Functional electrical stimulation (FES) has been considered to be an effective technique for aiding quadriplegic persons. However, the musculoskeletal system is highly nonlinear and varies with time; moreover, it is subjected to fatigue and has variable reflex characteristics. It is also difficult to stably and accurately control the limbs using FES. In this paper, we propose a simple FES method that is consistent with the motion control mechanism observed in humans. We focus on joint motion by a pair of agonist-antagonist muscles of the musculoskeletal system, and define the “electrical agonist-antagonist muscle ratio (EAA ratio)” and “electrical agonist-antagonist muscle activity (EAA activity)” in the light of the agonist-antagonist muscle ratio and agonist-antagonist muscle activity, respectively, proposed to extract the equilibrium point and stiffness from Electromyography (EMG) signals. The agonist-antagonist muscle ratio and agonist-antagonist muscle activity are based on the theory that the equilibrium point and stiffness of the agonist-antagonist motion system controlled by the central nerve system. We derived the transfer function between the input EAA ratio and force output of the end-point; this transfer function model is expressed as a cascade-coupled dead time system and secondary delay system. Through the agonist-antagonist muscle stimulation pattern determined by this transfer function model, high-speed, high-precision, and smooth control of the hand force were achieved.

抄録全体を表示

This paper concerns the design problem of discrete-time Gain-Scheduled Output Feedback (GSOF) controllers, in which inexact scheduling parameters are presupposed to be provided, for discrete-time Linear Parameter-Varying (LPV) systems. An iterative algorithm, which has already been proposed for the same problem with unstructured disturbance input and performance output, is revised to accommodate the structures of those signals by introducing constant scaling matrices for the reduction of conservatism. Flight controllers for the lateral-directional motions of an airplane are designed by using our method and a design method without consideration of the uncertainties in the provided scheduling parameters. Flight tests using the two controllers confirm that the GSOF controller with consideration of the uncertainties realized model-matching as well as disturbance suppression as assured in its design; however, the GSOF controller without consideration of the uncertainties could not achieve the assured control performance in its design.

抄録全体を表示

In this paper, we consider a stabilization problem of a mobile robot in an unknown plane at an unknown destination. For the problem, we propose a reinforcement learning based feedback controller design method on the unknown plane. We can design a semi-global CLF with an optimal action-value function obtained by reinforcement learning and local CLFs via multilayer minimum projection method. The feedback controller is designed by the semi-global CLF. The effectiveness of the proposed method is shown by computer simulation.

抄録全体を表示

This paper considers the problem of controlling both the position and the attitude of a quad-rotor helicopter. The quad-rotor helicopter is described by a set of nonlinear equations and some parameters of the dynamics are subjected to uncertainties. This paper proposes a new input-output linearization method based on an adaptive law, and presents a new adaptive tracking controller with on-line parameter estimation to make the tracking error zero asymptotically for the position and the yaw angle. The design problem of the controller is reduced to that of a linear system, and the design method is simple and straightforward. Moreover, a disturbance sensitivity reduction problem is also solved by applying a standard H-infinity control method. A numerical simulation is performed to evaluate the effectiveness of the proposed controller.

抄録全体を表示

In many engineering problems including control problems, optimization of the policy for multiple conflicting criteria is required. However this is very challenging if there exist noise, which may be input dependent, and/or the restriction in the number of evaluations, which is induced in the case where the experiments are expensive in time and/or money. This paper presents a multiobjective optimization (MOO) algorithm for expensive-to-evaluate noisy functions. By incorporating a heteroscedastic Gaussian process regression method as well as standard Gaussian process regression, the algorithm creates suitable surrogate functions from noisy samples and finds the point to be observed at the next step. This algorithm is compared against an existing MOO algorithm, and then applied to optimize the sidewinding gait of a snake robot.

抄録全体を表示

X-by-wire systems have brought many significant benefits, such as improvement in fuel consumption, pre-crash safety, crash safety, vehicle architectural flexibility, and so on. Therefore, a replacement of all mechanisms of a vehicle by X-by-wire, such as steer-by-wire, shift-by-wire, and brake-by-wire, has been attempted. However, there are many issues with safety if a failure occurs. The difficulty of securing safety prevents X-by-wire products from being popularized. In this paper, we propose a design of a fault tolerant robust control system which is capable of mitigating steering and drive-train system failures. By utilizing the steering system and the drive-train system complementarily, the proposed method does not need an additional redundancy for the fault tolerant purpose. For the steering system failure which should be detected and managed rapidly, the proposed method utilizes the disturbance observer idea to eliminate the fault detection. On the other hand, for the drive-train system failure, a fault detection is equipped. The effectiveness of this method has been verified in simulation.

抄録全体を表示

An attitude of a rigid body is defined on SO(3); the fact leads to problems on attitude control of a rigid body. We propose a new discontinuous control law for globally asymptotic stabilization of the attitude of the rigid body by minimum projection method. We prove that the system always has Carathéodory solutions under our control method. Moreover, we show an advantage of the proposed method over quaternion-based hybrid feedback.

抄録全体を表示