Fast Convergence of Multi-quadrotor Cooperation Using Weighted-Neighbor-based Control

Abstract

SUMMARY In this paper, a weighted-neighbor-based cooperation control of multi-quadrotor systems is investigated. A formation tracking problem is treated, where the reference formation trajectory (RFT) is not given a priori. The RFT is only available to some of the quadrotors (i.e. the leaders). In order to attain the fast convergence of the agents, we propose an algorithm to calculate the neighbors’ weights in decentralized way. Then, the weights are used to compose the formation controller. Compared to the widely used average-neighbor-based control method, the proposed control protocol can increase the convergence speed of the cooperation error. Since the formation control is improved in topological scale, the utilization of the proposed algorithms can be extended on any multi-robot systems. We show the improvement of the proposed control protocol by theoretical proof, simulation, and real-time experiments.

FAT-based robust adaptive control of cooperative multiple manipulators without velocity measurement

“…For example, high speed and accuracy make industrial manipulators widely used in the production processes. With the advancement of technology and the need to increase the quality of products, cooperating robots, including arms, mobile robots, or UAVs, have become the center of attention [1][2][3][4]. Performing industrial tasks such as welding, transportation, painting, packaging, and assembly requires high maneuverability, so they cannot be done by a single robot and require the use of a team of robots.…”

Consensus in networks of uncertain robot manipulators without using neighbors’ velocity information

“…Consensus means that all agents in a network reach an agreement in some qualities. Nowadays, consensus problem has found applications in variety fields such as spacecrafts attitude synchronization [1], frequency synchronization in power grids [2], and unmanned aerial vehicles cooperation [3,4]. Primary works in this subject consider a simple linear dynamical model for agents with a single integrator [5] or double integrator [6].…”