Design of consensus protocol for nonholonomic systems under directed communication topology

Abstract

In the real world, many physical systems are subject to nonholonomic constraints. In this paper, the problem of output consensus is investigated for such kind of systems in chained form. First, the consensus controller is developed for the strongly connected topology using the backstepping design technique. Then, the result is extended to the general directed topology via graph decomposition where consensus achievement among different strongly connected components is analyzed by means of input-to-state stability (ISS). Finally, a simulation example is provided to illustrate the effectiveness of the proposed controller.

Position centroid rendezvous and centroid formation of multiple unicycle agents

“…In this subsection, by adding bias term as in [14,26], we extend the obtained centroid rendezvous strategy to formation control of unicycle agents. The formation control law is designed as…”

“…For the cooperative control of general non-holonomic agents, Dong and Farrell presented systematic methods and control algorithms in [13,14] to ensure the entire state of each agent tracks a moving target point, and analysed the effects of time-delays on the algorithms. For the output consensus of general non-holonomic agents, Xu et al [26] combined the graph decomposition and the input-to-state stability theory to construct a dynamic control law such that the output of each agent converges to a common value and the internal states tend to zero. Different from [1,13,14,[18][19][20][21][22]26] where the control law is discontinuous or time-varying, the static smooth time-invariant distributed laws are designed in [6] and [22,27] to asymptotically achieve rendezvous for non-holonomic chained agents and unicycle agents, respectively.…”

“…For the output consensus of general non-holonomic agents, Xu et al [26] combined the graph decomposition and the input-to-state stability theory to construct a dynamic control law such that the output of each agent converges to a common value and the internal states tend to zero. Different from [1,13,14,[18][19][20][21][22]26] where the control law is discontinuous or time-varying, the static smooth time-invariant distributed laws are designed in [6] and [22,27] to asymptotically achieve rendezvous for non-holonomic chained agents and unicycle agents, respectively. Moreover, an image-based law [28] was constructed to drive the non-holonomic agents to a desired pattern by using an uncalibrated flying camera homograph.…”

“…The existing control laws have two categories: 'leader-following control laws' [1,2,[13][14][15][16][17] and 'non-leader ones' [18][19][20][21][22][23][24][25][26][27][28]. The leader-following control laws are designed such that the entire/portion states of each agent track the reference trajectory, sometimes with a desired team shape.…”

Nearest Neighbor-Based Rendezvous for Sparsely Connected Mobile Agents

“…Almost all the work on rendezvous for nonholonomic mobile robots seems specific for the unicycle model [20][21][22][23]. The unicycle model is an oversimplification used in modeling nonholonomic agents.…”