Vibration Control of a
Multi Degree-of-Freedom System by Swarm Intelligence
岡﨑 菜琳, 石津 諒太, 森下 信
Yokohama National University
A concept of swarm intelligence was applied to an adaptive vibration control system composed of sensor and actuator units and their network. The swarm intelligence may be observed in a group of insects or animals in nature; ants can search the shortest route to food using special pheromone, sardines can avoid being prey by converging, or wolves have high success rate of hunting by allocating tasks, for example. The expression of “swarm intelligence” was first used by G.Beni in the context of cellular robotic systems. In these systems, many simple robot agents may generate self-organized patterns through mutual interactions. In this paper, the sensor units were equipped on a multi degree-of-freedom lumped mass system and an information network was constructed on these sensor units. This network got the learning ability of vibration amplitude at each node in response to the excitation, such as an artificial neural network. Furthermore, the mobile actuator units with dampers included in the information network of sensor units could move to the node with maximum amplitude, and succeeded in suppressing vibration of the lumped mass system in simulation.
The Newmark-β method was used to simulate the vibration of a multi-degree-of-freedom (DOF) mass point system regarded as a structure.
It was confirmed that the vibration of the structure could be suppressed by adaptively adding actuator units consisting of a dynamic damper to the structure.