Stable Walking Research Articles

Reference Motion Quality and Design Choices for Bipedal Walking with PPO

This paper investigates the influence of reference motion quality and other design choices on the performance of deep reinforcement learning for bipedal walking with Proximate Policy Optimization (PPO). We use parametrized Cartesian quintic splines to generate reference actions for an omnidirectional walk policy. By using parameter sets with different qualities, we show that the performance of the trained policy correlates to the quality of the reference motion. We also show that a policy in Cartesian space performs superior to a joint-space-based one if an advantageous representation of orientation is chosen. Additionally, we show that using an initial bias for the policy speeds up the training and leads to higher performances for policies using position control. We also show that we can achieve a stable omnidirectional walk on a wide variety of simulated humanoid robots.

Analysis and Experiment of a Bioinspired Multimode Octopod Robot

Legged robots use isolated footholds to support, which have the merit of good terrain trafficability but lack speed ability. In contrast, wheeled robots have the advantages of high speed and efficiency but only run on flat roads. To improve the moving speed and terrain adaptability of the legged robot, this paper proposes a bioinspired multimode octopod robot with rolling, walking, and obstacle-surmounting modes. First, inspired by the multimode locomotion of the Cebrennus rechenbergi spider, the high-speed mobility of the legged robot is realized in involute kick-rolling mode through the extendable appendages. Then, the foot and appendage trajectories are analyzed by kinematic method and optimized for walking stability. Based on the static and the kinematic analyses, the terrain adaptability is improved by adhesive obstacle-surmounting mode with the assistance of the appendages affiliated to the main feet. The deformable trunk with one DoF is designed to switch between three modes. Finally, a series of dynamic simulations and experiments are carried out to verify the theoretical analyses of the adhesive obstacle-surmounting mode and the mobility of the involute kick-rolling mode. It is shown that the multimode octopod robot can integrate the advantages of high speed and good terrain trafficability from different types of robots and is suitable for performing tasks in unstructured terrains.

Sharp changes in fractal basin of attraction in passive dynamic walking

A passive dynamic walker is a mechanical system that walks down a slope without any control, and gives useful insights into the dynamic mechanism of stable walking. This system shows specific attractor characteristics depending on the slope angle due to nonlinear dynamics, such as period-doubling to chaos and its disappearance by a boundary crisis. However, it remains unclear what happens to the basin of attraction. In our previous studies, we showed that a fractal basin of attraction is generated using a simple model over a critical slope angle by iteratively applying the inverse image of the Poincaré map, which has stretching and bending effects. In the present study, we show that the size and fractality of the basin of attraction sharply change many times by changing the slope angle. Furthermore, we improved our previous analysis to clarify the mechanisms for these changes and the disappearance of the basin of attraction based on the stretching and bending deformation in the basin formation process. These findings will improve our understanding of the governing dynamics to generate the basin of attraction in walking.

An Energy-Based Framework for Robust Dynamic Bipedal Walking Over Compliant Terrain

Abstract Bipedal locomotion over compliant terrain is an important and largely underexplored problem in the robotics community. Although robot walking has been achieved on some non-rigid surfaces with existing control methodologies, there is a need for a systematic framework applicable to different bipeds that enables stable locomotion over various compliant terrains. In this work, a novel energy-based framework is proposed that allows the dynamic locomotion of bipeds across a wide range of compliant surfaces. The proposed framework utilizes an extended version of the 3D Dual Spring-Loaded Inverted Pendulum (Dual-SLIP) model that supports compliant terrains, while a bioinspired controller is employed to regulate expected perturbations of extremely low ground-stiffness levels. An energy-based methodology is introduced for tuning the bioinspired controller to enable dynamic walking with robustness to a wide range of low ground-stiffness one-step perturbations. The proposed system and controller are shown to mimic the vertical Ground Reaction Force (GRF) responses observed in human walking over compliant terrains. Moreover, they succeed in handling repeated unilateral stiffness perturbations under specific conditions. This work can advance the field of biped locomotion by providing a biomimetic method for generating stable human-like walking trajectories for bipedal robots over various compliant surfaces. Furthermore, the concepts of the proposed framework could be incorporated into the design of controllers for lower-limb prostheses with adjustable stiffness to improve their robustness over compliant surfaces.

The effect of prolonged walking on leg muscle activity patterns and vulnerability to perturbations

Understanding the consequences and ecological relevance of muscle fatigue is important to guide the development of strategies to preserve independence. However, few studies have examined walking-related fatigue and the effects on walking instability. Our purpose was to investigate the effects of prolonged walking on leg muscle activity and vulnerability to balance perturbations. Eighteen healthy young adults completed a 30-min walking trial at their preferred walking speed while leg muscle activities were recorded. Before and after the 30-min walk, participants responded to five 5% body weight lateral force perturbations. Time-frequency analysis with wavelet transformation and principal component analyses assessed neuromuscular adaptations of muscles to prolonged walking. Following prolonged walking, we observed a time-dependent increase in EMG intensities at slower frequencies for the soleus and tibialis anterior and a decrease in mean amplitudes for the soleus, lateral gastrocnemius, and semitendinosus. Mean mediolateral CoM displacement following perturbations averaged 21% larger after the 30-min walk. Our results suggest that walking for 30 min at a comfortable speed elicits complex neuromuscular adaptations indicative of local muscle fatigue and an increased vulnerability to walking balance perturbations. These findings could inform fatigue monitoring systems or walking assistive devices aimed at reducing walking-related fatigue and maintaining independent mobility.

Mediolateral foot placement control can be trained: Older adults learn to walk more stable, when ankle moments are constrained.

Falls are a problem, especially for older adults. Placing our feet accurately relative to the center-of-mass helps us to prevent falling during gait. The degree of foot placement control with respect to the center-of mass kinematic state is decreased in older as compared to young adults. Here, we attempted to train mediolateral foot placement control in healthy older adults. Ten older adults trained by walking on shoes with a narrow ridge underneath (LesSchuh), restricting mediolateral center-of-pressure shifts. As a training effect, we expected improved foot placement control during normal walking. A training session consisted of a normal walking condition, followed by a training condition on LesSchuh and finally an after-effect condition. Participants performed six of such training sessions, spread across three weeks. As a control, before the first training session, we included two similar sessions, but on normal shoes only. We evaluated whether a training effect was observed across sessions and weeks in a repeated-measures design. Whilst walking with LesSchuh, the magnitude of foot placement error reduced half-a-millimeter between sessions within a week (cohen's d = 0.394). As a training effect in normal walking, the magnitude of foot placement errors was significantly lower compared to the control week, by one millimeter in weeks 2 (cohen's d = 0.686) and 3 (cohen's d = 0.780) and by two millimeters in week 4 (cohen's d = 0.875). Local dynamic stability of normal walking also improved significantly. More precise foot placement may thus have led to improved stability. It remains to be determined whether the training effects were the result of walking on LesSchuh or from repeated treadmill walking itself. Moreover, enhancement of mechanisms beyond the scope of our outcome measures may have improved stability. At the retention test, gait stability returned to similar levels as in the control week. Yet, a reduction in foot placement error persisted.

Effects of protective step training on proactive and reactive motor adaptations in Parkinson's disease patients.

The aim of this study was to investigate to what extent PD affects the ability to walk, respond to balance perturbations in a single training session, and produce acute short-term effects to improve compensatory reactions and control of unperturbed walking stability. Understanding the mechanism of compensation and neuroplasticity to unexpected step perturbation training during walking and static stance can inform treatment of PD by helping to design effective training regimens that remediate fall risk. Current rehabilitation therapies are inadequate at reducing falls in people with Parkinson's disease (PD). While pharmacologic and surgical treatments have proved largely ineffective in treating postural instability and gait dysfunction in people with PD, studies have demonstrated that therapy specifically focusing on posture, gait, and balance may significantly improve these factors and reduce falls. The primary goal of this study was to assess the effectiveness of a novel and promising intervention therapy (protective step training - i.e., PST) to improve balance and reduce falls in people with PD. A secondary goal was to understand the effects of PST on proactive and reactive feedback responses during stance and gait tasks. Multiple-baseline, repeated measures analyses were performed on the multitude of proactive and reactive performance measures to assess the effects of PST on gait and postural stability parameters. In general, the results indicate that participants with PD were able to use experiences with perturbation training to integrate and adapt feedforward and feedback behaviors to reduce falls. The ability of the participants with PD to adapt to changes in task demands suggests that individuals with PD could benefit from the protective step training to facilitate balance control during rehabilitation.

Altered functional activity of the precuneus and superior temporal gyrus in patients with residual dizziness caused by benign paroxysmal positional vertigo.

Benign paroxysmal positional vertigo (BPPV) is a common clinical vertigo disease, and the most effective treatment for this disease is canal repositioning procedures (CRP). Most patients return to normal after a single treatment. However, some patients still experience residual dizziness (RD) after treatment, and this disease's pathogenesis is currently unclear. The purpose of this study is to explore whether there are abnormal brain functional activities in patients with RD by using resting-state functional magnetic resonance imaging (rs-fMRI) and to provide imaging evidence for the study of the pathogenesis of RD. The BPPV patients in the Second Affiliated Hospital of Xuzhou Medical University had been included from December 2021 to November 2022. All patients had been received the collection of demographic and clinical characteristics (age, gender, involved semicircular canal, affected side, CRP times, BPPV course, duration of RD symptoms, and whether they had hypertension, diabetes, coronary heart disease.), scale assessment, including Dizziness Handicap Inventory (DHI), Hamilton Anxiety Inventory (HAMA), Hamilton Depression Inventory (HAMD), rs-fMRI data collection, CRP treatment, and then a one-month follow-up. According to the follow-up results, 18 patients with RD were included. At the same time, we selected 19 healthy individuals from our hospital's physical examination center who matched their age, gender as health controls (HC). First, the amplitude of low-frequency fluctuations (ALFF) analysis method was used to compare the local functional activities of the two groups of subjects. Then, the brain regions with different ALFF results were extracted as seed points. Functional connectivity (FC) analysis method based on seed points was used to explore the whole brain FC of patients with RD. Finally, a correlation analysis between clinical features and rs-fMRI data was performed. Compared to the HC, patients with RD showed lower ALFF value in the right precuneus and higher ALFF value in the right superior temporal gyrus (STG). When using the right STG as a seed point, it was found that the FC between the right STG, the right supramarginal gyrus (SMG), and the left precuneus was decreased in RD patients. However, no significant abnormalities in the FC were observed when using the right precuneus as a seed point. In patients with RD, the local functional activity of the right precuneus is weakened, and the local functional activity of the right STG is enhanced. Furthermore, the FC between the right STG, the right SMG, and the left precuneus is weakened. These changes may explain the symptoms of dizziness, floating sensation, walking instability, neck tightness, and other symptoms in patients with RD to a certain extent.

Simulating human walking: a model-based reinforcement learning approach with musculoskeletal modeling.

Recent advancements in reinforcement learning algorithms have accelerated the development of control models with high-dimensional inputs and outputs that can reproduce human movement. However, the produced motion tends to be less human-like if algorithms do not involve a biomechanical human model that accounts for skeletal and muscle-tendon properties and geometry. In this study, we have integrated a reinforcement learning algorithm and a musculoskeletal model including trunk, pelvis, and leg segments to develop control modes that drive the model to walk. We simulated human walking first without imposing target walking speed, in which the model was allowed to settle on a stable walking speed itself, which was 1.45 m/s. A range of other speeds were imposed for the simulation based on the previous self-developed walking speed. All simulations were generated by solving the Markov decision process problem with covariance matrix adaptation evolution strategy, without any reference motion data. Simulated hip and knee kinematics agreed well with those in experimental observations, but ankle kinematics were less well-predicted. We finally demonstrated that our reinforcement learning framework also has the potential to model and predict pathological gait that can result from muscle weakness.

The Effects of Structural Characteristics of the Rollator on the Elderly’s Gait Strategies in Various Walking Environments

A rollator, one of the most widely used among walking assistance devices, can assist the elderly with stable walking in their daily lives. In this study, we investigated how the structural characteristics of two types of rollators affect the upper and lower extremity muscle activity and plantar pressure of the elderly in various walking environments. We quantified muscle activity (upper and lower limbs) and plantar pressure (mean force, peak pressure, and contact area) of 11 older adults walking in various environments (flat, obstacle, uneven, and sloped terrain) using two types of rollators. Upper extremity muscle activity was highest in the obstacle terrain and the uneven terrain, and a significant difference was found due to the structural differences of the rollator. Additionally, it was observed that lower extremity muscle activity and plantar pressure patterns appeared in accordance with the gait strategy to maintain stability in an unstable or inclined walking environment. In other words, it was confirmed that the weight of the rollator, the size of the wheel, grip type, and the auxiliary tools had a great effect on the upper and lower extremity muscle activity and plantar pressure of the elderly during walking. From the results of this study, it can be suggested that it is absolutely necessary to consider the biomechanical characteristics of the elderly and the structure of the rollator, which appear differently depending on the walking environment, in the development of walking aids. In the future, more clinical data will be collected, and based on this a rollator that can safely assist the elderly in various walking environments will be developed.

Green Function for an Asymptotically Stable Random Walk in a Half Space

We consider an asymptotically stable multidimensional random walk S(n)=(S1(n),…,Sd(n))\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$S(n)=(S_1(n),\\ldots , S_d(n) )$$\\end{document}. For every vector x=(x1…,xd)\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$x=(x_1\\ldots ,x_d)$$\\end{document} with x1≥0\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$x_1\\ge 0$$\\end{document}, let τx:=min{n>0:x1+S1(n)≤0}\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\ au _x:=\\min \\{n>0: x_{1}+S_1(n)\\le 0\\}$$\\end{document} be the first time the random walk x+S(n)\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$x+S(n)$$\\end{document} leaves the upper half space. We obtain the asymptotics of pn(x,y):=P(x+S(n)∈y+Δ,τx>n)\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$p_n(x,y):= {\ extbf{P}}(x+S(n) \\in y+\\Delta , \ au _x>n)$$\\end{document} as n tends to infinity, where Δ\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\Delta $$\\end{document} is a fixed cube. From that, we obtain the local asymptotics for the Green function G(x,y):=∑npn(x,y)\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$G(x,y):=\\sum _n p_n(x,y)$$\\end{document}, as |y|\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$|y |$$\\end{document} and/or |x|\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$|x |$$\\end{document} tend to infinity.

Effects of ankle-foot orthosis on gait pattern and spatiotemporal indices during treadmill walking in hemiparetic stroke.

Ankle-foot orthosis (AFO) is known to correct abnormal gait patterns and improve walking stability and speed in patients with hemiparesis. To quantify these benefits in post-stroke gait, a three-dimensional motion analysis of gait pattern was conducted. Forty patients with hemiparesis were enrolled. A three-dimensional motion analysis system was used to analyze patients' treadmill walking with/without an AFO. Outcome measures were 12 abnormal gait indices (forefoot contact, knee extensor thrust, retropulsion of the hip, flexed-knee gait, medial whip in the stance phase, circumduction gait, hip hiking, insufficient knee flexion during the swing phase, excessive lateral shifting of the trunk, contralateral vaulting, excessive hip external rotation, and posterior pelvic tilt), calculated using kinematic data and spatiotemporal indices, and the symmetry index of double-stance and single-stance time and step length. Forefoot contact (without AFO vs. with AFO: 71.0 vs. 65.8, P < 0.001), circumduction gait (65.0 vs. 57.9, P < 0.001), and contralateral vaulting (78.2 vs. 72.2, P = 0.003) were significantly reduced, whereas excessive hip external rotation (53.7 vs. 62.8, P = 0.003) significantly increased during walking with an AFO. Hip hiking (77.1 vs. 71.7) showed marginal reduction with the use of AFO ( P = 0.096). The absolute symmetry index of double-stance time (21.9 vs. 16.1, P = 0.014) significantly decreased during walking with an AFO. AFO effectively mitigates abnormal gait patterns typical of hemiparetic gait. A 3D motion analysis system with clinically oriented indices can help assess intervention efficacy for gait abnormalities.

Clinical profile and sensory pattern in lateral medullary syndrome

Background: Lateral Medullary syndrome is an acute ischemic stroke syndrome characterised by occlusion of the vertebral artery/posterior inferior cerebellar artery. The study aimed to describe the sensory pattern and clinical profile of patients admitted with lateral medullary syndrome. Methods: This cross-sectional descriptive study was done in the neurology department at Thanjavur Government Medical College, Tamil Nadu. Twenty patients with acute stroke clinically suspected of lateral medullary syndrome were taken into this study for one year. Magnetic Resonance Imaging (MRI) of the brain with an MR Angiogram was done to confirm clinical suspicion of Lateral medullary syndrome. They were examined in detail for clinical profile and sensory pattern, and relevant investigations were done. Results: Among the 20 LMS patients, 85% were males, and the mean age of occurrence was 53.2 years. All of them had acute infarcts in the lateral medulla. Hypertension, diabetes mellitus and dyslipidemia were the most common systemic risk factors. The most common sensory pattern observed was classical, with loss of sensation in the ipsilateral face and contralateral body in 45% of patients. The most common clinical presentation was walking instability, bulbar symptoms, and a longer hospital stay. The second most common pattern was only sensory impairment in the ipsilateral face, seen in 30%. Their clinical presentation was unsteadiness while walking, with no bulbar symptoms, shorter hospital stays, and better recovery. Conclusions: LMS is an acute stroke syndrome common in males, often after 50 years, with hypertension, diabetes, and dyslipidemia as systemic risk factors.

Effects of a soft robotic exosuit on the quality and speed of overground walking depends on walking ability after stroke

BackgroundSoft robotic exosuits can provide partial dorsiflexor and plantarflexor support in parallel with paretic muscles to improve poststroke walking capacity. Previous results indicate that baseline walking ability may impact a user’s ability to leverage the exosuit assistance, while the effects on continuous walking, walking stability, and muscle slacking have not been evaluated. Here we evaluated the effects of a portable ankle exosuit during continuous comfortable overground walking in 19 individuals with chronic hemiparesis. We also compared two speed-based subgroups (threshold: 0.93 m/s) to address poststroke heterogeneity.MethodsWe refined a previously developed portable lightweight soft exosuit to support continuous overground walking. We compared five minutes of continuous walking in a laboratory with the exosuit to walking without the exosuit in terms of ground clearance, foot landing and propulsion, as well as the energy cost of transport, walking stability and plantarflexor muscle slacking.ResultsExosuit assistance was associated with improvements in the targeted gait impairments: 22% increase in ground clearance during swing, 5° increase in foot-to-floor angle at initial contact, and 22% increase in the center-of-mass propulsion during push-off. The improvements in propulsion and foot landing contributed to a 6.7% (0.04 m/s) increase in walking speed (R2 = 0.82). This enhancement in gait function was achieved without deterioration in muscle effort, stability or cost of transport. Subgroup analyses revealed that all individuals profited from ground clearance support, but slower individuals leveraged plantarflexor assistance to improve propulsion by 35% to walk 13% faster, while faster individuals did not change either.ConclusionsThe immediate restorative benefits of the exosuit presented here underline its promise for rehabilitative gait training in poststroke individuals.

An overview on bipedal gait control methods

AbstractBipedal gait control has always been a very challenging issue due to the multi‐joint and non‐linear structure of humanoid robots and frequent robot–environment interactions. To realise stable and robust bipedal walking, many aspects including robot modelling, gait stability and environmental adaptivity should be considered to design the gait control method. In this paper, a general description of bipedal gait and the corresponding evaluation indicators are introduced. Moreover, the existing bipedal gait control methods are classified into model‐based gait, stability criterion‐based gait and learning strategy‐based gait and a comprehensive review is conducted. Finally, the existing challenges and development trends of bipedal gait control are presented.

Daily walking kinematic characteristics of the elderly in different residential settings: experimental study on Chinese community-living elderly and long-term nursing home residents.

Long-term nursing home (NH) care helps NH residents with their daily activities and improves their quality of life, but negatively affects their independent physical activities and increases the risk of dangerous events. Dangerous events in the elderly usually occur in the conversion of walking periods when forward striding has already happened, but the body has not yet entered a completely steady walking. Compare the gait characteristics in Chinese long-term NH residents and community-living elderly during the walking Transitional Period (TP) and Stabilization Period (SP). 32 long-term NH residents and 33 age- and sex-matched community-living elderly were recruited. The 30-Second Chair Stand Test (30-s CST), Timed Up and Go Test (TUGT), and Modified Falls Efficacy Scale (MFES) were used to assess their body function. The Xsens MVN BIOMECH system was used to collect and analyze the gait parameters of participants. Compared to community-living elderly, NH residents had fewer numbers of 30-s CST, took more time to complete TUGT, and lower MEFS scores. NH residents showed slower gait speed (P < 0.001), less peak hip flexion (P = 0.022) and extension (P = 0.003), knee internal rotation (P = 0.023), and ankle plantarflexion (P = 0.001) and internal rotation (P = 0.007) angles during walking. When walking progressed from TP to SP, NH residents showed increased ankle dorsiflexion (P < 0.001), decreased hip internal rotation (P < 0.001), and community-living elderly had increased hip extension (P = 0.005) angles. Chinese long-term NH residents had reduced lower extremities strength and postural balance, and higher fear of falling compared to community-living elderly. Their walking performance also showed high fall risk. Besides, long-term NH residents adopted a distal strategy to propel the body forward, which may be a compensatory measure to compensate for inadequate proximal joint control from forward walking to stable walking, and long-term NH residents have reduced postural stability during this process.

Hybrid learning mechanisms under a neural control network for various walking speed generation of a quadruped robot

Legged robots that can instantly change motor patterns at different walking speeds are useful and can accomplish various tasks efficiently. However, state-of-the-art control methods either are difficult to develop or require long training times. In this study, we present a comprehensible neural control framework to integrate probability-based black-box optimization (PIBB) and supervised learning for robot motor pattern generation at various walking speeds. The control framework structure is based on a combination of a central pattern generator (CPG), a radial basis function (RBF) -based premotor network and a hypernetwork, resulting in a so-called neural CPG-RBF-hyper control network. First, the CPG-driven RBF network, acting as a complex motor pattern generator, was trained to learn policies (multiple motor patterns) for different speeds using PIBB. We also introduce an incremental learning strategy to avoid local optima. Second, the hypernetwork, which acts as a task/behavior to control parameter mapping, was trained using supervised learning. It creates a mapping between the internal CPG frequency (reflecting the walking speed) and motor behavior. This map represents the prior knowledge of the robot, which contains the optimal motor joint patterns at various CPG frequencies. Finally, when a user-defined robot walking frequency or speed is provided, the hypernetwork generates the corresponding policy for the CPG-RBF network. The result is a versatile locomotion controller which enables a quadruped robot to perform stable and robust walking at different speeds without sensory feedback. The policy of the controller was trained in the simulation (less than 1 h) and capable of transferring to a real robot. The generalization ability of the controller was demonstrated by testing the CPG frequencies that were not encountered during training.

Humanoid Walking Control Using LQR and ANFIS

Humanoid robots possess remarkable mobility and adaptability for diverse environments. Nonetheless, accurate walking pattern tracking remains challenging, especially when employing the linear quadratic regulator (LQR) due to delays in high-mobility setpoint tracking. We propose a novel control approach to address this limitation by integrating an artificial neuro-fuzzy inference system (ANFIS) with the LQR to enhance pattern tracking. The research contributes to developing a control system that combines LQR and ANFIS to enable humanoid robots to follow various walking patterns with increased precision and efficiency and also the scheme to incorporate LQR and ANFIS. The study involves four experiments: step response, walking phase, static straight walking, and varied straight walking. Each test runs for 5 seconds with a 100-millisecond sampling rate, repeated five times, and employs the Integral Absolute Value (IAE) metric for evaluation. The LQR-ANFIS method exhibits superior performance, achieving a maximum overshoot of 0%, a rise time of 0.3 seconds, a settling time of 0.3 seconds, and a steady-state error of 0% in the step response experiment. The proposed control system also enables stable walking with step periods ranging from 0.15 to 4 seconds and step ranges of 0.05 to 0.03 meters. In conclusion, the integration of ANFIS with the LQR significantly enhances the mobility of humanoid robots, enabling them to navigate diverse environments and accurately track various walking patterns proficiently.

A meta-heuristics based framework of cluster label optimization in MR images using stable random walk

A meta-heuristics based framework of cluster label optimization in MR images using stable random walk

Meta-analysis of the effect of whole-body vibration training on the improvement of limb function in patients with Parkinson's disease.

This study aimed to systematically evaluate the effect of whole-body vibration (WBV) training on the improvement of limb function in patients with Parkinson's disease (PD). An electronic search was performed on the databases, including PubMed, Web of Science, Scopus, SCI-hub, ScienceDirect, Embase, IEEE, Medline, Wiley, ClinicaIKey, CNKI, Wanfang, VIP database, Chinese Medical Association, and CBM Database from inception to May 2022 to collect randomized controlled studies on whole-body vibration training for patients with Parkinson's disease. Two researchers independently screened the literature, extracted the data, and evaluated the quality of the literature, then used ReviewManager 5.4 software for quantitative statistical analysis, including heterogeneity test, sensitivity analysis, risk of bias assessment, combined outcome index effect size and effect size inspection. A total of 9 studies were included in the meta-analysis, including 346 patients, 174 patients in the control group and 172 patients in the observation group. Meta-analysis results showed that, compared with conventional physical therapy or drug therapy alone, whole-body vibration reduced the Unified Parkinson's Disease Rating Scale (UPDRSIII) score and significantly improved the motor function of patients with Parkinson's disease [MD=-2.39, 95% CI (-4.41, -0.37), Z=2.14 (p=0.23)]. Moreover, whole-body vibration significantly improved the walking stability of Parkinson's patients [MD=-1.96, 95% CI (-2.71-1.21), Z=1.17 (p=0.03)]. However, its improvement in balance ability [MD=-0.06, 95% CI (-0.77, 0.65), Z=1.07 (p=0.19)] and daily living ability [MD=0.03, 95% CI (-1.68, 0.74), Z=0.24 (p=0.87)] of patients, it was not t statistically significant. Compared with conventional therapy, WBV has certain advantages in improving the balance function and gait performance of PD patients, but the effect on balance ability and daily living ability is not significant. Thus, more high-quality research is required for further verification.