Behavioral coordination, structural congruence and entrainment in a simulation of acoustically coupled agents.

Social coordination is studied in a simulated model of autonomous embodied agents that interact acoustically. Theoretical concepts concerning social b ehavior are presented from a systemic perspective and their usefulness is e valuated in interpreting the results obtained. Two agents moving in an unst ructured arena must locate each other, and remain within a short distance o f one another for as long as possible using noisy continuous acoustic inter action. Evolved dynamical recurrent neural networks are used as the control architecture. Acoustic coupling poses nontrivial problems lit discriminati ng 'self' from 'non-self' and structuring production of signals in time so as to minimize interference. Detailed observation of the most frequently ev olved behavioral strategy shows that interacting agents perform rhythmic si gnals leading to the coordination of movement. During coordination, signals become entrained in an anti-phase mode that resembles turn-taking. Perturb ation techniques show that signalling behavior not only performs an externa l function, but it is also integrated into the movement of the producing ag ent, thus showing the difficulty of separating behavior into social and non -social classes. Structural congruence between agents is shown by exploring internal dynamics as well as the response of single agents in the presence of signalling beacons that reproduce the signal patterns of the interactin g agents. Lack of entrainment with the signals produced by the beacons show s the importance of transient periods of mutual dynamic perturbation wherei n agents achieve congruence.