Study On Distribution Formation Control And Modeling Of Robotic Swarm

Swarm robotics is a new interdisciplinary field inspired by biological swarm intelligent (SI) behaviors. The study content is very extensive covering coordination behavior control, system modeling and actual system design and so on. The extendable study also includes the individual's self-adaptive ability, communication design between the members and so on. The final purpose is to develope the robotic swarms containing large numbers of simple robots where the global behaviors will emerge through individuals' local interactions. Robotic swarms have many potential applications in military war, industrial production, health care and environmental protection, and daily life.To address the problems of formation control and the lack of system modeling in the status of swarm robotics, a part of the work is conducted here. The main work and innovative results are as following:(1) Following the principle that individual's behavior should have the tendency to benefit for achieving overall mission, individual's motion during forming E structure is divided into two stages of the adjustment process and cohesive process. Interaction control algorithm (TFS) is presented for three neighboring individuals to form E structure and the stability of E structure formation is analyzed. Simulation results show that three neighboring individuals can achieve the E structure through TFS algorithm.(2) The local interaction control algorithm (L-TFS), the extension version of TFS algorithm, is proposed for making robotic swarm to form multi-E network. Becase of exsting holes in final multi-E network under L-TFS algorithm, the modified local interaction control algorithm (ML-TFS) improved from L-TFS is proposed to achieve uniform multi-E network. The stabilities of (uniform) multi-E network formation of robotic swarms using (M)L-TFS are analyzed. Simulation results show that robotic swarms using (M)L-TFS can form (uniform) multi-E network from random initial positions and ML-TFS is scalable and robust for uniform multi-E network formation. (3) The control equation (RTFS) for RT formation of four neighboring robots in3D space is proposed from the inspiration of virtual spring mesh principle. The stability of RT formation by four neighbors using RTFS is analyzed. Furthermore, multi-RT network formation of robotic swarm in3D space is realized using the extended TRFS. Single channel communication of Propeller chip based on point-to-point is extended to multi-channel communication for underwater robotic swarms, and the proper transmission rate of laser in general water body (sea water or flash water) is resulted.(4) To improve the local mimima distribution in traditional environment potential field, an environment potential function with multiplication and addition configuration is proposed based on two kinds of simple goal and obstacle potential functions. The analysis shows that the proposed environmental potential field still can ensure robotic swarms to achieve stable aggregation. Some typical cases easily causing local minimas are considered and accordingly the rangs of proper potential function parameters are resulted.(5) Based on the task background of swarm collaborative foraging, the swarm mathematical models for assistance and division-work foraging are established respectively. Simulation results show the effectiveness of the proposed mathematical models. The dynamic properties of individual numbers along with various different states accordingly, and the effects of related parameters and control strategies to collaborative behaviors of swarm foraging are also analyzed.Finally, the results of the full text are summarized, and the prospect of future work is conducted.