The paper presents a hybrid formalism to analyze a decentralized control policy for a network of agents characterized by a class of non-linear heterogeneous dynamics and biased measurements. Each agent has a continuous-time dynamics and tracks a piecewise constant impulsive reference. The reference jumps depend on the relative distance with respect to some time-varying neighbors. The jumps take place at some updating instants that are decided independently by each agent. Our results provide input-to-state stability guarantees as far as a minimum and maximum dwell-time condition between consecutive updates is satisfied. A numerical example inspired by the formation realization problem for non-holonomic vehicles illustrates the theoretical results.