.png){kind=icon}
返回
Posture-based motion planning: Applications to grasping
.png){kind=icon}
PRE
AI
DOI：10.1037//0033-295X.108.4.709.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
This article describes a model of motion planning instantiated for grasping. According to the model, one of the most important aspects of motion planning is establishing a constraint hierarchy-a set of prioritized requirements defining the task to be performed. For grasping, constraints include avoiding collisions with to-be-grasped objects and minimizing movement-related effort. These and other constraints are combined with instance retrieval (recall of stored postures) and instance generation (generation of new postures and movements to them) to simulate flexible prehension. Dynamic deadline setting is used to regulate termination of instance generation, and performance of more than one movement at a time with a single effector is used to permit obstacle avoidance. Old and new data are accounted for with the model.
Keyword:
EQUILIBRIUM-POINT CONTROL
CORIOLIS-FORCE PERTURBATIONS
ARM MOVEMENTS
TRAJECTORY FORMATION
FUNCTIONAL-ORGANIZATION
NEURAL MODEL
MOTOR
PREHENSION
COORDINATION
ACCURACY
