Recently, an architecture called multiprocessor system-on-a-chip (MP-SoC) is paid much attention to. MP-SoC is composed of several processing elements (PEs) interconnected with a network-on-a-chip. It is important for MP-SoC not only to reduce the computational time of the executed task but also to avoid the rise of the temperature. In this paper, we formulate an optimization problem of clock frequencies of PEs where an objective function evaluates both the computational time and the temperature of MP-SoC. We transform the optimization problem into a potential game and propose a distributed algorithm named a round-robin update algorithm for optimizing the frequencies of PEs. By using simulations, we also compare the proposed algorithm with an existing algorithm and show advantages of the proposed algorithm.
This paper proposes a remote control method for a linear system with pointwise-in-time state and/or control constraints. Our technique guarantees achievement of tracking performance and constraint fulfillment under simultaneous occurrence of some phenomena over communication links, such as arbitrary and unforeseeable delays and lots of packet losses. Our ideas are to feedback information about not a state but a state-existing area and to share maximal output admissible sets with reference governor and switching control strategy that consist of the proposed technique. Through numerical and experimental examples, we demonstrate effectiveness of the proposed method.
This paper aims at demonstrating the effectiveness of discrete-time noncausal linear periodically time-varying scaling through control experiments with a cart inverted pendulum whose pendulum length can be set from three different values. We regard the variation in the pendulum length from its middle value as an uncertainty, which leads to a structured uncertainty on the plant model. By applying a robust performance synthesis method based on the above scaling technique to an appropriately constructed generalized plant, we demonstrate that we can successfully design a single linear periodically time-varying controller attaining high control performance regardless of the length of the pendulum.
In this paper, we construct a nonlinear reduced order model of a plasticization cylinder of injection machines, which has a spatially distributed nonlinear dynamics. First, a distributed parameter model for the overall control system is derived based on the physical laws. Next, we attempt to reduce the model complexity focusing on the specific structure of the nonlinearity, which arises from a temperature-dependency of the rate of heat loss of heaters subjected to natural convection. This enables us to obtain a 28 dimensional model (via an 808 dimensional spatially discretized model) while theoretically guaranteeing a practically satisfactory accuracy. The obtained model is examined experimentally by using a prototype system.