Dynamic Control Of Robot Manipulators Research Articles

Deep adaptive control with online identification for industrial robots

Derivation of control equations from data is a critical problem in numerous scientific and engineering fields. The inverse dynamic control of robot manipulators in the field of industrial robot research is a key example. Traditionally, researchers needed to obtain the robot dynamic model through physical modeling methods before developing controllers. However, the robot dynamic model and suitable control methods are often elusive and difficult to tune, particularly when dealing with real dynamical systems. In this paper, we combine an enhanced online sparse Bayesian learning (OSBL) algorithm and a model reference adaptive control method to obtain a data-driven modeling and control strategy from data containing noise; this strategy can be applied to dynamical systems. In particular, we use a sparse Bayesian approach, relying only on some prior knowledge of its physics, to extract an accurate mechanistic model from the measured data. Unmodeled parameters are further identified from the modeling error through a deep neural network (DNN). By combining the identification model with a model reference adaptive control approach, a general deep adaptive control (DAC) method is obtained, which can tolerate unmodeled dynamics. The adaptive update law is derived from Lyapunov’s stability criterion, which guarantees the asymptotic stability of the system. Finally, the Enhanced OSBL identification method and DAC scheme are applied on a six-degree-of-freedom industrial robot, and the effectiveness of the proposed method is verified.

Model Predictive Control With Integral Compensation for Motion Control of Robot Manipulator in Joint and Task Spaces

The motion control of robot manipulators is a crucial problem concerning automatically controlled robots. In this work, the model predictive control method with an integral compensation (MPC-I), which compensates for the matched uncertainties due to unmodeled dynamics, is proposed to solve the trajectory tracking problem of robot manipulators in joint and task spaces. First, this paper decouples the joint variables of the robot manipulator using a computed torque control method. The MPC-I method is, thereafter, derived to realize the motion control of the robot manipulators in joint space. To realize the motion control of the robot manipulator in task space, the task space is, thereafter, converted into the joint space, in which the MPC-I method is executed, afterward, to control the robot. Furthermore, an MPC-I variation, in which the inverse kinematics is calculated indirectly, is proposed to achieve the motion control in task space. The novelty of this paper is to propose the MPC-I method and the method of converting task space to joint space with indirect inverse kinematics calculation. The former is suitable for the dynamic control of the robot manipulators in the joint space, and the latter can extend the MPC-I method to dynamic control in task space. To evaluate the performance of the proposed control method, motion control simulations are performed in the task and joint spaces, respectively. Simulation results and comparisons verify the effectiveness of the proposed control approach for the dynamic control of the UR5 robot manipulator.

Simulation of Manipulator with Flexible Joint

Joint flexibility is the key factor during dynamic control of robot manipulator. Accurate dynamic model is the fundamental of manipulator system design, analysis and control. This paper adopts Lagrange method to accomplish two degrees freedom manipulator modeling, and then design Backstepping control law according to a single-link manipulator. For the above control law, the proof of the Lyapunov stability is given and simulations are done. The simulated result suggested that the static error is decreased.

Controlling robot manipulators by dynamic programming

A certain number of considerations should be taken into account in the dynamic control of robot manipulators as highly complex non-linear systems. In this article, we provide a detailed presentation of the mechanical and electrical implications of robots equipped with DC motor actuators. This model takes into account all non-linear aspects of the system. Then, we develop computational algorithms for optimal control based on dynamic programming. The robot's trajectory must be predefined, but performance criteria and constraints applying to the system are not limited and we may adapt them freely to the robot and the task being studied. As an example, a manipulator arm with 3 degress of freedom is analyzed.

再構成可能並列プロセッサと知能ロボット制御への応用

In the sensor feedback control of intelligent robots, the delay time must be reduced for a large number of arithmetic operations. In addition to many multiply-additions, the arithmetic operations such as absolute value calculation and maximum value selection are often used in the dynamic control of robot manipulators. To reduce the delay time for multi-operand arithmetic operations, the architecture of the reconfigurable parallel processor is proposed. In each processor element, a switch circuit is used to change the connection between the multipliers, adders and arithmetic units (AUs), so that the multi-operand AUs having desired numbers of operands can be reconfigured. Since the data transfer is accomplished by the direct connection between the multipliers, adders and AUs, the overhead for data transfer is reduced. The chip evaluation based on 0.8 [μm] CMOS design rule shows that the delay time for differential inverse kinematics computation and inverse dynamics computation of a six-degrees-of freedom manipulator become 1.7 [μsec] and 11.5 [μsec], respectively.

Parallel computation of robot inverse dynamics for high speed motions

The computation of the highly coupled dynamic equations has always posed a bottleneck in real-time dynamic control of robot manipulators. Recent advances in VLSI technology make it possible to implement new algorithms that compute these equations and meet real-time constraints. Parallel processing techniques can now be used to reduce the computation time for models of a highly mathematical nature such as the dynamical modelling of robot manipulators. In this paper an attempt is made to review the subject of the parallel computation of robot dynamics. Moreover, a new and highly efficient technique is introduced to solve this problem. A simplified form of the Lagrange-Euler is divided into subtasks and distributed on to a parallel processing system. The development (parallel processor) system employs several INMOS transputers running the OCCAM concurrent programming language. Further, the system is used to introduce parallelism to robot dynamics through different task allocation strategies. These strategies flow naturally from the Lagrange-Euler formulation. The cost effectiveness and speed of the algorithm are demonstrated by a case study (Stanford arm). Comparisons are made between uniprocessor (von Neumann) and parallel implementations of the algorithm. Several measures such as utilisation, efficiency, and speed up are used to evaluate the performance of the employed networks and task allocations.

Application of DSP to real-time computation for dynamic control of robot manipulators

In this paper, the application of a DSP (digital signal processor) to real time computation for low-cost dynamic control of robot manipulators is proposed. In order to realize dynamic control in re...

Application of a General Learning Algorithm to the Control of Robotic Manipulators

In this paper, we discuss the use of a general learning algo rithm for the dynamic control of robot manipulators. Unlike some other learning control schemes, learning is based solely on observations of the input-output relationship of the system being controlled and is independent of control objectives. Information learned previously can be applied to new control objectives as long as similar regions of the system state space are involved. The control scheme requires no a priori knowl edge of the robot dynamics and is easy to apply to a particu lar control problem or to modify to accommodate changes in the physical system. The control scheme is computationally efficient and well suited to fixed-point implementation. The learning controller is evaluated in a series of computer simu lations involving a two-axis-articulated robot arm during simulated repetitive and nonrepetitive movements. We inves tigate the effects of varying learning algorithm parameters as well as control system performance in the presence of obser vation noise and changing manipulator payloads. The learn ing control system presented promises to provide good dy namic performance in complex situations at a reasonable cost as measured in terms of both hardware and software devel opment.