Assist-as-needed control is underlain by the aim of replacing skillful therapists with rehabilitation robots. The objective of this research was to introduce a smart assist-as-needed control system for the elderly or partially paralyzed individuals. The main function of the proposed system is to assist the patients just in the required sub phases of the motion. To ensure that a smart and compliant system is developed, the target admittance gains of the controller was adapted according to the concept of energy The admittance gains were modified so that an exoskeleton reduces interaction energy in cases wherein users have sufficient strength for task execution and maximizes the interaction energy in the required subphases. The results of simulations and an experimental investigation on a novel exoskeleton showed that the proposed adaptive admittance control improves performance to a level substantially higher than that achieved with constant impedance control. Copyright © Cambridge University Press 2017.