Yohei Hoshino

Approaches based on particle swarm optimization for problems of vibration reduction of suspended mobile robot with a manipulator

Mobile robots with suspension systems can absorb vibration induced by rough roads, but due to center-of-gravity (CG) shift, the suspended platform is subject to vibration when the platform moves with acceleration. This paper presents approaches based on the particle swarm optimization (PSO) algorithm to overcome the following vibration problems: (1) when the suspended platform moves with a static manipulator, vibration of the suspended platform occurs due to CG shift, (2) when the suspended platform and the manipulator move simultaneously, the vibration is caused by the dynamic manipulator. For the first problem, a method for the optimization of multi-input shapers using PSO is adopted with chaos to reduce the residual vibration. For the second problem, an approach based on PSO with chaos is developed to suppress the vibration by searching for the time-jerk synthetic optimal trajectories of the manipulator. Finally, the authors perform the resulting shapers and optimal trajectories on the presented models and demonstrate the vibration can be controlled to a desired level effectively in both problems.

Map building from laser range sensor information using mixed data clustering and singular value decomposition in noisy environment

This paper presents the results of using mixed clustering (k-means and DBSCAN clustering) with singular value decomposition to build map in a noisy environment from laser range sensor information. Sensors are prone to errors, moreover, environmental and other factors may affect the sensor sensitivity. This may generate a lot of noise which must be removed before building the map. The study shows how density based clustering techniques can, without losing critical information, greatly reduce noise from sensor data. Further applying k-means clustering, the results with various cluster sizes are discussed. Singular value decomposition on the centroids obtained with the k-means clustering is applied to obtain straight regression lines. The %error in the generated maps were analyzed with different sizes of clusters. The experimental results confirmed that the proposed approach can generate accurate maps even in noisy environments.

Slope detection based on orthogonal assumption

This paper presents a method to detect slope by using two-dimensional laser range finder (LRF) for robot navigation tasks in indoor environment. Considering the slope existing indoor environments, ordinary two-dimensional scanning map is not enough since the parameters of the slope and the condition of the floor also become extremely important for robot movement. In our research, one laser range finder (LRF1) is used for the regular horizontal scan while another one (LRF2) is used for the vertical scan. We extracted lines in each vertical scanning section in a rapid way and then distribute points into different classes based on the parameters of their corresponding lines. In order to simplify the algorithm of slope detection and parameters extraction, we introduce orthogonal assumption which supposes the environment around the slope is well structured so that walls and passageways are parallel or vertical to each other. Based on this reasonable assumption, we can extract all the required parameters of the floor and the slope from classified points in two-dimensional way. It is enough for robot navigation and the computational requirement is not high.

Trajectory tracking of wheeled mobile robot with a manipulator considering dynamic interaction and modeling uncertainty

This paper proposes an adaptive control strategy for trajectory tracking of a Wheeled Mobile Robot (WMR) which consists of a suspended platform and a manipulator. When the WMR moves in the presence of friction and external disturbance, the trajectory can hardly be tracked accurately by applying the backstepping approach. For addressing this problem, considering the dynamic interaction, a dynamic model of the system is constructed by using Direct Path Method (DPM). An adaptive fuzzy control combined with backstepping approach based on the dynamic model is proposed. To track the trajectory accurately, a fuzzy compensator is proposed to compensate modeling uncertainty such as friction and external disturbance. Moreover, to reduce the approximation error and ensure the system stability, a robust term is added to the adaptive control law. Simulation results show the effectiveness and merits of the proposed control strategy in the counteraction of modeling uncertainty and the trajectory tracking.

Minimum-order system identification by subspace method and frequency-domain least-squares approximation for vibration isolation tables

A system identification technique to obtain the minimum-order model for a vibration isolation table is studied in this paper. The vibration isolation table is supported by four pneumatic actuators at corners of the square rigid table. A 3DOF mathematical model of the vibration isolation table is constructed. The regulator for the pneumatic actuator is driven by a stepping motor. Proposed procedure is described as follows: First, the state space model is identified from input-output data by subspace method. Next, the low-order curve fitted model is calculated by Frequency-Domain Least-Squares Approximation (FDLSA). And then, the characteristic eigenvalues near the eigenvalues of FDLSA model is picked out from the N4SID model, and the minimum-order system that has the characteristic eigenvalues of N4SID model is reconstructed. And the procedure can be fully automated. Availability and usefulness of this approach is illustrated by experiments.Copyright

Hybrid sliding mode control with optimization for flexible manipulator under fast motion

The modeling and vibration analysis of a flexible manipulator considering a nonlinearity and effect from the gravity imply a singular problem. In order to avoid the singularity, the dynamic equation is decomposed into two subsystems, including flexible dynamic subsystem and rigid dynamic subsystem. A combined feed-forward and feedback control scheme is presented to design the controller of the flexible manipulator. In the combined control, an optimization is applied to obtain the desired trajectory based on the flexible dynamic subsystem. As we know, the optimization is dependent on the accuracy of model but there are inevitably errors of model and external disturbances. The feedback control is expected with high robustness and fast convergence to overcome the problem. In order to improve the performance of control, a hybrid sliding mode control (HySMC) is proposed to track the desired trajectory and further suppress the residual vibration. This paper presents the theoretical derivation and experimental verification of the proposed controller.

Decomposed dynamic control for flexible manipulator in consideration of nonlinearity — Rigid dynamic control

A model of a flexible manipulator is developed with considering the geometrical nonlinearity and the effect of gravity. The model can be divided into a flexible dynamic subsystem and a rigid dynamic subsystem, and a decomposed dynamic control (DDC) including flexible dynamic control and rigid dynamic control is proposed for a controller design of the flexible manipulator. The flexible dynamic control has been investigated for a desired trajectory using an optimization, and the optimization is valid method to deal with nonlinear problems but dependent on the accuracy of models. There are errors in models and other factors causing disturbances inevitably, so that the rigid dynamic control is expected with enough robustness to overcome the uncertain problem. In this paper, the proposed controller does not only track the desired trajectory, but also further suppresses the residual vibration to improve performance of control. A hybrid sliding mode control (HySMC) is proposed to track the desired trajectory and provide a compensator for further vibration suppression. This paper mainly presents the theoretical design and experimental verification of the proposed controller.

Development of Control System by Nonlinear Compensation Using Digital Acceleration Control

Proportional-Integral-Derivative (PID) control has been most commonly used to operate mechanical systems. In PID control, however, there are limits to the accuracy of the resulting movement because of the influence of gravity, friction, and interaction of joints. We have proposed a digital acceleration control (DAC) that is robust over these modeling errors. One of the most practicable advantages of DAC is robustness against modeling errors. However, it does not always work effectively. If there are modeling errors in the inertia term of the model, the DAC controller cannot control a mechanical system properly. Generally an inertia term is easily modeled in advance, but it has a possibility to change. Therefore, we propose an online estimation method of an inertia term by using a system identification method. By using the proposed method, the robustness of DAC is considerably improved. This paper shows the simulation results of the proposed method using 2-link manipulator.Copyright

Apply nonlinear filter ESDS to quantized sensor data

Proportional-Integral-Derivative (PID) control is widely used to control mechanical systems. In PID control technique, however, there are limits to the accuracy of the resulting movement because of the influence of gravity, friction, and interaction of joints caused by modeling errors. Digital acceleration control has robustness for the modeling errors. But it requires position, velocity, and acceleration of a controlled object to construct a controller. In this paper, we use the novel digital differentiator, ESDS. It enables digital acceleration control without increasing the number of sensors. Furthermore, the proposed method works effectively for quantized sensor data. The validity of the proposed method is confirmed by simulations and experiments using 2-link manipulator.

Fuzzy Control of Motion and Force for Flexible Master-Slave Systems

This is the study of the motion, vibration and contact force control of a flexible master-slave system (FMSS). In this study, the master arm is a one-link arm that consists of a rigid body and the slave arm is a one-link arm that consists of a flexible link. Bilateral control based on passivity and a type-1 optimal servo system based on a linear quadratic regulator (LQR) are applied to the system. Fuzzy control is employed to reinforce the advantages of these control methods. A state observer is employed to construct these controllers, however, the accuracy of state estimation deteriorates while the slave-arm is in contact with objects. Therefore improvements in the accuracy of the state estimation are attempted by applying the disturbance-state observer. The performance of the present method is evaluated by experimental results.Copyright