Swarm Robot Multitarget Search Strategy Based on Triangular Cones in a Complex Dynamic Nonconvex Obstacle Environment

Previous studies have not extensively researched swarm robot multiobjective search in a complex environment sunch as unknown dynamic nonconvex (UDNC) obstacle environment to address collision prediction and low search efficiency. In addition, research on unnecessary excessive robot rotation and the loss caused by rotation during movement has not received enough attention. In this paper, a multitarget search strategy of a robot swarm based on triangular cones (MSTC) is proposed to satisfy the previously mentioned challenges. First, to address the problem of obstacle avoidance in a UDNC environment, an expanding triangular cone (ETC) method is proposed. This method not only considers the size of the robot but also enables the robot to avoid obstacles and other robots effectively and safely. Second, a roaming direction cone (RDC) and a target search cone (TSC) are proposed to improve the robot’s search ability. The RDC improves the ability to exploit the global unknown environment, while the TSC effectively enhances the local search capabilities. Furthermore, the three methods can avoid the time consumption, energy consumption and mechanical loss due to unnecessary direction rotation through a simple geometric relationship to improve the robot service life. Finally, the three proposed methods are combined to construct an MSTC strategy to simulate the multitarget search process of swarm robots. Simulation experiments are carried out using a simplified virtual force (SVF), a straight-line search method different from the nearest neighbour (LSDN), and a particle swarm optimization algorithm with kinematic constraints (KCPSO). The results show that the time consumption, the road consumption, the degree of rotation and the number of rotations have all made significant improvements.