Gait reference trajectory generation at different walking speeds using LSTM and CNN

Semwal, Vijay Bhaskar; Jain, Rahul; Maheshwari, Pushkar; Khatwani, Saksham

doi:10.1007/s11042-023-14733-2

Cited by 21 publications

(10 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, using the above-mentioned deep algorithms alone to capture and simulate human gait predictions will have large errors. Instead, the classification and simulation prediction of human gait can be made more accurate by using CNN combined with LSTM [7] [8], multiscale learning model (MSL) [9], deep neural network library with long and short-term memory (cuDNN LSTM) integrated with improved RNN [10], multilayer perceptron using nonlinear dynamics [11] and a two-layer feedforward neural network combination using WS classifier and PGC predictor [12] for simulating individual gait behavior.…”

Section: Gait Recognition Using Deep Learning Algorithmsmentioning

confidence: 99%

“…The data inputs for the above combined approach are mainly divided into neurophysiological signals such as EEG and EMG [8] [9] [12], 3D motion capture information [7] [10] [11] and kinetic data information [12]. And the use of the above combined approach has a large requirement for the initial input data.…”

Section: Gait Recognition Using Deep Learning Algorithmsmentioning

confidence: 99%

“…CNNs mainly perform feature extraction and dimensionality reduction on images through convolutional and pooling layers [8], thus reducing the number of parameters and easing the complexity of the computation. RNNs capture the temporal information in sequential data through cyclic layers and state transfer, thus modeling long-term dependencies, and have problems such as gradient disappearance and gradient explosion while effectively reducing the number of parameters and training time of the model [7]. LSTM effectively mitigates the gradient vanishing problem of RNNs by introducing mechanisms such as forgetting gates, input gates and output gates [8].…”

Section: Gait Recognition Using Deep Learning Algorithmsmentioning

confidence: 99%

See 2 more Smart Citations

An analysis of gait capture and simulation techniques for lower limb exoskeleton robots for stroke rehabilitation

Jin

2023

ACE

View full text Add to dashboard Cite

Lower extremity rehabilitation-assisted exoskeleton robots bring together multiple disciplines such as biomechanics, control engineering, robotics, and computer science. The main role of lower extremity rehabilitation exoskeleton robots is to help patients and rehabilitators to maintain or restore the mobility of lower extremities, therefore, proper research and discussion of human gait analysis is the basis for establishing and improving such exoskeleton robots. To analyse and evaluate the positive effects of modern technology on stroke patients, the development of human gait capture and simulation technologies will be mainly summarized, the latest human gait capture and simulation technologies will be classified and evaluated. And through a comprehensive review and analysis of research advances in these areas, the usefulness of lower extremity rehabilitation-assisted exoskeleton robots for stroke populations is evaluated. This paper is informative in studying the usefulness of a lower limb exoskeleton robot combined with gait capture and simulation technology for rehabilitation training of stroke patients.

show abstract

Section: Gait Recognition Using Deep Learning Algorithmsmentioning

confidence: 99%

Section: Gait Recognition Using Deep Learning Algorithmsmentioning

confidence: 99%

Section: Gait Recognition Using Deep Learning Algorithmsmentioning

confidence: 99%

See 1 more Smart Citation

An analysis of gait capture and simulation techniques for lower limb exoskeleton robots for stroke rehabilitation

Jin

2023

ACE

View full text Add to dashboard Cite

show abstract

“…Gait analysis is a methodical examination of human locomotion that gained attention within the area of rehabilitation engineering [1][2][3]. The objective of rehabilitation engineering is to create advanced assistive devices that offer technological support to individuals with disabilities.…”

Section: Introductionmentioning

confidence: 99%

Analyzing Lower Limb Dynamics in Human Gait Using Average Value-Based Technique

Sunny,

Sivanandan,

Parameswaran

et al. 2024

Mathematical Problems in Engineering

View full text Add to dashboard Cite

The motivation of this study is to develop effective and economical assistive technologies for people with physical disabilities. The novelty in this manuscript is the application of the average value-based technique to accurately represent the involved biomechanics of the lower limb joints during the human gait cycle. This mathematical formulation of lower limb joints’ biomechanics forms the first objective for modeling and final exoskeleton prototype development. To account for modeling the characteristics of human locomotion, the nth-order linear differential equation with constant coefficients is considered with appropriate modification. The physical characteristics of an individual are represented by the constant coefficients (P0, P1, P2, and P3) of the modified infinite series, which are obtained by processing experimental data collected using an optical technique. The differential terms of the infinite series are replaced by difference terms (δbavg, δ2bavg, and δ3bavg) since the data were captured as a set of digital values. The work presented here is based on the experimental results of individuals suitably categorized according to their physical nature like age and other physical structure. The optically monitored positional values of the lower limb joints of the individual subjects while they are completing the gait cycles are used for getting values of different terms of the model. The data collected through the conduct of experiments are used for finding the values of the terms of the differential equation. The model is effectively validated through experimental results. It was determined that the representation’s accuracy fell within the ±5% acceptable tolerance limit. The model is prepared for healthy as well as disabled persons, through which the disability is quantified. The resulting model can be used to develop assistive devices for people with physical disabilities. This results in the rehabilitation process and thereby helps the reintegration into society, subsequently allowing them to lead a normal life.

show abstract

“…Consequently, in the work of GPM, it is necessary to establish the connection between body parameters and gait, as well as predict gait characteristics. [10] have successfully predicted gait reference trajectories at various speeds using Long Short Term Memory Neural Networks (LSTM) and Convolutional Neural Networks (CNN). Similarly, [11] utilized an SAE-LSTM network to predict gait reference trajectories by employing motion acceleration data from the thighs, calves, feet, as well as angular velocity, angle, and other nine information quantities, along with three plantar pressure information.…”

Section: Introductionmentioning

confidence: 99%

Personalized lower limb gait reconstruction modeling based on RFA-ProMP

Chunhong,

Kang,

Zhiqin

et al. 2023

Preprint

View full text Add to dashboard Cite

To generate personalized gait reference trajectories for lower limb rehabilitation exoskeletons tailored to patients with different body types, we propose a combination of a random forest algorithm (RFA) with an attention mechanism and a mimicry learning method called Probabilistic Movement Primitives (ProMP). This approach aims to establish a personalized lower limb gait reconstruction model. Firstly, the RFA is utilized to map the individual body characteristic parameters and the relationship between gait speed and gait. It predicts the gait characteristics under a specified gait speed. Then, the ProMP is introduced to learn the probability distribution of normal gait data and perform the lower limb gait reconstruction based on the predicted gait characteristics. Experimental results demonstrate that the hip and knee curves reconstructed by RFA-ProMP exhibit the smallest error. Furthermore, the generated reference trajectory aligns with the original probabilistic reference trajectory. This method proves effective in generating personalized gait curves for the lower limbs.

show abstract

Gait reference trajectory generation at different walking speeds using LSTM and CNN

Cited by 21 publications

References 27 publications

An analysis of gait capture and simulation techniques for lower limb exoskeleton robots for stroke rehabilitation

An analysis of gait capture and simulation techniques for lower limb exoskeleton robots for stroke rehabilitation

Analyzing Lower Limb Dynamics in Human Gait Using Average Value-Based Technique

Personalized lower limb gait reconstruction modeling based on RFA-ProMP

Contact Info

Product

Resources

About