This paper proposes a design method of a robust controller with the ability to adjust control performances adaptively. The controller having a time-varying parameter is based on the parametrization of H∞ controllers. The time-varying parameter takes the place of the so-called Q-parameter to reduce a Lyapunov function of the resulting closed-loop system. This time-varying parameter changes according to the estimations of plant parameters which are derived from a realtime nonlinear calculation with a disturbance observer. The closed-loop system is guaranteed to be quadratically stable.
It is concerned with a multi-objective resource division problem (MRDP), where resources are allocated to activities so as to optimize plural objectives. The genetic algorithm is used to solve the MRDP and the genes are selected using a relative efficiency measure. This approach can easily reflect the preference of decision maker on plural objectives and find the Pareto optimal solutions efficiently. The MRDP is applied to an allocation problem of tools and parts in flexible manufacturing systems, associating with the concept of group technology. Two objectives such as to equalize the processing load for each machining cell and to minimize the number of assigned tools are considered, and some Pareto optimal solutions are provided.
Two types of cognitive tasks experiments, one through visual and the other through auditory presentation form were conducted to examine the correlation between eyeblinks and cognitive process, from the viewpoint of different nature of task phase. Blink rate and eyeblink intervals were analyzed statistically. As far as blink rate is concerned, by counting the number of eyeblinks based on the task phase, the index of blink rate to represent regularity of eyeblink occurrence was presented. Concerning eyeblink intervals, a new analysis approach using trend tests and correlation coefficients was verified. As trend tests for evaluations of statistical independence and underlying trend, reverse arrangement test in addition to conventional run test was applied to the obtained sequences of eyeblinks from the viewpoint of point process. Furthermore, under the guarantee of stationarity of an eyeblink sequence, correlation coefficients were calculated. The experimental results obtained by analysis of the above mentioned indices for eyeblink show that blink activity well synchronizes with the rhythm of the task phase.
This paper describes the measurement of drug density which is proposed for a newly developed ultrasound (US) drug delivery system (DDS). In this DDS, drug is carried in microcapsule and delivered to the lesion of a body. Here, an experiment was performed to estimate a microcapsule density that corresponds to the drug density from the brightness on an US B-mode image. The results indicate that the brightness on a US B-mode image is an index for a microcapsule density and that the our proposed method has the possibility to measure a drug density non-invasively.
Product allocation management in manufacturing and distribution systems is conventionally based on manual operations and requires great time and efforts. Market mechanism solves the product distribution problem by allocating the scheduled resources according to market prices. We formulate manufacturing and distribution system models as a discrete resource allocation problem, and demonstrate the applicability of economic analysis to this framework. The proposed algorithm facilitates a sophisticated product allocation system, which conducts a Pareto optimal solution in the product allocation problem.
In this paper we consider a design problem of robust servo systems with quadratic stability using observer-based output feedback and treat the case where there exist uncertainties both in the state matrix and the input matrix. This robust feedback controller is designed so that (1) the closed loop system is quadratically stable, (2) its output tracks a command step input, and (3) desirable output responses can be achieved. First, we correct those mistakes found in our previous results related to this problem. Next, we derive a design algorithm of robust servo systems with quadratic stability. With this design algorithm a specification of tracking characteristics and improvement of robust stability of the resulting system can be performed almost independently. Finally, a design example is shown based on the algorithm.
In this paper we consider a scheduling problem for a two-machine robotic cell of jobshop type. The problem is an extension of the classical two-machine jobshop scheduling addressed by Jackson 6), but is NP-hard. We propose a greedy heuristic algorithm and test for its performance with numerical experiments.