Improved poly-clonal artificial immune network for multi-robot dynamic path planning

Abstract

The most challenge of dynamic path planning lies in that the high unpredictability of environmental information. With the strong space search ability and learning ability, artificial immune network (AIN) has been used for path planning. Polyclonal artificial immune network (PCAIN) solves the problems of immature convergence and local minima with the increasing diversity of antibodies. In this paper, we propose improved polyclonal artificial immune network (IPCAIN) for multi-robot path planning with moving obstacles and moving goals in unknown environment. The antibody concentration is computed with taking other robots and moving obstacles into account. Moreover, memory units are used for preserving antibodies in the specific situations. The memory ability increases the initial concentration of specific antibodies, thus, reduces the response time for dynamic path planning. Extensive simulation experiments validate the proposed method can search the optimal path for multiple robots in dynamic unknown environment.