Nowadays, multicore processor is common presence and available everywhere. However, when we try to speedup non-numerical programs, which are used by ordinary users much especially on both desktop and mobile computers, by introducing parallel computation on multicore processors, it is necessary to strictly keep both control and data dependencies between parallel tasks in order to get the correct computation result. In many cases, parallel computation is difficult for non-numerical programs on multicore processors, since these dependencies are complicated. To solve this problem,we propose a method of speculative thread-level parallel computation along the hot path, the most frequently executed control flow, on entire program, assuming the use of the support hardware of speculative multithreading. This utilizes the general property of program codes, that is, the control flow of program tends to be highly biased to quite a small number of path even if the program code contains many possibilities of paths. Our method divides conventional serial program into thread codes along the hot path, and the thread codes are executed in parallel on each processor core on multicore processor. Preliminary performance evaluation by trace-based simulation using SPEC CINT2000, that is the collection of practical programs for performance valuation purpose, shows that the speedup of 2.48 times at maximum and 1.94 times on average can be expected by using our method, as compared to the serial execution.
This paper addresses state estimation problems for parametric uncertain nonlinear systems with linear measurements. A new Robust Extended Kalman Filter (REKF) which dose not involve augmented systems is devised. The REKF is more effective than the conventional EKF for uncertain systems. However, if the REKF is applied to nonlinear systems without parameter uncertainties, its accuracy will be lower than that of the conventional EKF. Then, we propose an Adaptive REKF(AREKF) to deal with this problem. Furthermore, we modify the predictive step of REKF in order to deal with continuous-discrete filtering problem. The validities of the proposed methods are illustrated in Monte Carlo simulations.
This paper develops a rendezvous control method for a mobile robot system with quantized sensing. First, a new control technique based on dithering is proposed. Then, the performance of the proposed technique is analyzed, which shows the expected deviation of the positions of two robots and an upper bound of the deviation in steady state.
This paper presents a mixed H2 / H∞ balanced realization and applies this to model reduction. Mixed H2 / H∞ balanced realization is a state space description such that solutions of H2 Lyapunov equation and H∞ Riccati equation become diagonal matrices simultaneously. It is shown that H∞ and H2 norms of any reduced model based on this realization are not greater than those of the original system respectively. Upper bounds of H∞ norm and H2 norm of the error system between the original model and the reduced model are also established.
A numerical computation method of the spectrum of the monodromy operator based on fast-sample/hold approximation is investigated. Through the numerical examples of the previous work with the zero-th, 1st and 3rd order polynomial hold functions, it is observed that the computational efficiency improves as the approximation order becomes higher. However, it is reasonable to expect that such a monotonous tendency will hit the ceiling as the approximation order grows. Motivated by this observation, our primary objective in this paper is to investigate the computational efficiency of higher order approximations. Since the order reflects the smoothness of the domain where the monodromy operator is considered, one must justify the approximation procedure for each function space (inductively) before developing the numerical algorithms. Due to the lack of scalability in the proofs of earlier results, this is the first challenge. Then the matrix formula for general order approximation is derived. After checking the error convergence property of the proposed method, we examine the computational efficiency of higher-order approximations through numerical examples.