Soft Robotic System Offers New Hope for Spinal Injury Patients
In a significant advancement for rehabilitation technology, researchers at the University of Shanghai for Science and Technology have developed a novel lower limb rehabilitation robot driven by parallel rope mechanisms. The system, designed to address the limitations of traditional rigid exoskeletons, leverages soft actuation principles to deliver safer, more adaptable gait training for patients recovering from spinal cord injuries (SCI). Published in the Journal of Biomedical Engineering Research, the study presents a comprehensive design, kinematic modeling, and simulation-based validation of the device, demonstrating its potential to improve therapeutic outcomes through enhanced human-machine interaction.
Spinal cord injuries remain a critical global health challenge, affecting millions worldwide and often resulting in long-term mobility impairment. According to epidemiological data cited in the study, the incidence of SCI in North America alone reaches 39 per million, with rising trends observed across developing nations. These injuries frequently lead to paraplegia or complete paralysis, severely diminishing patients’ quality of life and independence. While neurorehabilitation strategies have evolved significantly, the restoration of walking function continues to be a primary therapeutic goal. Clinical evidence supports the efficacy of repetitive, task-specific gait training in promoting neuroplasticity and functional recovery. However, conventional robotic systems used in such training—often based on rigid linkages—face inherent challenges in flexibility, safety, and patient comfort.
Existing commercial solutions such as the HAL exoskeleton developed at the University of Tsukuba in Japan and the LOKOMAT system by Swiss medical technology firms have demonstrated clinical utility. The HAL suit, for instance, uses bioelectric signals to initiate movement, enabling patients to engage in assisted walking. Despite its technological sophistication, the system’s high structural stiffness limits its operational workspace and may restrict natural movement patterns. Similarly, the LOKOMAT provides dynamic body-weight support and guided gait training but suffers from limited compliance during active participation. When patients exert effort during training, the applied torque can alter the intended trajectory, reducing the system’s adaptability and potentially compromising the therapeutic effect.
To overcome these limitations, the Shanghai-based team led by Zou Renling has turned to soft robotics, specifically rope-driven parallel mechanisms. Unlike rigid actuators, flexible cables offer inherent compliance, allowing for smoother force transmission and greater tolerance to misalignment between the human limb and the machine. This compliance is particularly advantageous in rehabilitation settings, where patient movement variability, muscle spasticity, and postural instability are common. By replacing metal joints and linkages with tensioned ropes, the new system reduces the risk of injury during interaction and enhances overall user comfort—key factors in promoting patient adherence to long-term therapy regimens.
The newly designed robot integrates two primary subsystems: a dynamic weight-support mechanism and a lower-limb motion unit. The weight-support component utilizes an electric cylinder coupled with a pulley system to provide adjustable vertical suspension, effectively reducing the gravitational load on the patient’s legs. This feature is crucial for individuals with partial or complete lower-limb paralysis, as it enables them to practice standing and stepping motions without full weight-bearing. The motion unit, responsible for driving hip and knee flexion and extension, employs a screw-driven linear module to actuate the ropes. These ropes are anchored at the center of mass of the thigh and shank, ensuring balanced force application and minimizing shear or rotational disturbances during movement.
One of the core innovations lies in the mechanical design of the cable routing. To maintain consistent tension and prevent slippage—a common issue in long-distance rope transmissions—the team implemented an S-shaped pulley configuration. This serpentine path increases the contact surface between the rope and the pulleys, enhancing friction and reducing the likelihood of sudden force fluctuations. Additionally, the frame is constructed from hollow stainless steel square tubing, selected for its high strength-to-weight ratio and excellent fatigue resistance. Reinforced with structural ribs at the base, the overall assembly ensures mechanical stability while remaining lightweight and durable.
From a biomechanical standpoint, the robot focuses on sagittal plane movements—the primary direction of motion during walking. The target range of motion includes hip flexion/extension between -20° and 20° and knee flexion from 0° to 70°, aligning with standard gait parameters observed in healthy individuals. Mechanical limit switches are incorporated at the extremities of these ranges to prevent overextension and ensure patient safety during automated training sessions. This constrained yet functional workspace is designed to replicate the essential kinematics of walking while minimizing the risk of joint strain or dislocation.
To model and control the robot’s motion, the research team established a detailed kinematic framework using the Denavit-Hartenberg (D-H) convention, a standard method in robotics for describing the spatial relationships between joints. By treating the human leg as a two-link kinematic chain—comprising the thigh and shank—they derived both forward and inverse kinematic solutions for the rope-driven system. The inverse kinematics problem, which determines the required rope lengths given a desired joint angle, was solved analytically using vector geometry. This allowed the researchers to compute the precise extension and retraction profiles needed for each actuator to achieve a target posture.
The forward kinematics, conversely, involves determining the joint angles based on measured rope lengths—a more complex, nonlinear problem due to the under-constrained nature of the system. To solve this, the team employed the Newton-Raphson iterative method, a numerical technique well-suited for converging on solutions within a specified tolerance. By initializing the algorithm with estimated joint angles and iteratively refining them based on the difference between actual and predicted rope lengths, the system could accurately reconstruct limb posture in real time. Validation tests in MATLAB showed angular estimation errors consistently below 0.001 radians, confirming the robustness and precision of the computational model.
A critical aspect of any robotic system is its operational workspace—the set of all reachable positions and orientations of the end effector, in this case, the ankle or foot. For rope-driven robots, workspace analysis is particularly challenging due to the unidirectional nature of cables (which can only pull, not push) and the need to maintain positive tension across all ropes. The researchers addressed this by applying Farkas’ Lemma, a mathematical principle used in convex analysis to determine the feasibility of force closure in under-constrained systems. Using a numerical approach, they sampled joint angles across the permissible range and computed the corresponding endpoint positions, generating a point cloud that mapped the achievable workspace.
The resulting distribution revealed a dense clustering of points around the central gait position, with gradual dispersion toward the extremes of motion—consistent with natural walking patterns. This non-uniform density reflects the biomechanical reality that most of the gait cycle occurs within a mid-range of joint angles, with rapid transitions at heel strike and toe-off. The coherence between the simulated workspace and physiological expectations further validates the mechanical and control design of the robot.
To evaluate the system’s performance under dynamic conditions, the researchers conducted multi-body simulations using Adams, a high-fidelity mechanical modeling software. A simplified human model was created with segment lengths set at 500 mm for the thigh and 400 mm for the shank, approximating average adult proportions. Revolute joints were assigned at the hip and knee, and the ropes were modeled as tension-only elements connected via fixed constraints. Motion profiles were driven by spline-based functions that replicated the temporal evolution of joint angles during a standard gait cycle.
The simulation outputs were then compared against the CGA (Clinical Gait Analysis) standard database, a widely accepted benchmark for normal walking kinematics. The results showed a close alignment between the robot-generated trajectories and reference data. The hip joint angle varied smoothly between -20° and 20°, while the knee flexed up to 70° during swing phase—both within clinically acceptable ranges. Although minor deviations were observed, particularly during transition phases, the overall root-mean-square error remained within permissible limits. These discrepancies were attributed to inherent differences between idealized models and real human movement, including soft tissue deformation, joint laxity, and neuromuscular noise.
Importantly, the high degree of trajectory fidelity suggests that the robot can deliver consistent, repeatable motion patterns—essential for effective motor relearning. In neurorehabilitation, repetition is not merely about quantity but also about quality. Erratic or inconsistent guidance can reinforce abnormal movement patterns, potentially hindering recovery. By minimizing trajectory error, the rope-driven system ensures that patients receive accurate sensory feedback, which is crucial for cortical reorganization and the rebuilding of neural pathways.
Another advantage of the design is its modular architecture. The separation between the weight-support system and the motion actuation unit allows for independent tuning of each subsystem. Clinicians can adjust the level of body-weight support based on the patient’s strength and endurance, while simultaneously modifying the speed, amplitude, or resistance of the leg movements. This adaptability makes the robot suitable for patients across the recovery spectrum—from those with minimal voluntary movement to those regaining independent ambulation.
Moreover, the use of ropes instead of rigid links reduces the overall inertia of the moving parts, leading to faster response times and lower energy consumption. This efficiency translates into quieter operation and less mechanical stress on both the device and the user. The reduced mass also simplifies transportation and installation, making the system more accessible for clinical deployment in hospitals, rehabilitation centers, and even home environments.
The research also highlights the importance of safety in human-robot interaction. Unlike rigid exoskeletons that can exert high contact pressures or cause pinching injuries, the compliant nature of rope actuation inherently limits peak forces. Even in the event of a control malfunction or sudden stop, the elasticity of the cables acts as a passive damper, absorbing kinetic energy and reducing the risk of trauma. This passive safety feature is particularly valuable when working with vulnerable populations, including elderly patients or those with spasticity.
While the current study is based on simulation and theoretical modeling, the team has laid the groundwork for future clinical testing. The successful validation of kinematic accuracy and workspace feasibility suggests that the prototype is ready for integration with real-time control systems and human subject trials. Future work may include the incorporation of electromyography (EMG) sensors to enable patient-initiated movement, or force sensors to implement impedance control—allowing the robot to adapt its stiffness based on user intent.
Additionally, the platform could be extended to bilateral training, enabling coordinated movement of both legs. Asymmetrical gait patterns are common in stroke and SCI patients, and a dual-sided system could provide targeted correction through differential assistance. Integration with virtual reality (VR) environments could further enhance engagement, turning repetitive exercises into immersive therapeutic games that improve motivation and cognitive involvement.
The implications of this research extend beyond spinal cord injury rehabilitation. The same rope-driven architecture could be adapted for stroke recovery, cerebral palsy, or post-surgical orthopedic rehabilitation. Its scalability and modularity make it a promising candidate for personalized medicine, where treatment protocols are tailored to individual patient needs. As healthcare systems increasingly emphasize cost-effective, home-based therapies, soft robotic devices like this one could play a transformative role in expanding access to high-quality rehabilitation.
In conclusion, the rope-driven lower limb rehabilitation robot developed by Wang Yiming, Wei Yishan, Hu Xiufang, Zou Renling, Xu Xiulin, Lin Haibo, and Bin Kehua at the University of Shanghai for Science and Technology represents a meaningful step forward in assistive technology. By combining mechanical innovation with rigorous computational analysis, the team has created a system that is not only technically sound but also clinically relevant. Its emphasis on safety, comfort, and biomimetic motion aligns with the evolving demands of modern rehabilitation. As the field continues to embrace soft robotics, this work stands as a compelling example of how engineering ingenuity can directly improve human health and restore lost function.
Soft Robotic System Offers New Hope for Spinal Injury Patients
Wang Yiming, Wei Yishan, Hu Xiufang, Zou Renling, Xu Xiulin, Lin Haibo, Bin Kehua, University of Shanghai for Science and Technology, Journal of Biomedical Engineering Research, DOI 10.19529/j.cnki.1672-6278.2021.04.10