| 摘要 |
Target motion and an instantaneous value of a target floor reaction force are determined based on a deviation between a target state amount relating to a posture of a robot (1) at least about a vertical axis or about an axis normal to a floor surface and an actual state amount of the robot (1), and on an allowable range of a to-be-limited amount that is a vertical component of floor reaction moment to be applied to the robot (1) or a component of the floor reaction moment, which component is in the direction normal to the floor surface. The determination is made, while keeping a to-be-limited amount corresponding to a target floor reaction force in an allowable range, such that the difference between floor reaction moment that balances in a dynamic model with target motion and floor reaction moment of the target floor reaction force causes the deviation to be closer to 0.
|
