The particle swarm optimization method (PSO) is one of popular metaheuristic methods for global optimization. Although the PSO is simple and shows a good performance of finding a desirable solution, it is reported that almost all particles sometimes converge to an undesirable local minimum for some problems. Thus, many kinds of improved methods have been proposed to keep the diversity of the search process. In this paper, we propose a novel multi-type swarm PSO which uses two kinds of particles and multiple swarms including either kind of particles. All particles in each swarm search for solutions independently where the exchange of information between different swarms is restricted for the extensive exploration. In addition, the proposed model has the restarting system of particles which initializes a particle with a sufficiently small velocity by resetting its velocity and position, and adaptively selects the kind of the particle according to which kind of particles contribute to improvement of the objective function value. Furthermore, through some numerical experiments, we verify the abilities of the proposed model.
Recently, a robotic hand with tactile sensors is developed all over the world. We also have developed a universal robot hand with tactile sensors and other sensors. Tactile sensors are very important for manipulating objects dexterously. However, array-type tactile sensor has many I/O, thus require much processing time. In this paper, we propose a hi-speed tactile sensing based on the genetic algorithm in order to measure the tactile information rapidly. The validity of the proposed method shows through some experiments. Moreover, a multi-object manipulation according to the tactile information is proposed.
In this paper, a real-time algorithm for nonlinear receding horizon control using the continuation/GMRES method (C/GMRES) is proposed. The continuation method is combined with a Krylov subspace method, GMRES, to update control input by solving a linear equation. In addition, we utilize singular value decomposition, which reduces the size of the linear equation, for the off-line optimization result to speed up on-line calculations, and a barrier function to deal with constraints on the control input. Simulation results show that the proposed algorithm is faster than the conventional C/GMRES.