ランダムサンプリングロボットシステムの安定性

抄録

We consider control issues of robotic systems composed of several subsystems which are connected one another through some communication channels. Such complex systems involve random delays due to communications in addition to casual ones, for example, in measuring and signal processing. Thus we have randomly sampled robotic systems.
In this paper, it is assumed that computed torque controllers are used for randomly sampled robotic systems and the asymptotic stability with probability one is studied. First, an approximate equation of error vectors from desired trajectories are derived. Using the equation, it is shown that randomly sampled robotic systems is always stabilizable by computed torque PD controllers and PID controllers if the distribution of sampling intervals are bounded. It is also shown that conventional critical damping controllers are robust for uniform distributions in terms of the stability, but they are not so for wide distributions in the communication. Hence a design method is proposed to select feedback gains in order to make the system asymptotically stable with probability one under given distribution of sampling intervals. Experimental results are also shown.