Neural Network For Feature Recognition in Automated Process Planning

The objective of this project is to study on the integration of Neural Network in Computer Aided Process Planning especially in feature recognition application. An introduction about process planning will be covered in brief. Following this, the concept of feature shall be discussed as an integral part in CAPP. The next section presents one of the feature recognition methods developed by Nezis and Vosniakos. The strategy their approach is to use a Neural Network system to recognize feature from a sub-graph. The sub-graph is decomposed from an Attributed Adjacency Graph of a solid model. In the last section of this report, a radial basis function network is proposed to substitute the current Neural Network type. Although, the feature recognition method presented in this report is not a new, it is still worthwhile to discuss as there are not many publications exploited this method. In addition, the advantages of the proposed NN over the current one seem to give a new atmosphere in this field. Thus, this can be a starting point and stimulator for further researches.