Self-Organizing Feature Maps for Modeling and Control of Robotic Manipulators

摘要: This paper presents a review of self-organizing feature maps (SOFMs), in particular, those based on the Kohonen algorithm, applied to adaptive modeling and control robotic manipulators. Through number references we show how SOFMs can learn nonlinear input–output mappings needed manipulators, thereby coping with important issues such as excess degrees freedom, computation inverse kinematics dynamics, hand–eye coordination, path-planning, obstacle avoidance, compliant motion. We conclude arguing that be much simpler, feasible alternative MLP RBF networks for function approximation design neurocontrollers. Comparison other supervised/unsupervised approaches directions further work field are also provided.