Real-Time On-Board Recognition of Continuous Locomotion Modes for Amputees With Robotic Transtibial Prostheses

Xu, Dongfang; Feng, Yanggang; Mai, Jingeng; Wang, Qining

doi:10.1109/tnsre.2018.2870152

Cited by 58 publications

(59 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A sliding window was selected to extract features of raw signals. Each window's length was 250 ms and the sliding increment was 10 ms (Zheng and Wang, 2016;Xu et al, 2018). Five time domain features were selected for this study.…”

Section: Signal Feature Extractingmentioning

confidence: 99%

“…SVM, QDA, and LDA were widely used in locomotion mode recognition (Liu et al, 2017). In the previous study, we have compared the on-board recognition performances of different algorithms (SVM, QDA, and LDA) (Xu et al, 2018). SVM algorithm could achieve high recognition accuracy, LDA could achieve good recognition time performance, and QDA could take count of accuracy and recognition time performance (Xu et al, 2018).…”

Section: Recognition Algorithmmentioning

confidence: 99%

“…In the previous study, we have compared the on-board recognition performances of different algorithms (SVM, QDA, and LDA) (Xu et al, 2018). SVM algorithm could achieve high recognition accuracy, LDA could achieve good recognition time performance, and QDA could take count of accuracy and recognition time performance (Xu et al, 2018). For multi-class (five locomotion modes in this study) recognition, a one vs. one strategy was adopted.…”

Section: Recognition Algorithmmentioning

confidence: 99%

“…Hargrove et al have also developed on-line recognition by fusing sEMG and mechanical sensors (Spanias et al, 2016(Spanias et al, , 2018. We have also developed realtime on-board recognition of continuous locomotion modes for amputees with robotic transtibial prostheses and got more than 93% recognition accuracy with just two IMUs (Xu et al, 2018).…”

Section: Introductionmentioning

confidence: 99%

“…As is known, real-time recognition based on the off-line trained models is meaningful for the on-line control; however, offline training brings a series of disadvantages for real-time recognition. Off-line training means bringing in other devices (e.g., a computer) to train the model, which is not convenient in integration with robotic prosthetics (Zhang et al, 2015;Xu et al, 2018). To improve the problem, Spanias et al (2018) conducted a model updated on an embedded micro-controller based on mechanical sensor and sEMG.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

On-board Training Strategy for IMU-Based Real-Time Locomotion Recognition of Transtibial Amputees With Robotic Prostheses

Wang

2020

Front. Neurorobot.

Self Cite

View full text Add to dashboard Cite

The paper puts forward an on-board strategy for a training model and develops a real-time human locomotion mode recognition study based on a trained model utilizing two inertial measurement units (IMUs) of robotic transtibial prosthesis. Three transtibial amputees were recruited as subjects in this study to finish five locomotion modes (level ground walking, stair ascending, stair descending, ramp ascending, and ramp descending) with robotic prostheses. An interaction interface was designed to collect sensors' data and instruct to train model and recognition. In this study, the analysis of variance ratio (no more than 0.05) reflects the good repeatability of gait. The on-board training time for SVM (Support Vector Machines), QDA (Quadratic Discriminant Analysis), and LDA (Linear discriminant analysis) are 89, 25, and 10 s based on a 10,000 × 80 training data set, respectively. It costs about 13.4, 5.36, and 0.067 ms for SVM, QDA, and LDA for each recognition process. Taking the recognition accuracy of some previous studies and time consumption into consideration, we choose QDA for real-time recognition study. The real-time recognition accuracies are 97.19 ± 0.36% based on QDA, and we can achieve more than 95% recognition accuracy for each locomotion mode. The receiver operating characteristic also shows the good quality of QDA classifiers. This study provides a preliminary interaction design for human-machine prosthetics in future clinical application. This study just adopts two IMUs not multi-type sensors fusion to improve the integration and wearing convenience, and it maintains comparable recognition accuracy with multi-type sensors fusion at the same time.

show abstract

Section: Signal Feature Extractingmentioning

confidence: 99%