Autonomous decentralized control has attracted considerable attention because it enables us to understand the adaptive and versatile locomotion of animals and facilitates the construction of truly intelligent artificial agents. Thus far, we have developed a snake-like robot (HAUBOT I) that is driven by a decentralized control scheme based on a discrepancy function, which incorporates phasic control. In this paper, we investigate a decentralized control scheme in which phasic and tonic control are well coordinated, as an extension of our previous study. To verify the validity of the proposed control scheme, we apply it to a snake-like robot (HAUBOT II) that can adjust both the phase relationship between its body segments and the stiffness at each joint. The results indicate that the proposed control scheme enables the robot to exhibit remarkable real-time adaptability over various frictional and inclined terrains. These findings can potentially enable us to gain a deeper insight into the autonomous decentralized control mechanism underlying the adaptive and resilient locomotion of animals.