首页>
外国专利>
Machine learning method and apparatus for predicting motion of an object in a robot's environment based on parameters of future robot motion in the environment based on an image capturing the object
Machine learning method and apparatus for predicting motion of an object in a robot's environment based on parameters of future robot motion in the environment based on an image capturing the object
PROBLEM TO BE SOLVED: To predict the movement (s) (if present) generated with respect to an object (s) in an environment of a robot in response to a specific motion of the robot in the environment. Provides a method and device for deep machine learning capable of. Deep to predict at least one transformation (if any) of an image of a robot's environment that should occur as a result of performing at least a portion of a particular movement of the robot in the environment. -Aimed at training neural network models. Deep neural network model training makes it possible to predict transformations based on inputs that include images and robotic motion parameters that define part of a particular motion. [Selection diagram] Fig. 1
展开▼
