Coordination difficulties of individuals who do not know when, where or how to act in a collaborative situation are frequent. However, with the development of service robots and teammates, it is necessary that a human and a robot can engage naturally and effectively in a joint activity in order to reach a common goal.To this purpose, we propose a theoretical review of the main mechanisms involved in the understanding and achievement of joint activities in children. More specifically, we aim to identify the different difficulties encountered in human-robot situations of collaboration, and how we can respond to them through developmental psychology by considering the different steps of development of joint attention or multimodal communication capabilities. By allowing the emergence and maintenance of mutual understanding between agents, these elements must indeed be better understood in order to improve human-robot joint action.