Generate Control Inputs Research Articles

Distributed Fault-Tolerant Control for Networked Robots in the Presence of Recoverable/Unrecoverable Faults and Reactive Behaviors

The paper presents an architecture for distributed control of multi-robot systems with an integrated fault detection, isolation and recovery strategy. The proposed solution is based on a distributed observer-controller schema where each robot, by communicating only with its direct neighbors, is able to estimate the overall state of the system; such an estimate is then used by the controllers of each robot to achieve global missions as, for example, centroid and formation tracking. The information exchanged among the observers is also used to compute residual vectors that allow each robot to detect failures on anyone of the teammates, even if not in direct communication. The proposed strategy considers both recoverable and unrecoverable actuator faults as well as it properly manages the possible activation of reactive local control behaviors of the robots (e.g., the activation of obstacle avoidance strategy) which generate control inputs different from those required by the global mission control. In particular, when the robots are subject to recoverable faults, those are managed at a local level by computing a proper compensating control action. On the other side, when the robots are subject to unrecoverable faults, the faults are isolated from anyone of the teammates by means of a distributed fault detection and isolation strategy; then, the faulty robots are removed from the team and the mission is rearranged. The proposed strategy is validated via numerical simulations where the system properly identifies and manages the different cases of recoverable and unrecoverable actuator faults, as well as it manages the activation of local reactive control in an integrated case study.

Online Learning Control of Hydraulic Excavators Based on Echo-State Networks

In some of recent advances in automation of construction equipment, much research has been conducted on the control of hydraulic excavators in both industry and academia for the benefit of safety and efficiency. However, most relevant works have employed model-based control approaches that require a mathematical representation of the target plant. For hydraulic excavators, obtaining a useful dynamic model for control can be challenging due to the nonlinearity of the hydraulic servo system. With this in mind, this paper investigates the feasibility of an online learning control framework based on echo-state networks (ESNs) to the position control of hydraulic excavators. While ESNs are a class of recurrent neural networks, the training of ESNs corresponds to solving a linear regression problem, thus making it suitable for online implementation. By exploiting the dynamic properties of ESNs, the deployed control framework uses the input and output signals of the plant to learn an inverse model, which is then used to simultaneously generate control inputs to track the desired trajectory. Experiments conducted on a 21-ton class hydraulic excavator show the promising results in that accurate tracking is achieved even in the absence of a dynamic model.

High-Density Electromyography and Motor Skill Learning for Robust Long-Term Control of a 7-DoF Robot Arm.

Myoelectric control offers a direct interface between human intent and various robotic applications through recorded muscle activity. Traditional control schemes realize this interface through direct mapping or pattern recognition techniques. The former approach provides reliable control at the expense of functionality, while the latter increases functionality at the expense of long-term reliability. An alternative approach, using concepts of motor learning, provides session-independent simultaneous control, but previously relied on consistent electrode placement over biomechanically independent muscles. This paper extends the functionality and practicality of the motor learning-based approach, using high-density electrode grids and muscle synergy-inspired decomposition to generate control inputs with reduced constraints on electrode placement. The method is demonstrated via real-time simultaneous and proportional control of a 4-DoF myoelectric interface over multiple days. Subjects showed learning trends consistent with typical motor skill learning without requiring any retraining or recalibration between sessions. Moreover, they adjusted to physical constraints of a robot arm after learning the control in a constraint-free virtual interface, demonstrating robust control as they performed precision tasks. The results demonstrate the efficacy of the proposed man-machine interface as a viable alternative to conventional control schemes for myoelectric interfaces designed for long-term use.

Optical Flow-Based Tracking of Needles and Needle-Tip Localization Using Circular Hough Transform in Ultrasound Images.

Image-guided interventions have become the standard of care for needle-based procedures. The success of the image-guided procedures depends on the ability to precisely locate and track the needle. This work is primarily focused on 2D ultrasound-based tracking of a hollow needle (cannula) that is composed of straight segments connected by shape memory alloy actuators. An in-plane tracking algorithm based on optical flow was proposed to track the cannula configuration in real-time. Optical flow is a robust tracking algorithm that can easily run on a CPU. However, the algorithm does not perform well when it is applied to the ultrasound images directly due to the intensity variation in the images. The method presented in this work enables using the optical flow algorithm on ultrasound images to track features of the needle. By taking advantage of the bevel tip, Circular Hough transform was used to accurately locate the needle tip when the imaging is out-of-plane. Through experiments inside tissue phantom and ex-vivo experiments in bovine kidney, the success of the proposed tracking methods were demonstrated. Using the methods presented in this work, quantitative information about the needle configuration is obtained in real-time which is crucial for generating control inputs for the needle and automating the needle insertion.

Application of Echo-State Networks to the Position Control of Shape-Memory Alloys

Shape-memory alloys (SMA) have the ability to generate strain in response to temperature change. However, the relationship between applied temperature and crystalline phase is hysteretic. Obtaining an explicit representation of this highly nonlinear phenomenon for the purpose of control consumes time and effort. Also, the identification process is subject to uncertainties, and furthermore, the dynamic properties of SMAs may change during its lifetime which reduces the reliability of the identification. With this in mind, we employ an online learning control framework for the position control of an SMA wire. The online learning control framework performs an inverse learning of the plant based solely on the input and output signals, and uses this information to generate control inputs. Thus, an explicit representation of the plant is not required and as a result from online learning, the controller adapts to changes of the plant. The specific method with which the inverse learning is employed, is by the use of echo-state networks (ESN). ESNs are a class of recurrent neural networks and are distinguished by their large number of hidden nodes often referred to as a dynamic reservoir. While the originally proposed method of constructing this dynamic reservoir relies on a stochastic sampling process, recent studies have suggested that using a simple and deterministic reservoir also provides sufficient performance. Here, we also investigate the impact of using such simple and deterministic reservoir structure within the online learning control framework. Experiments of the online learning control framework conducted on an SMA wire are presented.

A Planning Based Framework for Controlling Hybrid Systems

The control of dynamic systems, which aims to minimize the deviation of state variables from reference values in a continuous state space, is a central domain of cybernetics and control theory. The objective of action planning is to find feasible state trajectories in a discrete state space from an initial state to a state satisfying the goal conditions, which in principle addresses the same issue on a more abstract level. We combine these approaches to switch between dynamic system characteristics on the fly, and to generate control input sequences that affect both discrete and continuous state variables. Our approach (called Domain Predictive Control) is applicable to hybrid systems with linear dynamics and discretizable inputs.

Event-based control of input-output linearizable systems

AbstractThis paper proposes a new event-based control method for nonlinear systems that are input-output linearizable. The control input generator uses a copy of a continuous reference system to generate an exponential control input that keeps the state of the disturbed plant in a bounded surrounding of the setpoint. An upper bound for the deviation of the event-based control loop to the reference system is derived, which depends on the event threshold of the event generator. Hence, by appropriately choosing the event threshold, the event-based control can be made to mimic the continuous control with arbitrary accuracy. The proposed event-based control method is applied to a cooling process.

Synchronization of networked Lagrangian systems with input constraints

AbstractThis paper addresses the synchronization problem of systems modeled by Euler-Lagrange equations and subject to input saturations. First, a new control design strategy, based on virtual systems, is proposed. This approach allows to generate control inputs that are a priori bounded in the presence of communication time-delays, regardless of the information flow topology between systems in the network. Second, we remove the requirement of the generalized velocities, leading to a velocity-free synchronization scheme with a priori bounded inputs. The effectiveness of the proposed control schemes is demonstrated through simulation examples on a network of four robot manipulator arms.

314 可変重力環境下の受動歩行を規範としたコンパス型2足ロボットの歩行速度制御

Recently, biped robots are flourished all over the world. Japan, as a leading country of biped robots, plans to send them into the moon. The project is magnificent but there are lots of problems to operate biped robots. One of the biggest problems is the difference of gravity between the moon and the earth. Therefore, it is necessary to develop control method concerning the gravity. This paper proposes a novel control method to generate gait under variable gravity. Previous research show a method to generate gait based on passive dynamic walking. Proposed method expands the previous study and enables to control robots' step length and velocity on any slope angle under variable gravity. The concept of the proposed method is to generate gait based on natural relativities of passive dynamic walking: the step length depends on the slope angle and the velocity depends on the gravity. To generate control inputs, we find the passive dynamic walking that corresponds to the desired gait in step length and velocity. We apply this referential passive dynamic walking to previous method and then we get control inputs. With proposed method, it is shown that simulation results under variable gravity conform to the desired gait.

Modelling and Control of Pressurized Molten Metal in Press Casting

AbstractThis paper presents a modeling and control of molten metal's pressure in a pressing process using an innovative iron casting developed by our group. In this method, molten metal is directly poured into the lower mold, and then pressed to fill the cavity by the upper mold being lowered down. For the complex liquid flow during pressing, the liquid's pressure change inside vertical path with various contraction and expansion geometries is newly modeled via the unstationary Bernoulli equation. The mathematical model is derived for a control design of pressing. To conduct the pressing velocity design algorithm, an unknown parameter of proposed model considering viscous flow is identified by using CFD: Computational Fluid Dynamics model with heat flow calculation. Control performance using a multi-switching velocity pattern is confirmed as an effective control design using the pressure model, because the pressure fluctuation has discontinuous variation points. Substituting detailed information for mold shape, poured volume and initial temperature into a developed control input generator, an optimum pressing velocity design and a robust design for defect-free production are proposed by the design algorithm based on the construction of an inverse system comprised of the sequential switching from higher to lower speed. Consequently, the effectiveness of the pressing control with reasonable pressure suppression has been demonstrated through the CFD simulation.

Optimum Pressure Control of Molten Metals for Casting Production Using a Novel Greensand Mold Press Casting Method

This paper presents a novel method of sand mold press casting. In this process, molten metal is poured into a drag mold, and then the cope mold is placed on top of the drag mold, and the two molds are matched. The mold design for casting, the pouring control and the velocity control of the press have been previously clarified as key factors in the manufacture of sound products. This paper presents methods for modeling and control of the molten metal’s pressure for novel sand press casting technique. Substituting detailed information for the complex mold shape, the poured volume and initial temperature into a developed control input generator, an optimum pressing velocity design and a robust design for defect-free production are proposed by the sequential control algorithm based on the construction of an inverse system comprised of a sequential switching from higher to lower speed. The effectiveness of adaptive press casting is demonstrated by using CFD model simulations.

Aplicación de Muestreo basado en Modelos de Control Predictivo a un Vehículo Autónomo Subacuático

Unmanned Underwater Vehicles (UUVs) can be utilized to perform difficult tasks in cluttered environments such as harbor and port protection. However, since UUVs have nonlinear and highly coupled dynamics, motion planning and control can be difficult when completing complex tasks. Introducing models into the motion planning process can produce paths the vehicle can feasibly traverse. As a result, Sampling-Based Model Predictive Control (SBMPC) is proposed to simultaneously generate control inputs and system trajectories for an autonomous underwater vehicle (AUV). The algorithm combines the benefits of sampling-based motion planning with model predictive control (MPC) while avoiding some of the major pitfalls facing both traditional sampling-based planning algorithms and traditional MPC. The method is based on sampling (i.e., discretizing) the input space at each sample period and implementing a goal-directed optimization (e.g., A*) in place of standard numerical optimization. This formulation of MPC readily applies to nonlinear systems and avoids the local minima which can cause a vehicle to become immobilized behind obstacles. The SBMPC algorithm is applied to an AUV in a 2D cluttered environment and an AUV in a common local minima problem. The algorithm is then used on a full kinematic model to demonstrate the benefits.

A state-feedback approach to event-based control

This paper proposes a new method for event-based state-feedback control in which a control input generator mimics a continuous feedback between two consecutive event times. The performance of the event-based control system is evaluated by comparing this loop with the continuous state-feedback loop. An upper bound of the difference between both loops is derived, which shows that the approximation of the continuous state-feedback loop by the event-based control loop can be made arbitrarily tight by appropriately choosing the threshold parameter of the event generator.

Fuzzy logic adaptive Kalman filtering in the control of irrigation canals

AbstractA fuzzy logic adaptive Kalman filtering methodology was developed for the automatic control of an irrigation canal system under unknown disturbances (water withdrawals) acting in the canal. Using a linearized finite difference model of open channel flow, the canal operation problem was formulated as an optimal control problem and an algorithm for gate opening in the presence of arbitrary external disturbances (changes in flow rates) was derived. Based on the linear optimal control theory, the linear quadratic regulator (LQR), assuming all the state variables (flow depths and flow rates) were available, was designed to generate control input (optimal gate opening). As it was expensive to measure all the state variables (flow rates and flow depths) in a canal system, a fuzzy logic adaptive Kalman filter and traditional Kalman filter were designed to estimate the values for the state variables that were not measured but were needed in the feedback loop. The performances of the state estimators designed using the fuzzy logic adaptive Kalman filter methodology and the traditional Kalman filtering technique were compared with the results obtained using the LQR (target loop function). The results of the present study indicated that the performance of the fuzzy logic adaptive Kalman filter was far superior to the performance of the observer design based upon the traditional Kalman filter approach. The obvious advantages of the fuzzy logic adaptive Kalman filter were the prevention of filter divergence and ease of implementation. As the fuzzy logic adaptive Kalman filter requires smaller number of state variables for the acceptable accuracy therefore, it would need less computational effort in the control of irrigation canals. Copyright © 2009 John Wiley & Sons, Ltd.

A method for determining control inputs for prescribed accelerations of the end‐effector of a motion‐simulator robot

AbstractDescribed in this paper is an approach for generating control inputs for an industrial Kuka robot such that it produces prescribed linear acceleration time histories for a given point of the end‐effector. The method takes into account the workspace limits, and uses the internal interpolation cycle of the robot controller to generate the corresponding trajectories that will produce the prescribed accelerations. The PI parameters of the internal interpolation are identified from test measurements. The control scheme consists in producing time‐varying values for the desired angle values of the interpolation algorithm, making the changes as the algorithm proceeds and thus generating smooth trajectories. The resulting acceleration of these trajectories is compared to the prescribed accelerations, and an optimization procedure is run such as to determine the required time‐histories of the desired angles for the interpolation algorithm such that both accelerations match. The approach is demonstrated for the example of a 500kg payload Kuka robot that is aimed at producing given motions as a physical motion simulator (e. g. for cars or complex working machines). It is shown that the desired accelerations can be generated accurately and with high repeatability, making the approach suitable for example for simulating car rollover tests. (© 2008 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)

PWM型制御入力に基づく安定化制御

This paper deals with stabilization of continuous-time systems whose control inputs take two distinct values. We consider to generate control inputs based on pulse-width-modulation (PWM) and derive a discrete-time system which governs the state transitions of the continuous-time system at sampling instants. Since PWM control inputs keep pulsating, stability of fixed points for the discrete-time system implies stability of limit cycles for the continuous-time system. Based on the discrete-time system, we show analysis results for the existence of the fixed point, local controllability and local stabilizability. Furthermore, we will propose methods to synthesize locally stabilizing state and output feedback controllers based on LMIs.

An Adaptive Neurocontroller Using RBFN for Robot Manipulators

In recent years, neural networks have fulfilled the promise of providing model-free learning controllers for nonlinear systems; however, it is very difficult to guarantee the stability and robustness of neural network control systems. This paper proposes an adaptive neurocontroller for robot manipulators based on the radial basis function network (RBFN). The RBFN is a branch of neural networks and is mathematically tractable. Therefore, we adopt the RBFN to approximate nonlinear robot dynamics. The RBFN generates control input signals based on the Lyapunov stability that is often used in the conventional control schemes. A saturation function is also chosen as an auxiliary controller to guarantee the stability and robustness of the control system under the external disturbances and modeling uncertainties.

A Stable Adaptive Control for Linear Plant with Unknown Relative Degree

This paper presents a method to realize a tracking type stable adaptive control without knowing the plant’s relative degree. A structure estimator for the unknown relative degree is introduced to the ordinary adaptive control which consists of the parameter estimator and the control input generator. The relative degree is estimated with a multiple identification model using a criterion whether the parameter estimation is worked well or not. The global stability will be ensured by combining the established method and the nature of the criterion. Through simulation study, it is verified that the adaptive control system works well even if the relative degree is changed suddenly.

Realtime Digital Computer State Estimator for a Hard Rock Milling Circuit

This paper describes the development of non-linear stochastic dynamic models of the unit operations in a large, complex milling circuit which are combined to form the basis for building a two-stage extended Kalman filter. The filter provides optimal parameter and state estimates of the operating plant. A description is given of the techniques used for implementing the estimator on a digital computer such that on-line estimates are generated in real time. Sufficient additional real-time computing ability is available to generate control inputs based on the plant estimates. The filter was shown to predict adequately the pebble load in the mills (this is the most useful variable for pebble feed control), water hold-ups, solids hold-ups and particle size distributions around the circuit from measurements of pebble mill power draft, overflow fraction -75 µm and specific gravity, rod mill feed rate and sump discharge specific gravity. The single most important difficulty encountered was the bias introduced by inaccurate modelling of the units and the filter segmentation or “decentralisation” that is necessary because of real-time considerations.

Dynamics of linear systems with semiactive force generators

A semiactive element produces control forces in a dynamic system by controlled variation of a resistive parameter. Control constraints exist making the output of a semiactive element complex and nonlinear even for linear feedback control laws. State equations are derived for a general linear system with an arbitrary number of semiactive elements. These equations are piecewise-linear and exhibit intermittent linear dependencies among elements of the state vector. Two special cases exist yielding simplified descriptions of the semiactive control and system dynamics. Semiactive vibration control is applied to a flexible vehicle on a rough guideway. Numerical results demonstrate performance superior to passive suspensions and comparable to fully-active control. PPLICATION of modern control theory and parameter optimization techniques to shock and vibration control permit the design of high-performance suspension systems for taxiing aircraft1 and ground vehicles.27 Due to their sim- plicity, low cost, and reliability, passive suspensions com- posed of spring and damper elements are used in all but the most critical applications. Although control using active force generators receives considerable attention in the literature, its superior performance over passive control is frequently offset in practical designs by increased cost, complexity, energy requirements, and weight. While passive devices are self- powered and self-contained, an active force generator requires a power source to generate control inputs along with measurement and signal-processing systems. At least the disadvantages of a power supply can be eliminated by parametric control in which the parameters of passive elements are varied in response to a control signal. When a resistive parameter is varied, large control forces in a dynamic system may be generated by controlled dissipation of system energy. This type of control, which has been called semiac- tive, 8 combines features of passive control with the capability of generating the sophisticated feedback controls usually reserved for active systems. Recently, semiactive control has been studied in shock and vibration applications.811 In linear and nonlinear systems with deterministic810 and random inputs,911 these studies have established that the performance of semiactive force generators approaches that of fully active elements. This is not too surprising because any stable vibration-control system must eventually dissipate system vibrational energy, something a resistor does quite well. Although these studies illustrate basic applications and principles of semiactive control, a general treatment of the dynamics of a system with an arbitrary number of semiactive elements is lacking. Semiactive control introduces complex, nonstationary dynamics in even simple, linear systems.8'9 Because control forces are produced by a dissipative process, an inequality constraint on control power exists causing complex behavior of the semiactive control force. In this paper, a general theory of the dynamics of systems with semiactive elements is developed. Two special cases are found that provide sim- plified representations of the semiactive control and system dynamics.