Embedded and supported vector machines for passability teaching

摘要

PROBLEM TO BE SOLVED: To provide a system and a method capable of adapting an autonomous system such as a robot to a new environment and improving the ability to perform a classification task in the new environment without offline retraining of a neural network. A system 100 having a memory module 140 for storing image data captured by a camera 104 and 102 communicatively coupled to the memory module, wherein an electronic controller is the image data captured by the camera. And implement a neural network that predicts the driveable part of the image data of the environment. Neural networks predict driveable parts of the environment's image data. The support vector machine classifies the predictable driveable part as driveable based on the hyperplane of the support vector machine and determines whether to output a display of the driveable part of the environment. [Selection diagram] Fig. 1