A locomotion intent prediction system based on multi-sensor fusion - PubMed (original) (raw)

A locomotion intent prediction system based on multi-sensor fusion

Baojun Chen et al. Sensors (Basel). 2014.

Abstract

Locomotion intent prediction is essential for the control of powered lower-limb prostheses to realize smooth locomotion transitions. In this research, we develop a multi-sensor fusion based locomotion intent prediction system, which can recognize current locomotion mode and detect locomotion transitions in advance. Seven able-bodied subjects were recruited for this research. Signals from two foot pressure insoles and three inertial measurement units (one on the thigh, one on the shank and the other on the foot) are measured. A two-level recognition strategy is used for the recognition with linear discriminate classifier. Six kinds of locomotion modes and ten kinds of locomotion transitions are tested in this study. Recognition accuracy during steady locomotion periods (i.e., no locomotion transitions) is 99.71% ± 0.05% for seven able-bodied subjects. During locomotion transition periods, all the transitions are correctly detected and most of them can be detected before transiting to new locomotion modes. No significant deterioration in recognition performance is observed in the following five hours after the system is trained, and small number of experiment trials are required to train reliable classifiers.

PubMed Disclaimer

Figures

Figure 1.

Figure 1.

The placement of the sensors on human body. Foot pressure insoles are placed in shoes of both feet and the signals are sampled by circuits embedded in FP module1 and FP module2. IMUs used for recording the movement data of the thigh, the shank and the foot are embedded in IMU module1, IMU module2 and FP module1, respectively.

Figure 2.

Figure 2.

(a) The structure of data transmission in the designed system using RS-485 bus. In each sensor module, the data stream is controlled by MCU; (b) The custom made foot pressure insoles used in this study. Positions of pressure sensors are marked as 1, 2, 3 and 4, respectively; (c) The IMU board embedded with an accelerometer (ACCEL), a gyroscope (GYRO) and a magnetometer (MAG); (d) The foot pressure sampling circuit.

Figure 3.

Figure 3.

Block diagram of the locomotion intent prediction system.

Figure 4.

Figure 4.

Average recognition accuracy (a) and adjusted prediction time (b) over seven able-bodied subjects with window size ranging from 100 ms to 200 ms. Color shades denote SEMs across subjects.

Figure 5.

Figure 5.

Average recognition error (a) and adjusted prediction time (b) over seven able-bodied subjects with the original majority voting and the modified post-processing approach. Error bars denote SEMs across subjects.

Figure 6.

Figure 6.

Average recognition accuracy (a) and adjusted prediction time (b) over seven able-bodied subjects with some time ranging from 15 min to 300 min after training. Color shades denote SEMs across subjects.

Figure 7.

Figure 7.

Average recognition accuracy (a) and adjusted prediction time (b) over seven able-bodied subjects with the number of training pairs ranging from 1 to 20. Color shades denote SEMs across subjects.

Figure 8.

Figure 8.

Recognition performances of online test experiments for Set-A experiment trial (a) and Set-B experiment trial (b). Data of 6 experiment pairs were used to train the system. The solid line denotes decision stream made by the locomotion intent recognition system. The dash line denotes the critical moment for each locomotion transition. Color shades denote locomotion transition periods of each trial.

References

    1. Au S., Weber J., Herr H. Powered ankle-foot prosthesis improves walking metabolic economy. IEEE Trans. Robot. 2009;25:51–66.
    1. Sup F., Varol H., Mitchell J., Withrow T., Goldfarb M. Preliminary evaluations of a self-contained anthropomorphic transfemoral prosthesis. IEEE/ASME Trans. Mechatron. 2009;14:667–676. -PMC -PubMed
    1. Hitt J., Sugar T., Holgate M., Bellmann R., Hollander K. Robotic transtibial prosthesis with biomechanical energy regeneration. Ind. Robot. 2009;36:441–447.
    1. Versluys R., Desomer A., Lenaerts G., van Damme M., Bey P., Van der Perre G., Peeraer L., Lefeber D. A pneumatically powered below-knee prosthesis: Design specifications and first experiments with an amputee. Proceedings of the 2nd IEEE/RAS-EMBS International Conference on Biomedical Robotics and Biomechatronics; Scottsdale, AZ, USA. 19–22 October 2008; pp. 372–377.
    1. Bellman R., Holgate M., Sugar T. SPARKy 3: Design of an active robotic ankle prosthesis with two actuated degrees of freedom using regenerative kinetics. Proceedings of the 2nd IEEE/RAS-EMBS International Conference on Biomedical Robotics and Biomechatronics; Scottsdale, AZ, USA. 19–22 October 2008; pp. 511–516.

Publication types

MeSH terms

LinkOut - more resources