Publications

This paper describes how variable structure control can be used to describe the overall behavior of multiple autonomous robotic vehicles with simple finite state machine rules. The importance of this result is that we can then begin to design provably asymptotically stable group behaviors from a set of simple control laws and appropriate switching points with variable structure control. The ability to prove convergence to a goal is especially important for applications such as locating military targets or land mines.

This paper discusses recent experiments in the manipulation and assembly of parts with 100 micron outside dimensions and submicron tolerances. The objective of this work is to develop a micromanipulation workcell which can automatically assemble LIGA (Lithography Galvonoforming Abforming) parts using an assembly plan and a CAD drawing of each of the components. The workcell consists of an AdeptOne robot, precision stages, long distance microscope, and a high aspect ratio modeled polysilicon tweezers for picking up the parts. Fourier optics methods are used to generate synthetic microscope images from CAD drawings. These synthetic images are used off-line to test image processing routines under varying magnifications an depths of field. They also provide reference image features which are used to visually servo the true part to the desired position.

For the past seven years, Sandia National Laboratories has been active in the development of robotic systems to help remediate DOE`s waste sites and decommissioned facilities. Some of these facilities have high levels of radioactivity which prevent manual clean-up. Tele-operated and autonomous robotic systems have been envisioned as the only suitable means of removing the radioactive elements. World modeling is defined as the process of creating a numerical geometric model of a real world environment or workspace. This model is often used in robotics to plan robot motions which perform a task while avoiding obstacles. In many applications where the world model does not exist ahead of time, structured lighting, laser range finders, and even acoustical sensors have been used to create three dimensional maps of the environment. These maps consist of thousands of range points which are difficult to handle and interpret. This paper presents a least squares technique for fitting range data to planar and quadric surfaces, including cylinders and ellipsoids. Once fit to these primitive surfaces, the amount of data associated with a surface is greatly reduced up to three orders of magnitude, thus allowing for more rapid handling and analysis of world data.

Sandia National Laboratories (SNL) has recently developed a 16 cm{sup 3} (1 in{sup 3}) autonomous robotic vehicle which is capable of tracking a single conducting wire carrying a 96 kHz signal. This vehicle was developed to assess the limiting factors in using commercial technology to build miniature autonomous vehicles. Particular attention was paid to the design of the control system to search out the wire, track it, and recover if the wire was lost. This paper describes the test vehicle and the control analysis. Presented in the paper are the vehicle model, control laws, a stability analysis, simulation studies and experimental results.

A prototype sensor fusion framework called the `Knowledge Assistant' has been developed and tested on a gantry robot at Sandia National Laboratories. This Knowledge Assistant guides the robot operator during the planning, execution, and post analysis stags of the characterization process. During the planning stage, the Knowledge Assistant suggests robot paths and speed based on knowledge of sensors available and their physical characteristics. During execution, the Knowledge Assistant coordinates the collection of data through a data acquisition `specialist.' During execution and post analysis, the Knowledge Assistant sends raw data to other `socialists,' which include statistical pattern recognition software, a neural network, and model-based search software. After the specialists return their results, the Knowledge Assistant consolidates the information and returns a report to the robot control system where the sensed objects and their attributes (e.g. estimated dimensions, weight, material composition, etc.) are displayed in the world model. This paper highlights the major components of this system.

A prototype sensor fusion framework called the {open_quotes}Knowledge Assistant{close_quotes} has been developed and tested on a gantry robot at Sandia National Laboratories. This Knowledge Assistant guides the robot operator during the planning, execution, and post analysis stages of the characterization process. During the planning stage, the Knowledge Assistant suggests robot paths and speeds based on knowledge of sensors available and their physical characteristics. During execution, the Knowledge Assistant coordinates the collection of data through a data acquisition {open_quotes}specialist.{close_quotes} During execution and postanalysis, the Knowledge Assistant sends raw data to other {open_quotes}specialists,{close_quotes} which include statistical pattern recognition software, a neural network, and model-based search software. After the specialists return their results, the Knowledge Assistant consolidates the information and returns a report to the robot control system where the sensed objects and their attributes (e.g., estimated dimensions, weight, material composition, etc.) are displayed in the world model. This report highlights the major components of this system.

This report discusses work performed under a Cooperative Research And Development Agreement (CRADA) with Corning, Inc., to analyze and test various techniques for controlling the motion of a high speed robotic arm carrying an open container of viscous liquid, in this case, molten glass. A computer model was generated to estimate the modes of oscillation of the liquid based on the shape of the container and the viscosity of the liquid. This fluid model was experimentally verified and tuned based on experimental data from a capacitive sensor on the side of the container. A model of the robot dynamics was also developed and verified through experimental tests on a Fanuc S-800 robot arm. These two models were used to estimate the overall modes of oscillation of an open container of liquid being carried by a robot arm. Using the estimated modes, inverse dynamic control techniques were used to determine a motion profile which would eliminate waves on the liquid`s surface. Experimental tests showed that residual surface waves in an open container of water at the end of motion were reduced by over 95% and that in-motion surface waves were reduced by over 75%.

This paper describes an algorithm for determining the optimal placement of a robotic manipulator within a workcell for minimum time coordinated motion. The algorithm uses a simple principle of coordinated motion to estimate the time of a joint interpolated motion. Specifically, the coordinated motion profile is limited by the slowest axis. Two and six degrees of freedom examples are presented. In experimental tests on a FANUC S-800 arm, the optimal placement of the robot can improve the cycle time of a robotic operation by as much as 25%. In high volume processes where the robot motion is currently the limiting factor, this increased throughput can result in substantial cost savings.

This paper describes two methods for controlling the surface of a liquid in an open container as it is being carried by a robot arm. Both methods make use of the fundamental mode of oscillation and damping of the liquid in the container as predicted from a boundary element model of the fluid. The first method uses an infinite impulse response filter to alter an acceleration profile so that the liquid remains level except for a single wave at the beginning and end of the motion. The motion of the liquid is similar to that of a simple pendulum. The second method removes the remaining two surface oscillations by tilting the container parallel to the beginning and ending wave. A double pendulum model is used to determine the trajectory for this motion. Experimental results of a FANUC S-800 robot moving a 230 mm diameter hemispherical container of water are presented.

Considerable research has been performed on Robotic Visual Servoing (RVS) over the past decade. Using real-time visual feedback, researchers have demonstrated that robotic systems can pick up moving parts, insert bolts, apply sealant, and guide vehicles. With the rapid improvements being made in computing and image processing hardware, one would expect that every robot manufacturer would have a RVS option by the end of the 1990s. So why aren`t the Fanucs, ABBs, Adepts, and Motomans of the world investing heavily in RVS? I would suggest four seasons: cost, complexity, reliability, and lack of demand. Solutions to the first three are approaching the point where RVS could be commercially available; however, the lack of demand is keeping RVS from becoming a reality in the near future. A new set of applications is needed to focus near term RVS development. These must be applications which currently do not have solutions. Once developed and working in one application area, the technology is more likely to quickly spread to other areas. DOE has several applications that are looking for technological solutions, such as agile weapons production, weapons disassembly, decontamination and dismantlement of nuclear facilities, and hazardous waste remediation. This paper will examine a few of these areas and suggest directions for application-driven visual servoing research.

This paper describes a collision avoidance system using Whole Arm Proximity (WHAP) sensors on a PUMA 560 robot arm. The capacitance-based sensors generate electric fields which can completely encompass the robot arm and detect obstacles as they approach from any direction. The directional obstacle information gathered by the WHAP sensors together with the sensor geometry and robot configuration is used to scale the commanded joint velocities of the robot. A linearized relationship between the WHAP sensor reading and the distance from the obstacle allows direct transformation of perturbations in WHAP readings to perturbations in joint velocities. The WHAP reading is used to directly reduce the component of the command input velocity along the normal axis of the sensor, allowing graceful reductions in speed as the arm approaches the obstacle. By scaling only the component of the velocity vector in the direction of the nearest obstacles, the control system restricts motion in the direction of obstacles while permitting unconstrained motion in other directions.

This paper presents an infinite impulse response (IIR) filtering technique for reducing structural vibration in remotely operated robotic systems. The technique uses a discrete filter between the operator's joy stick and the robot controller to alter the inputs of the system so that residual vibration and swing are reduced. A linearized plant model of the system is analyzed in the discrete time domain, and the filter is designed using pole-zero placement in the z-plane. This technique has been successfully applied to a two link flexible arm and a gantry crane with a suspended payload.

This paper describes current research and development on miniaturized sensing systems for use during in situ characterization of nuclear waste storage tanks, buried waste sites, and decommissioned production facilities. Each miniaturized sensor system will consist of a suite of chemical, radiological, and physical properties sensors integrated into a compact package which will be mounted on the end of a robotic arm and/or vehicle. While the specific size of this remote sensor head and the types of sensors included will depend on site needs, the supporting generic computing system may be used for other waste characterization applications. This computing system will contain all necessary hardware and software to acquire, combine, interpret, display, and archive a wide range of sensor data. This paper describes the present status of the project, the lessons learned from the first prototype, and planned future designs of the next generation system. 7 refs.

This paper describes current research and development on a miniaturized sensing system for use during in situ characterization of nuclear waste storage tanks. Sandia is designing this sensing system as a tool for a large robotic arm that is deployed through an access port in the top of a storage tank. While the robot arm scans the sensing package over the waste, a distributed computing system acquires sensor data, correlates the data with the position of the robot, and produces maps of the chemical and radiological contents of the tanks in real time. We have built and demonstrated a first prototype system containing eight sensors. 53 refs.

The experimental results of model-based and sensor-based control strategies for a two-link, flexible, planar robot arm are analyzed. Off-line optimization determines a minimum-time, straight-line tip trajectory which stays within the torque constraints of the motors and ends in a quiescent state, i.e., no vibrational transients. An efficient finite-element model is used in the optimization to approximate the flexible arm dynamics. Control schemes tested include an open-loop torque profile, a closed-loop proportional-derivative (PD) joint controller, and a feedforward controller. This last scheme uses the torque profile created from the optimization, along with PD feedback of the joint angles and proportional feedback of the strain gages, to correct for errors in the model. The results show that the best performance is achieved with this feedforward approach.

The long length and relatively small cross sectional area of the robotic arms envisioned for use inside of the underground nuclear waste storage tanks will require the control of flexible structures. This will become an important problem in the characterization and remediation of these tanks. We are developing control strategies to actively damp residual vibrations in flexible robotic arms caused by high speed motion and abrupt external forces. A planar, two-link flexible arm is currently being used to test these control strategies. In this paper, two methods of control are discussed. The first is a minimum-time control approach which utilizes a finite element model and and optimization program. These tools plan the motor torque profiles necessary for the tip of the arm to move along a straight line, in minimum time, within the motors' torque constraints, and end in a quiescent state. To account for modeling errors in the finite element model, errors in joint angles, velocities, and link curvatures are added to the optimal torque trajectory. Linear quadratic Gaussian (LQG) regulatory design theory is used to determine the feedback gains. The second method of control is a teleoperated joystick controller which uses an input shaping technique to alter the commands of the joystick so as to reduce the residual vibration of the fundamental modes. Approximating the system as linear, the natural frequency and damping ratio are estimated on-line for the complete system, which includes the structure plus a lower level proportional derivative controller. An input shaping filter is determined from the estimated natural frequency, estimated damping ratio, and the desired transfer function of the system. 11 reps., 9 figs.