Design of a control framework for lower limb exoskeleton rehabilitation robot based on predictive assessment

https://doi.org/10.1016/j.clinbiomech.2022.105660Get rights and content

Highlights

  • Biceps femoris short head muscle has a great impact on the knee swing phase in gait cycle when walking.

  • Pathological gait was simulated by applying forward dynamics.

  • Established an impedance control model for lower limb exoskeleton rehabilitation robot.

  • Achieve customized adjustment of the exoskeleton rehabilitation robot motion trajectory.

Abstract

Background

Patients suffering from lower limb dyskinesia, especially in early stages of rehabilitation, have weak residual muscle strength in affected limb and require passive training by the lower limb rehabilitation robot. Anatomy indicates that the biceps femoris short head muscle has a strong influence on knee motion at the swing phase of walking. We sought to explore how it would influence on gait cycle in optimization framework. However, the training trajectory of conventional rehabilitation robots performing passive training usually follows gait planning based on general human gait data, which cannot simultaneously ensure both effective rehabilitation of affected limbs with varying severity pathological gait and comfort of the wearer within a safe motion trajectory.

Methods

To elucidate the effects of weakness and contracture, we systematically introduced isolated defects into the musculoskeletal model and generated walking simulations to predict the gait adaptation due to these defects. An impedance control model of the rehabilitation robot is developed. Knee joint parameters optimized by predictive forward dynamics simulation are adopted as the expected values for the robot controller to achieve customized adjustment of the robot motion trajectory.

Findings

Severe muscle contracture leads to severe knee flexion; severe muscle weakness induces a significant posterior tilt of the upper trunk, which hinders walking speed.

Interpretation

Our simulation results attempt to reveal pathological gait features, which may help to reproduce the simulation of pathological gait. Furthermore, the robot simulation results show that the robot system achieves a speedy tracking by setting a larger stiffness value. The model also allows the implementation of different levels of damping or elasticity effects.

Trial registration: The method proposed in this paper is an initial basic study that did not reach clinical trials and therefore retains retrospectively registered.

Introduction

At present, with a rapidly aging population and the increase of physical movement disorder patients caused by various diseases, rehabilitation robot technology has become a new research field which combines the medicine and engineering (Wang and Liu, 2020). Paralyzed patients with movement disorders require extensive rehabilitation programs to restore ambulatory function. Traditional manual therapy methods are labor-intensive and have limited therapeutic effects, thus promoting the development of robot-assisted rehabilitation therapy (Wu et al., 2016; Wu et al., 2018). Jezernik et al. (Jezernik et al., 2003) invented the robotic orthotic Lokomat, a rehabilitation robot developed to train spinal cord injury and stroke patients on a treadmill as a motor automation device. The Ekso GT exoskeleton (Pransky, 2014) developed by Ekso Bionics is a wearable exoskeleton suit designed for the assistance and rehabilitation of patients with various levels of lower extremity weakness. It is suitable for rehabilitation training for patients with lower limb mobility.

Patients with lower extremity movement disorders, especially in the early stages of rehabilitation, have weak residual muscle strength in the affected limb and require passive training of the affected limb by a lower extremity rehabilitation robot (Gilbert et al., 2016; Shi et al., 2018). Lei et al. (Lei et al., 2016), in a study of sensorimotor adaptation, found that subjects achieved better rehabilitation results when combining movement observation induction with passive training induction. H. J. Appell (Appell, 1997) found that damaged skeletal muscles entering the rehabilitation process in a weakened state should avoid high-mechanical loads and that electrical stimulation can be applied to promote recovery, which otherwise leads to impaired muscle function and often exhibits disturbances in innervation patterns. Another study by H. J. Appell (Appell, 1990) found that the regenerating fibers of the affected limb cannot mechanically withstand the generated intramuscular forces, especially during isometric movements, and therefore may suffer secondary damage. Therefore, it is recommended to start retraining with very moderate exercises and, when using isokinetic devices, to avoid maximum efforts, especially at slow angular velocities. In the passive training stage, the rehabilitation robot mainly drives the affected limb to perform rehabilitation training according to the preset gait trajectory. However, the motion trajectory of conventional rehabilitation robots performing passive training follows a gait plan established based on normal human gait data (Chen et al., 2016). Serrancolí et al. (Serrancoli et al., 2019) proposed that the prediction of human-exoskeleton forces is crucial for assessing user comfort and the effectiveness of interaction. This singularity in the motion trajectory of the rehabilitation robot during the passive training stage may not simultaneously ensure both effective rehabilitation and safe motion trajectories for various severity pathological gait and may cause patient discomfort during this rehabilitation training. The mentioned simulations have been successfully applied to estimate human motions which are not directly observable, such as muscle forces or joint moments (Delp et al., 2007). Although these dynamic simulations provide useful insights into human motion, they rely on existing data and cannot predict new behavior (Seth et al., 2011). Moreover, motion capture requires expensive hardware equipment, which makes it difficult to apply to the rehabilitation treatment stage.

Biceps femoris short head (BFSH) weakness or contracture, usually occurs in conditions such as cerebral palsy, stroke and secondary dysfunction after knee osteoarthritis surgery. Although these deficits are the pathologies cause of the observed knee flexion contractures, this phenomenon is usually accompanied by neurological deficits and biomechanical factors, making it difficult to confirm the correspondence. Carmichael F. Ong et al. (Ong et al., 2019a) applied the predictive dynamics simulation method to investigate how SOL and GAS muscles affect the plantarflexor of the ankle joint and generalized the relationship between them. Kirsten Veerkamp et al. (Veerkamp et al., n.d.) evaluated how gastrocnemius hyperreflexia affects gait kinematics by using predictive simulations. The isometric length and force of the muscles and tendons have a large effect on the movement of the knee joint (Campbell and Trudel, 2020). Knee flexion contracture and semimembrane tendon and biceps femoris tendon affect each other (Nakagawa et al., 2020). Anatomy indicates that the BFSH muscle has a strong impact on knee joint motion during the swing phase of walking[16,17].

In this paper, the pathological gait is simulated by the predictive forward dynamics with energy efficiency as the high-level goal. The gait controller uses a combination of state machines and low-level control laws to determine the excitation, aiming at calculating the optimal motion trajectory to accomplish the given task and ultimately seeking the relationship between the impact of the tendon system on the pathological gait. The paper systematically introduces the isolated BFSH muscle into the musculoskeletal model as an example of knee dyskinesia in order to predict gait adaptation due to the lack of this muscle group. We apply mild, moderate and severe muscle weakness or contracture to the BFSH and retrained the model to walk at a self-selected speed to recreate pathological gait without experimental data and obtain more realistic changes in knee parameters. The framework attempts to obtain a pathological gait model by studying the effect of muscle weakness or contraction for injured muscle on walking. The optimized parameters (angle, angular velocity, torque, etc.) are then fed back into the robot controller. This paper proposed a new approach to clinical gait analysis with the aim of both improving the effectiveness of rehab training and increasing the comfort of the wearer and reducing patient resistance.

Section snippets

Methods

Our control framework consists of a predictive dynamic simulation and a lower limb exoskeleton rehabilitation robot. Comparison of predictive forward dynamic simulation and lower limb exoskeleton rehabilitation robot are provided in the supplemental material. Predictive dynamics uses shooting methods to solve the dynamic optimization problem for generating gait simulations. We implement the musculoskeletal model in OpenSim 3.3 (NCSRR, USA), which is actuated by 18 Hill-type musculoskeletal

Simulating walk with BFSH contracture or weakness

According to the forward predictive simulation framework (Fig. 2), we proceed to analyze the BFSH muscle of the lower extremities in contracture and muscle weakness, respectively. Comparison of simulation results with experimental data collected in the literature (Moissenet and Armand, 2015; Wang et al., 2012), knee angle for the normal model in the late swing was lag behind experimental data, knee moments were larger than the experimental data during swing (Fig. 4). The reason was probably

Discussion

We have shown that the musculoskeletal assessment framework for the lower extremities can successfully predict the simulation results of gait adaptation for the BFSH muscle without tracking experimental data. Predicted gait adaptation was verified to be basically accurate using the EMG sensors. Our work aimed to obtain simulation results of pathological gait adaptation due to different degrees of deficiencies in the BFSH muscle through a predictive assessment framework. In addition, the

Conclusion and future works

We present a control framework for a lower limb exoskeleton rehabilitation robot based on predictive simulation assessment. Firstly, the musculoskeletal model of the affected limb is modified using Opensim based on medical advice. The modified model is simulated by SCONE for predictive optimization, aiming to reappear the pathological gait. In this paper, taking the BFSH muscle as an example, the effects of BFSH muscle contracture and weakness on the knee joint motion are studied separately for

Ethics approval and consent to participate

The surface EMG signal acquisition experiments involved in this paper were performed on the author's leg, therefore, this article is not applicable.

Consent for publication

Not applicable.

Availability of data and materials

The data presented in this study are available on request from the corresponding author.

Declaration of Competing Interest

The authors declare no conflict of interest.

Acknowledgments

This work was supported in part by 2021 Guangxi University's Young and Middle-aged Teachers' Basic Research Ability Improvement Project (Grant No.2021KY0442), Guangxi Ship Digital Design and Advanced Manufacturing Engineering Technology Research Center of Beibu Gulf University, Open Fund Project of State Key Laboratory of Hydraulic Engineering Simulation and Safety (No. HESS-1801).

References (32)

  • Q. Wu et al.

    Development of an RBFN-based neural-fuzzy adaptive control strategy for an upper limb rehabilitation exoskeleton

    Mechatronics

    (2018)
  • H.-J. Appell

    The muscle in rehabilitation

    Orthopade

    (1997)
  • Hans-Joachim Appell

    Muscular Atrophy Following Immobilisation: A Review

    Sports Med.

    (1990)
  • S. Armand et al.

    Gait analysis in children with cerebral palsy

    EFORT Open Rev.

    (2016)
  • S.L. Delp

    OpenSim: open-source software to create and analyze dynamic simulations of movement

    IEEE Trans. Biomed. Eng.

    (2007)
  • B.J. Fregly

    Grand challenge competition to predict in vivo knee loads

    J. Orthop. Res.

    (2012)
  • Cited by (5)

    All authors were fully involved in the study and preparation of the manuscript and the material within has not been and will not be submitted for publication elsewhere.

    View full text