摘要

Self-replication(or self-reproduce) is an ultimate way for machine system to long-termly survive and be useable. Current self-replication motion planning are mainly based on evolved algorithm, since the replication process are too complex for ordinary method to deal with. However, we argue that deterministic algorithm should also be considered in some extent, especially for those robotics whose kinematic laws of motion could be modeled. As the first step, we proposed a designed self-replication motion strategy for 2-D modular robotics, after its motion laws are analyzed and modeled. The experiments show that the proposed approach outperformed the traditional evolved method both in search time and search space.

  • 出版日期2011

全文