Abstract: Three kinds of cloud models were designed, which are the nodding times approving cloud, the maximal nodding angle approving cloud and the nodding speed approving cloud. With the guidance of respective positive cloud generation algorithms of the three models, three parameters in a nodding action serial, which are the nodding times, the maximal nodding angle and the nodding speed, were generated. The three parameters follow normal distributions around expected values of the three cloud models respectively. The proposed CBNCA algorithm generates control curves of nodding actions with disciplinary differences between different nodding action serials. In order to describe differences of actions in one nodding action serial, CBNCA+ algorithm was proposed. It can realize uncertainty control of nodding action both between different action serials, and between different actions in one serial. Experiment results prove the effectiveness of the uncertainty control for nodding action of the two cloud-based algorithms. Feature values of the three kind cloud models were proposed.

Key words: robot technology, synthesis of nod, cloud model, uncertainty control, fuzziness, randomness