Control For Self-organized Collective Motions Of Swarm Robotic Systems

The creatures in nature live mostly in the form of self-organized collective motions, the significance of which is to achieve the purpose of growing and multiplying with the help of the group advantages including the information interaction between individuals, protection against invasion, higher foraging efficiency and low energy consumption, etc. Currently, cooperative control of multi-robotic system have been discussed to address the various actual issues of collective behaviors in engineering. However in the case of collective system consisting of large number of individuals in the smaller scale, the existing control strategies cannot perfectly meet the actual demand of collective motions, such as random switches of interconnection between individuals and cancellation and incorporation of special individuals as well as self-adaptation to complex environments. And the complicated designs for interconnection between individuals will encounter some issues about destabilization of interconnection and explosion of information, which would cause the failure of large scale collective motion. In contrast, the self-organized collective control strategy shows the unique advantages for large scale collective control with good adaptability, redundancy and fault tolerance and without communications or computing bottleneck caused by centralized data processing, because of excellent characteristics, such as equality, independence and no need to one-to-one correspondence in individual relationship as well as arbitrary cancellation and incorporation.In our research, two representative collective systems including macroscopic agent system and micro/nano robotic system are introduced as the objects of the study and the self-organized collective control strategy based on Loose Preference Rule (LP-Rule) and interaction with magnetization and microflow are respectively modelled to study characteristics of collective behaviors, such as synchronization property, cooperative behavior, intelligence performance and emergent properties, etc. The collective motions are divided into the aggregative and scattered behaviors in the direction and vertical direction to the target, respectively. With the role of the uniform targeted guiding and interconnection between individuals as the main line, the main work and innovations in this thesis are as follows.First, taking the interactive relationship-two nearest neighbors as interconnected objects, 2D line (chain)-shaped LP-Rule is modeled. With the help of uniform targeted guiding, formation control, cooperative handling and hunting are selected as the collaborative tasks of large scale collective system to be achieved by the self-organized control strategy based on LP-rule.Next, the interactive relationship is further expanded from the planar line type to three dimensional network type, and is only considered to the effects of collective motion by removing the attractive effects to individuals from the target. In the present of LP-rule based on topological interconnection, the emerging condition and operating mechanism of spiral-shaped collective motion are analyzed systematically by changing the initial position, interconnection and the number of individuals in the group, and then it is proven that the behavior emergence is able to be controlled effectively by the topological trap which occurs with increasing and deceasing of the number of individuals.Then, for collective motion control of micro/nano robot system, without any interconnections but only with external uniform targeted magnetic field, the synchronous movements of large scale helical microrobots on the plane are achieved by improving the fabricating process and the control signals of the external rotating magnetic field, and micro-manipulations are done by the control of collaboration between individuals with different swimming performances.Finally, aiming at targeted cargo delivery by the magnetic controlled microcarrier system, the issues on adhesion in the process of collective movement was solved by established the interaction of magnetic attraction and fluid repulsion, therefore the self-organized behaviors of large scale helical microrobots are discussed further in the present of the external uniform targeted magnetic field.The researches in this paper indicate that 1) the collective control strategy based on the self-organized rules can be effectively employed to achieve the cooperative tasks of large scale systems; 2) the interactive relationship among agents determine behavior emergences of helical shaped trajectory in collective self-organized motions, which can be controlled to re-organize the emerged behaviors by "topological trap"; 3) the external magnetic field has been proven to be effective at controlling the synchronized motion and cooperative micromanipulations, and 4) the establishment of interaction with magnetization and microflow makes the self-organized targeted drug delivery using swarms of microhelices in microfluidic channel possible.