Today, we want to bring back an interesting paper from the recent past. The article “Design and Validation of an Ambulatory User Support Gait Rehabilitation Robot: NIMBLE”, developed by the BeST group at Universidad Rey Juan Carlos, addresses a long-standing question in rehabilitation robotics. How can we combine body weight support, overground gait training and active assistance in a single coherent system? The NIMBLE gait rehabilitation robot integrates a mobile robotic frame, a partial body weight support (PBWS) system, a lower-limb Exo-H3 exoskeleton and a cable-driven pelvis actuator. The goal is to train gait in realistic scenarios, while preserving safety and therapy personalization.
System architecture and role of the Exo-H3 in the NIMBLE gait rehabilitation exoskeleton
From an engineering point of view, the NIMBLE gait rehabilitation exoskeleton is conceived as a three-layer modular platform. The mobile frame, the exoskeleton and the pelvis actuator are coordinated through a ROS 2 based architecture and a hierarchical control scheme. The Exo-H3 (Technaid) provides active mobilization of hip, knee and ankle in the sagittal plane, with six actuated degrees of freedom and integrated sensing for joint position, interaction forces and plantar pressure. This enables the implementation of gait patterns and stability strategies already tested in other rehabilitation contexts. The PBWS frame guarantees partial unloading of body weight and trajectory tracking, while the cable-driven pelvis unit assists the centre of mass to train lateral balance and weight transfer with lower mechanical stiffness than rigid robotic solutions.
Experimental validation: kinematics and frame control
The study focuses on two main experimental questions. First, how much does the exoskeleton–frame coupling alter gait kinematics? Second, how effective is the frame tracking strategy under realistic disturbances? To isolate the pure mechanical effect of the coupling, the authors analyse treadmill walking in five healthy subjects under three conditions. These are free walking, walking with the Exo-H3 corset alone and walking with the corset attached to the frame, without the leg bars of the exoskeleton. Two-dimensional kinematics are extracted with Kinovea and processed in Matlab to obtain joint ranges of motion, centre of mass displacement and step width, together with specific tests of the frame controller response to sudden misalignments.
Results show that the impact of the frame on gait kinematics is quantitatively small. In the sagittal plane, the most visible change appears in hip extension, but the additional effect of the frame is around two degrees when compared to the corset condition alone. The total deviation with respect to free walking remains within clinically acceptable limits. In the frontal plane, variations in range of motion and centre of mass displacement are on the order of two degrees and 0.6 cm. Step width does not show significant changes between conditions. At the same time, the frame velocity controller corrects maximum desynchronisation in less than 0.5 seconds and maintains a root mean square error around 4.8 cm at low walking speeds, which supports the feasibility of the approach for overground therapy.
Context within rehabilitation robotics and contribution beyond the state of the art
This work must be read in the context of a large body of literature on body weight support systems and robot-assisted gait training. Many devices have shown promising clinical results, but with heterogeneous outcomes when compared with conventional physiotherapy. Previous reviews on exoskeletons and PBWS systems have emphasised the need for more flexible solutions that allow overground gait, dynamic balance and weight shift training, instead of only predefined treadmill patterns. In comparison with platforms such as Andago, HYBRID, MLLRE or Walktrainer, NIMBLE stands out by combining weight support, centre of mass training and active exoskeleton assistance in a single system, while keeping mechanical and control complexity at a reasonable level thanks to the cable-driven pelvis design.
For researchers in robotics and rehabilitation, the added value of this study lies in the quantitative validation of the trade-off between movement freedom and system complexity. The kinematic data show that it is possible to add meaningful functionality, such as balance and weight shift training, without severely degrading the user’s gait pattern or introducing excessive mechanical constraints. At the same time, the ROS 2 architecture and modular design invite the integration of new control layers. Examples include adaptive control, human-in-the-loop optimisation or the use of physiological biomarkers based on EMG or additional sensing. In this sense, the NIMBLE gait rehabilitation robot emerges as a suitable experimental platform for future clinical trials and advanced control studies, with the Exo-H3 exoskeleton as the core lower-limb assistance device.
Full publication: https://www.technaid.com/2024-09-design-and-validation-of-an-ambulatory-user-support-gait-rehabilitation-robot-nimble/