Experiment-free exoskeleton assistance via learning in simulation

doi:10.21203/rs.3.pex-2632/v1

Download PDF

Method Article

Experiment-free exoskeleton assistance via learning in simulation

https://doi.org/10.21203/rs.3.pex-2632/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 12 Jun, 2024

Read the published version in Nature →

Version 1

posted

You are reading this latest preprint version

Our experimental protocol details the human subject evaluation of a novel, simulation-based approach for optimizing exoskeleton control strategies. Utilizing multi-layer perceptron neural networks, this approach achieves: modeling human kinematics, optimizing muscle coordination, and automating exoskeleton assistance. This method capitalizes on the strengths of data-driven machine learning to refine control strategies, significantly reducing metabolic costs across various activities including walking, running, and stair climbing. This experiment protocol aims to collect metabolic rate and kinematic data during 3 activities (walking, running, stair climbing) in 3 conditions (No Exo, Assist On, Assist Off) to demonstrate the controller’s ability to generate smooth and synergistic assistance to different locomotion activities. The entire protocol can be completed in approximately 30 minutes for each activity of each subject (n=8).

experiment-free

exoskeleton assistance

VO2 Master Analyzer, VO2 Master Health Sensors Inc., Canada
LPMS-B2 Series 9-Axis Inertial Measurement Units, LP-Research Inc., Japan
RMD-X8 1:9, Microactuator Technology Co.
Teensy 4.1, PJRC, USA
Raspberry Pi Zero W, Raspberry Pi Foundation, UK
Bluetooth microcontroller, ItsyBitsy nRF52840 Express, Adafruit, USA
Li-Po battery, 270 g, 22.2 V, 1800 mAh

Activity-adaptive versatile control:

To demonstrate the controller’s ability to adapt to different locomotion activities, we conducted a treadmill experiment for walking and running at three different speeds. The weights and biases in the controller network are taken directly from the simulation, and the inputs are thigh angles and angular velocities measured from one IMU sensor mounted on each thigh. These wearable sensor inputs are used to decode human intention and generate the desired assistive torque of the exoskeleton. The controller consists of a 3-layer neural network and is implemented on a desktop computer running Simulink Realtime. Our method does not require intermediate activity detection or gait cycle detection. The assistance torque is generated in real-time at each time step (100 Hz, namely, 0.01 seconds) using the current thigh angle and thigh angular velocity plus the history data from the past 0.03 seconds (corresponding to three time steps). Through training in the simulation, our controller learned to treat human movement as a continuous process and produces appropriate real-time assistance torque that is synergistic to the user movement during not only steady-state movements but also the transition phases. We conducted treadmill experiments (n=8) for walking at 0.75 m s^-1, 1.25 m s^-1, 1.75 m s^-1, and running at 2 m s^-1. We chose to test across gait speeds to demonstrate the generalizability of the control policy. The assistive torque generated by the controller learned in the simulation is adaptive to walking and running at different speeds.

Continuous assistive torque profile:

To demonstrate the controller’s ability to generate smooth and synergistic assistance for three activities and their transitions, we conducted an activity-varying experiment in the real world for one participant. The participant started from slow walking at approximately 0.8 m s^-1, accelerated to running at approximately 2 m s^-1, then decelerated and finally began stair climbing (7 stairs) in a smooth manner.

Versatile assistive control with metabolic rate reduction:

For each activity: walking (n = 8), running (n = 8), and stair climbing (n = 8), we tested three different conditions, namely, Assist On, Assist Off, and No Exoskeleton. Two 9-axis IMU sensors mounted on both thighs were used to measure the joint angles and angular velocities.

1. Measure the metabolic rate during the Rest condition for 5 min.

2. Measure the metabolic rate, joint angle, joint angular velocity, encoder angle and velocity during the No Exo condition for 5 min.

3. Rest for 5 min.

4. Measure the metabolic rate, joint angle, joint angular velocity, encoder angle and velocity during the Assist On condition for 5 min.

5. Rest for 5 min.

6. Measure the metabolic rate, joint angle, joint angular velocity, encoder angle and velocity during the Assist Off condition for 5 min.

1. Ding, Y., Kim, M., Kuindersma, S. & Walsh, C. J. Human-in-the-loop optimization of hip assistance with a soft exosuit during walking. Science Robotics 3, eaar5438 (2018).

2. Gordon, D. F., McGreavy, C., Christou, A. & Vijayakumar, S. Human-in-the-loop optimization of exoskeleton assistance via online simulation of metabolic cost. IEEE Transactions on Robotics 38, 1410-1429 (2022).

3. Lim, B., Lee, J., Jang, J., Kim, K., Park, Y.J., Seo, K. and Shim, Y. Delayed output feedback control for gait assistance with a robotic hip exoskeleton. IEEE Transactions on Robotics, 35, 1055-1062 (2019).

4. Kim, J., Heimgartner, R., Lee, G., Karavas, N., Perry, D., Ryan, D.L., Eckert-Erdheim, A., Murphy, P., Choe, D.K., Galiana, I. and Walsh, C.J., 2018, May. Autonomous and portable soft exosuit for hip extension assistance with online walking and running detection algorithm. In 2018 IEEE International Conference on Robotics and Automation (ICRA) (pp. 5473-5480). IEEE.

5. Zhang, X., Tricomi, E., Missiroli, F., Lotti, N., Bokranz, C., Nicklas, D. and Masia, L., 2022. Enhancing gait assistance control robustness of a hip exosuit by means of machine learning. IEEE Robotics and Automation Letters, 7(3), pp.7566-7573.

6. Cao, W., Chen, C., Hu, H., Fang, K. & Wu, X. Effect of hip assistance modes on metabolic cost of walking with a soft exoskeleton. IEEE Transactions on Automation Science and Engineering 18, 426-436 (2020).

7. Kim, J., Quinlivan, B.T., Deprey, L.A., Arumukhom Revi, D., Eckert-Erdheim, A., Murphy, P., Orzel, D. and Walsh, C.J., Reducing the energy cost of walking with low assistance levels through optimized hip flexion assistance from a soft exosuit. Scientific Reports, 12(1), p.11004 (2022).

8. Seo, K., Lee, J. & Park, Y. J. Autonomous hip exoskeleton saves metabolic cost of walking uphill. International Conference on Rehabilitation Robotics (ICORR). 246-251 (2017).

9. Nasiri, R., Ahmadi, A. & Ahmadabadi, M. N. Reducing the energy cost of human running using an unpowered exoskeleton. IEEE Transactions on Neural Systems and Rehabilitation Engineering 26, 2026-2032 (2018).

10. Ning, C., Li, Y., Feng, K., Gong, Z. & Zhang, T. Soochow Exo: a Lightweight Hip Exoskeleton Driven by Series Elastic Actuator with Active-type Continuously Variable Transmission. IEEE International Conference on Mechatronics and Automation (ICMA). 1432-1437 (2022).

11. Kim, D.S., Lee, H.J., Lee, S.H., Chang, W.H., Jang, J., Choi, B.O., Ryu, G.H. and Kim, Y.H., 2018. A wearable hip-assist robot reduces the cardiopulmonary metabolic energy expenditure during stair ascent in elderly adults: a pilot cross-sectional study. BMC geriatrics, 18(1), pp.1-8.

12. Kim, J., Lee, G., Heimgartner, R., Arumukhom Revi, D., Karavas, N., Nathanson, D., Galiana, I., Eckert-Erdheim, A., Murphy, P., Perry, D. and Menard, N., 2019. Reducing the metabolic rate of walking and running with a versatile, portable exosuit. Science, 365(6454), pp.668-672.

Download PDF

Journal Publication

published 12 Jun, 2024

Read the published version in Nature →

Version 1

posted

You are reading this latest preprint version

Experiment-free exoskeleton assistance via learning in simulation

Status:

Journal Publication

Version 1

Abstract

Equipment

Procedure

References

Associated Publications

Status:

Journal Publication

Version 1

Privacy Policy

Terms of Service

Experiment-free exoskeleton assistance via learning in simulation

Status:

Journal Publication

Version 1

Abstract

Equipment

Procedure

References

Associated Publications

Status:

Journal Publication

Version 1

Privacy Policy

Terms of Service

Manage Cookie Preferences