In this paper, we discuss coordination problems of a group of autonomous agents, including the target aggregation to a convex set and the state agreement. The aggregation of the whole agent group, consisting of leaders (informed agents) and followers, to a given set is investigated with switching interconnection topologies described by the connectivity assumptions on the joint topology in the time interval [t, + infinity) for any time t, and then the state agreement problem is studied in a similar way. An approach based on set stability and limit set analysis is given to study the multi-agent convergence problems. With the help of graph theory and convex analysis, coordination conditions are obtained in some important cases, and the results show that simple local rules can make the networked agents with first-order nonlinear individual dynamics achieve desired collective behaviors. (C) 2009 Elsevier Ltd. All rights reserved.