration feedback algorithm was used which approximates a two degree of freedom dynamic vibration absorber where the damping coefficient can be adjusted by the acceleration feedback gain. The resulting damping ratio of the fundamental mode was increased by factors of between 18 and 191.

During the 1995 DOE Robotics Forum at Albuquerque, New Mexico, Georgia Tech's Intelligent Machine Dynamics Laboratory demonstrated remote display capabilities of its telemanipulated long-reach manipulator testbed through the Internet. This paper describes the testbed and addresses the methods, limitations, and capabilities of remote operations.

This work investigates issues related to vibration control in a micro/macro manipulator. This paper provides experimental results obtained by the combination of two dissimilar flexible control techniques to a micro/macro flexible link testbed. Inertial damping and command filtering techniques are implemented simultaneously to form a robust controller that results in minimal residual vibration due to commanded movements or external excitations. Experimental results show the effectiveness of both control techniques in their individual state as well as the improved performance resulting from their combination. The experimental results of the combined controller clearly show the advantages of each technique.

A single-link flexible manipulator with a rotary actuator at one end and a mass at the other is modeled using the Lagrangian method coupled with an assumed modes vibration model. A SIMO state space model is developed by linearizing the equations of motion and simplified by neglecting natural damping. Laplace domain pole-zero plots between torque input and tip position show nonminimum phase behavior. Nonminimum phase behavior causes difficulty for both conventional and artificial neural network (ANN) inverse-model control. The most promising ANN method for the control of flexible manipulators does not appear to converge to a solution when the system is lightly damped. To overcome this limitation, a modified cost function is proposed. Simulations show that the ANN is able to converge to a solution even in the case of no damping. The modified approach fails, however, for beams exceeding some critical length measure. Identification of the critical length and proposals for extending the result are discussed.

Distributed flexibility of the links is a severe obstacle for the endpoint position control of lightweight manipulators. In order to accomplish with satisfactory performance certain tasks involving a controlled interaction of the tip of the robot with the worksurfaces, a combined control of the motion and the contact forces can provide some advantages. This paper presents a general and systematic model of a flexible robot interacting with a rigid environment. A force/position control scheme based on this model is also introduced. Results obtained simulating the constrained motion of an existing 2-DOF flexible arm are given.

Impedance control is a popular method of controlling the dynamic response a robot has to external forces. The advantage of such a control paradigm is the ability to manipulate in both unconstrained and constrained environments. Unlike hybrid control methods that attempt to control forces and motions in orthogonal directions, impedance control consists of a single control law that accommodates external forces. One problem associated with such a controller is the selection of the robot's target impedance. This paper illustrates, through analysis and experimentation, the shortcoming of improper target impedance selection. The integration of a real time estimate of the robot's environment with the impedance controller resolves this problem. Experimental results illustrate improved performance during the transition from unconstrained to constrained manipulation.

In this paper, the authors examine optimal path planning for the motion of a wheel which minimizes some combination of time and effort expended. Several techniques are examined but a combination of continuation and multi-dimensional Newton-Raphson shooting is shown to be effective, within limits.>

This paper proposes a bracing strategy for micro/macro manipulators. The bracing micro/macro manipulator can provide advantages in accurate positioning, large work-space, and contact-task capability. However, in exchange for improvement in performance, one must accept the complex control problem along with the complex dynamics. This research develops a control scheme for a bracing manipulator which makes multiple contacts with the environment. Experimental results show the feasibility of the proposed ideas for real world applications.>

The state-of-the-art is summarized, focusing on interdisciplinary approaches and positions, and future directions in the design, analysis, and control of lightweight robotic and telerobotic motion systems for space application are discussed. The emphasis is on providing a logical connection between the special demands of space applications and the design of the motion system. Flexibility is presented as a natural consequence of these demands. A number of technologies are relevant to extending feasible performance into regions of the design space previously avoided due to the resulting flexibility of the structures and drives. Control technology is considered foremost, but passive damping, structural materials, structural design, operational strategy and sensor technology are closely related. Numerous references are presented for those wishing to employ these technologies.>

The performance limitations of manipulators under joint variable-feedback control are studied as a function of the mechanical flexibility inherent in the manipulator structure. A finite-dimensional time-domain dynamic model of a two-link, two-joint planar manipulator is used in the study. Emphasis is placed on determining the limitations of control algorithms that use only joint variable-feedback information in calculations of control decisions, since most motion control systems in practice are of this kind. Both fine and gross motion cases are studied. Results for fine motion agree well with previously reported results in the literature and are also helpful in explaining the performance limitations in fast gross motions.>

Modeling results are presented for a RALF (robotic arm, large and flexible) consisting of two ten-foot-long links and an actuator link that is parallel to the first link and drives the second link. A reduced-order model of the RALF system, which consists of the first two component modes of each link, was derived using the modal cost analysis approach. The original models are based on the first five component modes of each link. Two sets of mode shapes, which result from different boundary conditions on the components (links) are considered. The reduced-order model is optimal from the control point of view. The reduced-order model is evaluated by a comparison between the frequency responses and the modal cost analysis results. From the component cost calculation results, the modeling error ratio was found to have little variation, in spite of the change in angles of the joints.>

This paper presents a new systematic algorithm to symbolically derive the full nonlinear dynamic equations of motion of multi-link flexible manipulators. Lagrange's-Assumed modes method is the basis of the new algorithm and adapted in a way suitable for symbolic manipulation by digital computers. It is applied to model a two-link flexible arm via a commercially available symbolic manipulation program. The advantages of the algorithm and simulation results are discussed.

This paper describes a linear state-space model for a flexible single link manipulator arm. The resultant model is compared to an experimental four foot long direct drive manipulator. The method employed to generate the model utilizes a separable formulation of assumed modes to represent the tranverse displacement due to bending. Lagrangian dynamics are applied to determine the kinetic and potential energies for the system. The resultant dynamic equations are then organized into a state space model suitable for use in linear control system design procedures. The performance of the model is considered for different model orders and assumed modes. Several important aspects of candidate mode selection, and results for different model order are discussed. The final section of the paper provides a brief summary and describes ongoing and future work.