Search

Electrooculography (EOG) can be used to measure eye movements while the eyelids are open or closed and to assist in the diagnosis of certain eye diseases. However, challenges in biosignal acquisition and processing lead to limited accuracy, limited resolution (both temporal and spatial), as well as difficulties in reducing noise and detecting artifacts. Methods such as finite impulse response, wavelet transforms, and averaging filters have been used to denoise and enhance EOG measurements. However, these filters are not specifically designed to detect saccades, and so key features (e.g., saccade amplitude) can be over-filtered and distorted as a consequence of the filtering process. Here we present a model-based fusion technique to enhance saccade features within noisy and raw EOG signals. Specifically, we focus on Westheimer (WH) and linear reciprocal (LR) eye models with a Kalman filter. EOG signals were measured using OpenBCI’s Cyton Board (at 250 Hz), and these measurements were compared with a state-of-the-art EyeLink 1000 (EL; 250 Hz) eye tracker. On average, the LR model-based KF produced a 47% improvement of measurement accuracy over the bandpass filters. Thus, we conclude that our LR model-based KF outperforms standard bandpass filtering techniques in reducing noise, eliminating artifacts, and restoring missing features of saccade signatures present within EOG signals.

Modeling and simulation of human eye movement have a wide range of applications in many domains. Various attempts have been made to model and simulate eye movements in a physically accurate manner. All the existing models show limitations and problems in simulating secondary and tertiary eye movements. Recent investigation of pulley models (passive and active hypotheses) in representing human eye motion has recognized mathematical complexity in modeling eye behavior. Sophisticated techniques of modeling are required to investigate eye movements. This paper presents a procedure for eye movement simulation through geometrical modeling (an OpenSim script with its recent MATLAB binding) for binocular vision. First order neural dynamics with Millard's muscle model are used to actuate six Extra Ocular Muscles (EOMs). Pulse-step inputs are used to generate the muscle forces around the eye globe. The implemented model is successful in simulating horizontal and vertical movements of the human eye with respect to the prescribed activation. The developed technique is evaluated using responses from lumped parameter models and EOG recordings.

Underactuated hands are able to achieve shape adaptation to conformally grasp a wide variety of objects, while keeping low undesirable hand attributes such as weight, size, complexity and cost. The available analytical and simulation studies of planar underactuated hands normally assume quasi-static conditions and a fixed object. In the present paper, a new quasi-dynamic analysis of the grasping process in the horizontal plane by a planar, two-finger, four-joint underactuated hand is presented. The study considers object movement during the grasping process, and also contact friction with a surface that supports the object. An extensive and versatile simulation program, based on the analysis, is developed to investigate the effects of various parameters of hand and object on the grasping process. A prototype hand has been developed and the simulation results are validated experimentally. An extensive and detailed study and optimization exercise is carried out using the developed simulation tool. Specifically, the study concerns a manipulative grasping process that moves the object to the hand centerline during the process. Important new findings on the influence of link dimensions, link angular speeds, friction with the supporting surface, object mass and object size on the grasping performance of the hand in this scenario are presented. The results are used to establish new design guidelines for the hand. In particular, the results indicate that in the case where there is limited information on the size and precise initial location of the object to be grasped, the optimal hand design would involve inner to outer phalange size ratios of approximately 3:1, and inner phalange joints that are very close to each other., peer-reviewed

A typical mechatronic problem (modeling, identification, and design) entails finding the best system topology as well as the associated parameter values. The solution requires concurrent and integrated methodologies and tools based on the latest theories. The experience on natural evolution of an engineering system indicates that the system topology evolves at a much slower rate than the parametric values. This paper proposes a two-loop evolutionary tool, using a hybrid of genetic algorithm (GA) and genetic programming (GP) for design optimization of a mechatronic system. Specifically, GP is used for topology optimization, while GA is responsible for finding the elite solution within each topology proposed by GP. A memory feature is incorporated with the GP process to avoid the generation of repeated topologies, a common drawback of GP topology exploration. The synergic integration of GA with GP, along with the memory feature, provides a powerful search ability, which has been integrated with bond graphs (BG) for mechatronic model exploration. The software developed using this approach provides a unified tool for concurrent, integrated, and autonomous topological realization of a mechatronic problem. It finds the best solution (topology and parameters) starting from an abstract statement of the problem. It is able to carry out the process of system configuration realization, which is normally performed by human experts. The performance of the software tool is validated by applying it to mechatronic design problems.

This is the final part of a three-part series on micro-electromechanical systems (MEMS) sensor technology. In the first part, a general introduction to MEMS sensing was given including the underlying principles [1]. Biomedical MEMS sensors were described and the principles of bio-sensing in a typical set of biologically inspired sensors were presented. The second part discussed mechanical sensors and their applications [2]. Some important issues of MEMS sensors were addressed, including the need to compensate for environmental effects, the Casimir effect, harvesting of energy for self-powered sensors, and the subject of sensor selection.

This paper presents a system framework to automate the design evolution of multi-domain engineering systems. The developed system integrates machine health monitoring with an expert system to monitor the performance of an existing mechatronic system and to take a decision as to whether and which part of the system a design improvement is necessary. A system model represented by Linear Graphs is evolved using genetic programming to obtain a design that optimally satisfies a specified level of performance. The developed approach allows exploration of the design space to find the optimum design solution in an automated manner. In order to control the arbitrary exploration of the design space, domain knowledge, expertise, and input from the machine health monitoring system are used. The design evolution algorithm is implemented using GPLAB, a MATLAB tool, and integrated with Simscape for modeling and simulation. The developed system is applied to redesign the electro-mechanical conveyor system of an industrial fish processing machine.

This paper outlines the inspiration received by the author from the Zadeh–MacFarlane–Jamshidi trio in his pursuit concerning the theory and Application of fuzzy logic. Beginning with Zadeh’s pioneering work, a hierarchical control system was developed, in collaboration with MacFarlane, for application in robotic manipulators. Subsequently, the work was extended to an analytical basis for controller tuning using fuzzy decision making. On the prompting of Jamshidi to address the issue of knowledge-base simplification, theorems were developed related to decoupling a fuzzy rule base. These developments provided a theoretical basis for applying single-context decision making to a problem governed by the knowledge base of coupled fuzzy rules. The developed theorems establish an analytical equivalence between the decisions made from a coupled set of fuzzy rules and an uncoupled set of fuzzy rules concerning the same problem domain. These developments have been applied to supervisory control of an industrial fish cutting machine. The paper presents the pertinent theory and illustrative examples.

This paper develops a robust vision-based mobile manipulation system for wheeled mobile robots (WMRs). In particular, this paper addresses the retention of visual features in the field of view of the camera, which is an important robustness issue in visual servoing. First, the classical approach of image-based visual servoing (IBVS) for fixed-base manipulators is extended to WMRs and a control law with Lyapunov stability is determined. Second, in order to guarantee visibility of visual features, an innovative controller with machine learning using Q-learning is proposed, which can learn its behavior policy and autonomously improve its performance. Third, a hybrid controller for robust mobile manipulation is developed to integrate the IBVS controller and the Q-learning controller through a rule-based arbitrator. This is thought to be the first paper that integrates reinforcement learning or Q-learning with visual servoing to achieve robust operation. Experiments are carried out to validate the approaches developed in this paper. The experimental results show that the new hybrid controller developed here possesses the capabilities of self-learning and fast response, and provides a balanced performance with respect to robustness and accuracy.

This paper presents a modified, distributed Q-learning algorithm, termed as sequential Q-learning with Kalman filtering (SQKF), for decision making associated with multirobot cooperation. The SQKF algorithm developed here has the following characteristics. 1) The learning process is arranged in a sequential manner (i.e., the robots will not make decisions simultaneously, but in a predefined sequence) so as to promote cooperation among robots and reduce their Q-learning spaces. 2) A robot will not update its Q-values with observed global rewards. Instead, it will employ a specific Kalman filter to extract its real local reward from the global reward, thereby updating its Q-table with this local reward. The new SQKF algorithm is intended to solve two problems in multirobot Q-learning: credit assignment and behavior conflicts. The detailed procedure of the SQKF algorithm is presented, and its application is illustrated using a prototype multirobot experimental system. The experimental results show that the algorithm has better performance than the conventional single-agent Q-learning algorithm or the team Q-learning algorithm in the multirobot domain.

This paper presents a new scheme for fault detection and diagnosis (FDD) in spacecraft attitude determination (AD) sensors. An integrated attitude determination system, which includes measurements of rate and angular position using rate gyros and vector sensors, is developed. Measurement data from all sensors are fused by a linearized Kalman filter, which is designed based on the system kinematics, to provide attitude estimation and the values of the gyro bias. Using this information the erroneous sensor measurements are corrected, and unbounded sensor measurement errors are avoided. The resulting bias-free data are used in the FDD scheme. The FDD algorithm uses model-based state estimation, combining the information from the rotational dynamics and kinematics of a spacecraft with the sensor measurements to predict the future sensor outputs. Fault isolation is performed through extended Kalman filters (EKFs). The innovation sequences of EKFs are monitored by several statistical tests to detect the presence of a failure and to localize the failures in all AD sensors. The isolation procedure is developed in two phases. In the first phase, two EKFs are designed, which use subsets of measurements to provide state estimates and form residuals, which are used to verify the source of the fault. In the second phase of isolation, testing of multiple hypotheses is performed. The generalized likelihood ratio test is utilized to identify the faulty components. In the scheme developed in this paper a relatively small number of hypotheses is used, which results in faster isolation and highly distinguishable fault signatures. An important feature of the developed FDD scheme is that it can provide attitude estimations even if only one type of sensors is functioning properly.

Due to their extensive utilization in engineering designs, various vibration controllers have been investigated with design specifications in mind. The optimization of vibration controller designs is a complicated multi-criteria problem. In this paper, the mechatronic design quotient (MDQ) approach and genetic algorithm (GA) are coupled together to determine this optimal design solution. The MDQ is presented to formulate an evaluation function for the passive vibration controller design of flexible beam structures, and the GA is then used to maximize this function so as to achieve a design solution with the highest MDQ value. For comparison, both the MDQ performance of passive vibration controller design using linear dampers and active vibration controller design using a linear quadratic regulator are provided. The latter is used as the performance benchmark to evaluate the optimal design solution of the former. Experimental results show that the linear dampers design with the proposed method can achieve similar performance to an optimal active controller.

A mechatronic system needs an integrated, concurrent, and system-based design approach due to the existence of interactions among its subsystems, and also the existence of interactions between the criteria involved in a realistic evaluation of a mechatronic product. This paper presents a systematic methodology for a detailed mechatronic design based on a mechatronic design quotient (MDQ). MDQ is a multicriteria index, reflecting a system-based evaluation of a mechatronic design, which is calculated using soft computing techniques, thereby accommodating interactions between criteria and human experience. A niching genetic algorithm is utilized to explore the huge search space raised due to concurrent and integrated design approach, with the aim to find the elite representatives of different possible configurations. To demonstrate the method, it is applied to an industrial fish cutting machine called the Iron Butcher-an electromechanical system that falls into the class of mixed or multidomain systems.

The principles and applications of dc motors, ac induction motors, stepper motors, solenoid actuators, and micromotors used in process automation are discussed. Their operating principles, characteristics control approaches are outlined. Several output devices typically used in industrial applications are defined. Microactuators, new members of the output equipment family, are also introduced. Several real industrial actuator products and their specifications are described. Computer-based displays and data loggers are introduced and and important technical trends are pointed out. It is clear that future applications of output devices will focus on the development of industrial actuators of smaller size, lighter weight, higher output power and torque, and better static and dynamic characteristics. Novel mechanisms for force generation and more sophisticated machining technique to facilitate the realization of these goals are expected.

This paper addresses the stability of a newly-developed control strategy for networked control systems (NCS). This control strategy hones the potential of constrained model predictive control (MPC) by buffering the predicted control sequence at the actuator in anticipation of typical data transmission errors associated with NCS. Closed-loop stability in the sense of Lyapunov is guaranteed for the controller in the linear case, by bounding the projected receding horizon costs by lower- and upper-bounding terms using a predetermined terminal cost. A stability theorem is developed, which provides a suboptimal measure for the controller in real time, and is sufficient to estimate the worst-case transmission delay that can be handled by the developed control buffering strategy. The stability conditions, as governed by the theorem, are validated through real-time implementation on an electro-hydraulic servo system of an industrial processing machine, through an Ethernet network.

A concurrent, integrated, and multicriteria methodology is presented for the conceptual design of mechatronic systems. It uses an evaluation model called mechatronic design quotient (MDQ) to facilitate decision-making in the design process. A nonlinear fuzzy integral is used for the aggregation of criteria in MDQ, thereby accommodating possible correlations among them. The performance of the developed methodology is validated by applying it to an industrial fish cutting machine called Iron Butcher-an electromechanical system that falls into the class of mixed or multidomain systems

In metal cutting processes, an effective monitoring system, which depends on a suitably developed scheme or set of algorithms can maintain machine tools in good condition and delay the occurrence of tool wear. In this paper, an approach is developed for fault detection and diagnosis based on an observer model of an uncertain linear system. A robust observer is designed, using the derived uncertain linear model, to yield the necessary and key information from the system. Subsequently, it is used as a state (tool wear) estimator, and fault detection is carried out by using the observed variables and cutting force. The developed approach is applied to milling machine center. Several linear models are identified based on different working conditions. A dominant model plus uncertain terms is derived from these model set and used as an observer. Threshold values are proposed for detecting the fault of the milling machine. Examples taken from experimental tests shown that the developed approach is effective for the fault detection. The approach can be used for fault detection of failures arising from sensor or actuator malfunction.

When using linear dampers for the suppression of vibration in flexible structures, their mounting locations and the damping coefficients have to be chosen properly. In this article, an approach to the optimal design of linear dampers for vibration control in flexible beams based on the mechatronic design quotient (MDQ) is presented. This approach strives to make the optimal concurrent design of the dampers come as close as possible to the performance of an optimal sequential design of uncoupled subsystems. The MDQ approach provides a practical way to evaluate the performance of multi-criteria design. It also provides an insight into the performance degradation of various subsystems caused by dynamic coupling effects in the overall system. In the approach developed here, a cost function, based on the modal settling times, is first established from the design objective. Next, the optimal performance indices from the sequential and concurrent designs are obtained by selecting the coefficients and locations of the vibration dampers. These performance indices are used to determine the design corresponding to the highest MDQ value, which gives the overall optimal design. For performance evaluation, optimal design of active vibration controllers is carried out, using the linear quadratic regulator (LQR), and its MDQ value is established. The results from two case studies show that optimal linear dampers designed using the MDQ approach can achieve the performance of an optimal active controller. This approach can also provide an insight into the performance degradation in each subsystem (or sub-problem) due to the dynamic coupling effects in the overall system.

Fault detection and isolation in mobile robots has become a challenging task primarily due to uncertain and dynamic operating environments. The design of model-based fault detection methods would not be a practical real-time solution in view of the dynamic and uncertain nature of the problem. Also, conventional single-sensor approaches have limitations in practical applications. In this paper, a method of fault detection and isolation (FDI) based on a multi-level data fusion and response (behavioral) analysis technique is presented. The proposed FDI scheme mainly consists of pre-processing, sensor-fusion, a conflict monitoring unit, a confidence level computation unit, a high-level information fusion unit and a fault isolation unit. The developed FDI method is implemented in a simulated robot environment employing IR/camera fusion for navigation and obstacle avoidance. The fusion-based FDI method is tested under faults in camera and IR sensor. With the developed approach, faults are detected in a timely manner and isolated accurately. Also, with the incorporation of sensor fusion, reliable and accurate sensor information is adaptively fused and fault tolerance is achieved under camera/IR sensor faults.

Fault tolerance is increasingly important for multirobot systems, especially those operating in an unknown and hazardous environments. This paper presents a framework for a fault tolerant multi-robot system that works cooperatively to achieve a common goal. The proposed methodology comprises four key phases: grouping of robots, task execution, fault detection, and fault tolerance. Initially robots are grouped based on their capabilities. Next, the process of task execution takes place during which components of the robots are monitored in real time to detect faults. Upon detection of partial or full failure of a robot, the other team members respond robustly, reliably, flexibly, and coherently while communicating with each other to regroup and perform remaining tasks in order to achieve the common goal. The proposed methodology is implemented in a group of simulated heterogeneous mobile robots that transport multiple objects cooperatively to a goal location in an unknown and unstructured environment.

There is a growing trend in developing the wireless machine-to-machine (M2M) communication for factory automation. Currently, most wireless protocols for infrastructureless mobile ad-hoc networks focus on the mobility issue and are mainly suitable for narrowband radio devices. The mechanisms for route update and discovery that work well for wireless mobile ad-hoc networks prove to be redundant and instead add to processing overheads in a static wireless factory setting. This paper presents the development and implementation of a new wireless routing protocol, Bluewave, that caters specifically to wireless communication between machines in a factory setting. Bluewave utilizes features of Bluetooth technology when performing the route setup. The merits of the proposed protocol, in terms of shorter initialization time and route setup time, are duly discussed. A case study illustrating the application of Bluewave to an online condition-based maintenance scenario is presented. Experimental results obtained are furnished to show the viability of Bluewave in industrial automation

This paper investigates the frequency-domain approach for some useful types of analysis and design related to problems of mechanical vibration. First, some attractive features of the frequency-domain analysis are noted. The possibility of a unified approach to analysis is indicated through the existence of analogies between mechanical and non-mechanical systems. The uses of mechanical impedance and transmissibility functions in the modeling, analysis, and design of mechanical vibrating systems are illustrated. Applications in the design of vibration isolators and in the analysis of material damping are given to demonstrate the attractive characteristics of some frequency-domain techniques in the field of mechanical vibration. The manner in which the material presented in the paper may be incorporated into an undergraduate core course in mechanical vibration is indicated.

Distributed control networks encounter nondeterministic delays in data communication between sensors, actuators, and controllers, including direct-feedback control and higher level supervisor control. This paper presents a novel strategy, which extends the Generalized Predictive Control (GPC) algorithm to compensate for these data-transmission delays. Communication lines between sensors to controller and controller to actuators are considered. The present strategy incorporates a minimum-effort estimator to estimate missing or delayed sensor data and a variable-horizon adaptive GPC controller to predict the required future control efforts to drive the plant to track a desired reference trajectory. Action buffers are introduced at the actuators to sequence the future control efforts. A parallel objective of this paper is to investigate the suitability of the Ethernet network, which is cost-effective and widely deployed for implementing networked control systems. An Ethernet-based client-server control architecture is developed. The developed scheme is implemented on the dual-axis hydraulic position system of an industrial fish-processing machine.

The present study focuses on the development and implementation of an intelligent hierarchical controller for the vibration control of a deployable manipulator. The emphasis is on the use of knowledge-based tuning of the low-level controller so as to improve the performance of the system. To this end, first a fuzzy inference system (FIS) is developed. The FIS is then combined with a conventional modal controller to construct a hierarchical control system. Specifically, a knowledge-based fuzzy system is used to tune the parameters of the modal controller. The effectiveness of the hierarchical control system is investigated through numerical simulation studies. Realistic examples are considered where the system experiences vibrations due to initial disturbances at the flexible revolute joint or due to maneuvers of a deployable manipulator. The results show that the knowledge-based hierarchical control system is quite effective in suppressing vibrations induced due to the above mentioned disturbances. Results suggest that performance of the modal controller could be significantly improved through knowledge-based tuning.

This paper presents a real-time open-architecture control system (ROACS) which has been developed for flexible manipulation of industrial robots. Flexible manipulation refers to robot manipulation that handles tasks with uncertainties; hence, decision-making based on feedback data is essential in realtime operation. A real-time open-architecture control system with the capacity of parallel processing of realtime events, extraction of information from realtime data, and intelligent decision-making, is developed. The entire system consists of a real-time subsystem which manages robot hardware and executes path planning and data processing, and an intelligent subsystem which performs intelligent decision-making and feedback task control. In the context of intelligent task control, information extraction, fuzzy-logic-based interpretation and decision-making, and a novel design of associated real-time robot task language (RTTL) are developed. The conflicts between high bandwidth requirements for real-time services and the undeterministic time length for intelligent decision-making are managed in a cooperative real-time intelligent system model. Client–server architecture is found quite suitable for implementation of the system. The entire system has been successfully developed, implemented, and demonstrated for a robotic salmon slicing task which requires online determination of the backbone position.

Mechatronic systems are electromechanical systems where an integrated design procedure is used with respect to mechanical and electrical aspects. These systems possess computational intelligence with capabilities such as perception, learning, reasoning, and making inferences (or, decisions) from incomplete information. Concepts of sensing and information acquisition are crucial for mechatronic systems in all stages of design, development, testing, operation, and maintenance. This paper outlines the considerations of sensing and information acquisition for intelligent mechatronic systems. A typical mechatronic system operates as a self-contained unit and involves direct information acquisition. There are situations, however, where a group of mechatronic systems have to operate in a coordinated manner, as in a mechatronic workcell. Also it may be necessary to monitor a mechatronic system remotely and to share common resources between several applications. Networked control is needed then. This issue is also addressed in the paper. Several industrial applications of intelligent mechatronics have been developed in our Industrial Automation Laboratory. Some of them involve cutting, inspection, and grading of fish products. The paper presents a representative application, by highlighting information acquisition and the integration of sensing and control.

Circular saws are widely used in wood processing for applications ranging from primary lumber manufacturing to furniture industry and home workshops. They directly contribute to production problems such as poor cutting accuracy, poor surface quality, short tool life, high noise levels, and excessive raw material wastage. Vibration of the saw blade during woodcutting has been identified as a key reason for poor wood recovery. In fact, about 12% of the raw material in woodcutting ends up as sawdust due to excessive sawing gap. Efficient wood sawing is being pursued with the objective of mitigating these problems; particularly, to reduce saw blade vibration and sawdust. In this paper, we present active control of saw blade vibration using linear quadratic Gaussian (LQG) control. We outline a test rig that has been developed for our experimental investigation. The system configuration is described and the control problem is formulated. The experimental procedure for identification of a system model is described. The implementation of the LQG control scheme is outlined, and typical results from the experimental control system are presented and discussed. The developed controller is shown to be very effective in the present application, as evident from the results that have been obtained. In particular, the amplitude of the saw blade vibration has been reduced by 66% on average using active control, compared to cases with no control. Also, the cutting gap (kerf) has been reduced by 25%, from 2.00 mm to 1.50 mm, through active control. In terms of 1995 prices, this would correspond to an increased revenue of $640,000 per year for a mill producing 100 MM fbm of lumber annually.

The paper focuses on a novel ground-based prototype manipulator. The system has two identical modules connected in a chain topology. Each module consists of two links: one free to slew while the other is permitted to deploy. To begin with, design, construction and integration of the manipulator are explained. This is followed by the development of a mathematical model for the manipulator using the Lagrangian procedure. Finally, a series of trajectories are tracked using the proportional-integral-derivative (PID) control and feedback linearization technique (FLT). The objective is the real-time implementation of the control algorithms, developed for the unique space manipulator, on a earth-based prototype system.

Advancement in food processing requires development of innovative sensor technologies that are flexible, fast, and accurate. A new approach has been developed for achieving optimal packaging of product portions so that the deviation of the portion weight from a specific target weight is minimized. The new portioning technology depends primarily on fast and accurate sensing of complex properties of the processed natural object. Conventional sensors would not meet the speed and cost constraints of on-line processing. A new model-based approach for fast and cost effective on-line sensing of a complex property such as the weight distribution of a natural object is presented in this paper. The approach involves off-line development of a model-base that contains non-dimensional models of the target property and a set of correlated simple features that can be measured easily at high speed and high accuracy. The on-line sensing process involves measurement of these simple features for each processed object and matching them to the appropriate property model in the model-base. The approach is applied here in an innovative can-filling process of fish to estimate the weight-rate function of the processed fish. Real data of fish are used to build the non-dimensional models of fish. Results presented in this paper indicate that the sensing approach is suitable for the fish processing industry with respect to the needed accuracy and speed.

Using a novel space platform-based manipulator with slewing and deployable links, the paper addresses two issues of considerable importance: (a) How important is it to model flexibility of the system? (b) How many modes are needed to adequately represent the elastic character? Results suggest that the fundamental mode is able to capture physics of the response quite accurately. Due to its massive character, the platform dynamics is virtually unaffected, even by severe maneuvers of the manipulator. Hence, treating the platform as rigid would save the computational cost without affecting the accuracy. Although the link flexibility does affect the manipulator's tip vibration, the joint and platform vibrations remain negligible. The revolute joint flexibility appears to be an important parameter affecting both the joint as well as tip responses. The information should prove useful in the design of this new class of manipulators.

Accurate and high-speed generation of a product portion for packaging or canning is of great industrial significance. Optimal portion control refers to making the portion as close as possible to a ...

This paper studies dynamics of a space platform—based mobile two-module manipulator with each module consisting of two links, one free to slew (revolute joint) and the other permitted to deploy and retrieve (prismatic joint). Both the platform and the manipulator are considered flexible. Using the computation ally efficient order- N Lagrangian formulation of the system, the paper presents results of a comprehensive parametric study aimed at assessing effects of initial conditions, system variables, and manipulator maneu vers on the uncontrolled response. In general, the mass of payload, speed of maneuver, and joint flexibility represent three important parameters governing the system dynamics. Results suggest significant coupling between the platform, link, and joint vibrations as well as the system libration. The system exhibited unac ceptable response under critical combinations of parameters, thus indicating a need for active control. Next, effectiveness of a composite control strategy, involving the feedback linearization technique (FLT) applied to the rigid degrees of freedom with flexible generalized coordinates regulated through the linear quadratic regulator, is assessed. The control procedure proves to be remarkably successful in damping even relatively large disturbances in a robust fashion. Finally, the numerical simulation results are complemented with the prescribed trajectory tracking data, obtained using a two-module (four links) ground-based prototype manip ulator in conjunction with the FLT and proportional integral derivative control. Considering the presence of friction and backlash at the joints, the tracking results may be considered satisfactory. The comprehensive study lays a sound foundation for the design of this class of novel manipulators.

A relatively general formulation for studying dynamics and control of a novel multi-module mobile manipulator, with slewing as well as deployable links, is presented. The deployment character leads to several advantages including reduced coupling, fewer singular configurations and ease of obstacle avoidance. The three-dimensional, order N , Lagrangian approach accounts for interactions between orbital, librational, slew, deployment and elastic degrees of freedom. The versatile character of the formulation makes it applicable to a large class of manipulator systems of contemporary interest. Of particular significance is the order N property which promises considerable saving in computational cost.

A rather general formulation for a novel manipulator system was developed in Part I. Here its versatility in studying dynamical response under wide variety of situations is demonstrated. To begin with, validity of the formulation and the associated computer code is established through conservation of energy. A planar parametric study follows which provides appreciation as to the influence of several important system variables, initial disturbances and maneuver profiles. Results indicate that the system flexibility can significantly affect the manipulator's performance. Under critical combinations of parameters, the system response can become unacceptable suggesting a need for control.

Application of the fuzzy logic intelligent control of space-based systems has received relatively less attention compared to conventional control procedures. Basic concepts involved in development of a hierarchical control system using fuzzy logic are explained. Next, fuzzy logic is applied to control a novel two-link robotic arm, composed of revolute and prismatic joints, supported by a mobile base traversing along an orbiting platform. The control system that is developed has three levels. A conventional controller, consisting of the feedback linearization technique combined with proportional-derivativecontrol is used in the bottom level of the hierarchical system, to control the servomotors of the robot. A second layer consists of a servo-expert that preprocesses the high-resolution information coming from joint encoders and extracts the status of the system. A third, intelligent layer is added at the top of the hierarchy to complete the control structure. The main purpose of the top level is to tune the parameters of the crisp (conventional) controller to improve the response of the system. A decision table is developed of ine, for tuning the parameters of the crisp (nonfuzzy) controller, which considerably reduces the real-time computational effort. Typical pick-and-place results from the robot are presented that show that extra tuning can signiŽ cantly improve the performance of the manipulator.

In the present context, optimal portion control refers to the process of preparing can-filling portions so that the deviation of the portion weight from a specified target weight is minimized. An approach has been developed for achieving this where a batch of objects is placed in a linearly overlapped optimal arrangement and then cut into portions using a series of parallel blades. The parameters of optimization are the arrangement order, orientation and degree of overlap of the objects. The approach has been demonstrated to produce impressive improvements in the application of fish canning. For this application, two approaches of optimal cutting are compared in the present paper. In one approach, the blade spacing is kept fixed and constant at a predetermined value. In the second approach, the blade spacing is varied for each portion after the objects are placed according to the optimal arrangement, where the target weight distribution is allowed to vary within a tolerance interval. The results presented in this paper indicate that the second approach produces a significantly higher percentage of acceptable portions than the first approach. What is presented are results from computer simulations, utilizing true data as measured from actual batches of fish. The paper demonstrates the potential benefit of the optimal portion control approach when applied in an industrial fish-canning process. Copyright © 2001 John Wiley & Sons, Ltd.

The main objective of portion control is to ensure that desired portions, usually specified by weight, are placed in the packages. In can-filling in particular, an optimal goal would be to minimize underfilling and overfilling. This paper develops an advanced packaging process for automated can-filling of fish, which achieves this goal. The overall automated system uses an innovative technique of optimal overlapping and cutting of fish. First, a batch of fish are overlapped in a linear orientation where the ordering sequence, the head orientation, and the degree of overlap between fish are the variables of optimization. The optimization is carried out with the objective of minimizing the absolute total weight of underfill and overfill of the produced cans. The optimal portioning method should possess a computational speed that is consistent with the process speed and the filling accuracy requirements. Several models of optimization have been developed. This paper follows a model development procedure that realizes a feasible and practical model. A numerical example that uses real data on a batch of salmon is presented to illustrate the approach and to demonstrate its feasibility in achieving both optimization objective and the processing speed. A comparison of several optimization models that have been developed is given, with respect to the computational speed and the filling accuracy. Results show that the optimal portioning method is able to achieve high production rates and improved filling accuracy in can-filling process of salmon.

Space manipulators present several features uncommon to ground-based robots: They are highly flexible, are often mobile, and have a degree of redundancy. As space robots become more complex, efficient algorithms are required for their simulation and control. The present study uses an order N algorithm, based on the Lagrangian approach and velocity transformations, to simulate the planar dynamics of an orbiting manipulator with arbitrary number of slewing and deployable flexible links. The relatively general formulation accounts for interactions between orbital, librational, slewing, deployment, and vibrational degrees of freedom and, thus, is applicable to a large class of manipulator systems of contemporary interest. A parametric analysis of the system dynamics suggests significant coupling between the rigid-body motion and structural vibrations. Obviously, this would affect the manipulator's performance. A nonlinear controller based on the feedback linearization technique is developed to regulate the rigid degrees of freedom. Nomenclature />/ = vector describing the position of F/ relative to F0 d\ = translation of the frame F/ from the tip of the (/th-l)body dm/ = mass of the infinitesimal element located on the /th body E AI , EIj = structural stiffness of the /th body in the

This paper outlines an innovative, automated system for portioning and filling of fish into cans in an optimal manner. The overall system includes the stages of pre-filling, filling, and post-filling. An integrated approach to optimal grouping and cutting, according to a weight-based portioning criterion, is central to the advantages of the approach. The associated technology includes fast and accurate estimation of the weight distribution of each fish, a portion optimization method, handling, conveying, and cutting devices, advanced sensor technology, and multi-layered intelligent control. The paper outlines the concepts of this advanced-technology system. An example is given to illustrate the yield improvements that are possible through an optimal portioning procedure.

Analysis and estimation of friction and compensation for its effects in the control of an electrohydraulic manipulator is addressed. The specific hydraulic manipulator is an integral part of an automated fish processing machine which has been developed in our laboratory. The analysis reveals that considerable static and dynamic friction exists in the system. An available nonlinear observer for Coulomb friction, is modified to simultaneously estimate friction, velocity, and acceleration. A novel observer-based friction compensating control strategy is developed for improved tracking performance of the manipulator. The approach is based on acceleration feedback control. Experimental investigations show that this controller significantly outperforms the conventional PD controller. The general approach presented in this paper, may be applied to compensate for friction in any servomechanism, particularly when the actuator dynamics is not negligible.

A laboratory workcell has been developed for fish cutting. The workcell, which uses a variety of sensors, actuators, and hardware for component interfacing and control, operates with the help of a dedicated supervisory control system. A layered architecture has been used for the system. It has several knowledge-based modules for carrying out tasks such as workcell monitoring, controller tuning, workcell conditioning, and product quality assessment. The paper describes the workcell and the development of the supervisory control system. Development of the knowledge-based modules of the system, which employ fuzzy logic for knowledge representation and decision making, is described.

A laboratory workcell has been developed for fish cutting. The workcell, which uses a variety of sensors, actuators, and hardware for component interface and control, operates with the help of a dedicated supervisory control system. A layered architecture has been used for the system. It has several knowledge-based modules for carrying out tasks such as workcell monitoring, controller tuning, workcell conditioning, and product quality assessment. Part I of the paper described the workcell and the development of the supervisory control system [de Silva, C.W. and Wickramarachchi, N., 1998. Knowledge-Based Supervisory Control System of a Fish Processing Workcell; Part I: System Development. Engineering Applications of Artificial Intelligence, This Issue]. Implementation of the overall system, including a versatile user interface, is presented here. Also, the performance of the knowledge-based supervisory system is studied through experiments of fish cutting.

This work presents the enhancement and application of a fuzzy classification technique for automated grading of fish products. Common features inherent in grading-type data and their specific requirements in processing for classification are identified. A fuzzy classifier with a four-level hierarchy is developed based on the "generalized K-nearest neighbor rules". Both conventional and fuzzy classifiers are examined using a realistic set of herring roe data (collected from the fish processing industry) to compare the classification performance in terms of accuracy and computational cost. The classification results show that the generalized fuzzy classifier provides the best accuracy at 89%. The grading system can be tuned through two parameters-the threshold of fuzziness and the cost weighting of error types-to achieve higher classification accuracy. An optimization scheme is also incorporated into the system for automatic determination of these parameter values with respect to a specific optimization function that is based on process renditions, including the product price and labor cost. Since the primary common features are accommodated in the classification algorithm, the method presented here provides a general capability for both grading and sorting-type problems in food processing.

The development and integration of sensor technology for herring roe grading is described. In this application, by and large, grading is presently done by hand. Shape, firmness/texture, color and weight all play key roles in determining the grade of a roe skein. Furthermore, number of these variables are quite subjective both in the way they are determined by human graders, and in the way they affect the overall grade of a skein. Thus, in the present system an "intelligent" decision-making system, with multiple levels of abstraction, is used to determine the final grade of the product. The system uses a model base and also a knowledge base developed through interactive acquisition of knowledge from grading experts and off-line experiments. These are implemented as a set of fuzzy-logic rules and model matching procedures. A prototype grading machine has been developed and tested at an industrial plant. Typical experimental results are presented and studied.

In process applications, fast and accurate extraction of complex information from an object for the purpose of mechanical processing of that object, is often required. In this paper, a general rule-based approach is developed using a database of measurable geometric "features" and associated complex information. The rules relate the features to the complex processing information. During the on-line processing, the object features are measured and passed into the rule base. The output from the rule base is the complex information that is needed to process the object. A methodology is developed to generate probabilistic rules for the rule base using multivariate probability densities. A knowledge integration scheme is also developed which combines statistical knowledge with expert knowledge in order to improve the reliability and efficiency of information extraction. The rule generation methodology is implemented in a knowledge-based vision system for process information recognition. As an illustrative example, the problem of efficient head removal in an automated salmon processing plant is considered.

This article presents a knowledge-based approach for firmness assessment of herring roe, for application in an automated grading machine. A low-level ultrasound sensor along with knowledge-based preprocessing and high-level decision making is interpreted as an intelligent firmness sensor, in the present context. Techniques of low-level sensing and information preprocessing are presented. The rule-based decision making system for quality assessment uses fuzzy logic concepts, which incorporates the expertise of experienced human graders. The membership functions are estimated for firmness resolution, using a probabilistic approach. It is found that this approach is superior to the use of trapezoidal membership functions.

This paper presents a knowledge-based approach for automatic load sharing among workcell components in a flexible production system (FPS). The approach is termed “dynamic structuring” and it enables an FPS to choose its configuration automatically to suit new situations and production demands. A method of incorporating fuzzy logic in representing the component-sharing knowledge, and in processing the associated decisions is outlined. A blackboard architecture is described for implementation of the restructuring system, using Prolog. An illustrative example of a multiple-workcell, automated, fish processing plant is simulated to demonstrate the application of the approach.

An intelligent sensor, based primarily on computer vision, for the assessment of quality of processed fish, as applied to an automated fish processing plant, is discussed. The role of information resolution on the degree of intelligence required for data processing is given major consideration. A hierarchy of data processing is introduced where the crisp, high resolution data of an image is transcribed into low resolution context that is more amenable to further intelligent processing at a higher level of the hierarchy. Experiments earned out in the laboratory with a prototype fish processing workcell are described and the results are summarised.

In a coupled rule base, more than one condition variable may link with more than one action variable within the same rule. This paper develops an analytical framework for decoupling such a rule base, with specific reference to fuzzy-logic control. The computational problem of fuzzy-logic is greatly simplified through the use of an uncoupled rule base. Conditions are derived, under which, decision making through an uncoupled rule base becomes equivalent to that through a coupled rule base. A laboratory prototype of a single-axis servo motor system which incorporates a knowledge-based tuner is employed to demonstrate the effective use of an uncoupled rule base that is equivalent to a coupled one. >

Robotics, machine vision, and intelligent control have been integrated in the technology development for flexible automation of fish processing. Robotic techniques are applicable in handling, moving, positioning, and cutting tasks. Machine vision is used for real-time gauging at a low level and in quality monitoring and grading at a high level. Intelligent control employing knowledge-based and hierarchical approaches are particularly useful in system tuning, self-organization and other high-level control tasks. A workcell for head removal of salmon has been developed, and an automated system for grading herring roe is being built. This paper outlines the recent developments pertaining to various projects within the framework of flexible automation of fish processing, that are being carried in the Industrial Automation Laboratory.