While rehabilitation of upper-limb motor function with human-interactive robots has been met with success, robot-aided locomotor rehabilitation has proven challenging. Conventional therapeutic robots for walking impose pre-planned repetitive kinematic trajectories on patients using high-gain trajectory tracking controllers. This approach may have discouraged active engagement of patients in making movements and interfered with the natural oscillatory dynamics of walking. To inform more effective approaches to robotic gait therapy, it is important to understand neuro-mechanical dynamics and control of unimpaired locomotion. Our previous studies reported that human gait entrained to periodic mechanical perturbations at the ankle when the perturbation period was close to preferred walking cadence. Moreover, entrainment was accompanied by synchronizing the perturbations to a constant gait phase, the same for all subjects, where they provided mechanical assistance. To test the generality of entrainment-based assistance, we are studying the behavior of unimpaired subjects walking overground while wearing a hip exoskeleton robot, the Samsung Gait Enhancing and Motivating Systems for hip (GEMS-H). The study would quantify important aspects of the nonlinear neuro-mechanical dynamics underlying the control of walking, which will inform the development of effective approaches to robotic walking therapy.