Adaptive robot mediated upper limb training using electromyogram-based muscle fatigue indicators

Abstract

Studies on improving the adaptability of upper limb rehabilitation training do not often consider the implications of muscle fatigue sufficiently. In this study, electromyogram features were used as fatigue indicators in the context of human-robot interaction. They were utilised for auto-adaptation of the task difficulty, which resulted in a prolonged training interaction. The electromyogram data was collected from three gross-muscles of the upper limb in 30 healthy participants. The experiment followed a protocol for increasing the muscle strength by progressive strength training, that was an implementation of a known method in sports science for muscle training, in a new domain of robotic adaptation in muscle training. The study also compared how the participants in three experimental conditions perceived the change in task difficulty levels. One task benefitted from robotic adaptation (Intervention group) where the robot adjusted the task difficulty. The other two tasks were control groups 1 and 2. There was no difficulty adjustment at all in Control 1 group and the difficulty was adjusted manually in Control 2 group. The results indicated that the participants could perform a prolonged progressive strength training exercise with more repetitions with the help of a fatigue-based robotic adaptation, compared to the training interactions, which were based on manual/no adaptation. This study showed that it is possible to alter the level of the challenge using fatigue indicators, and thus, increase the interaction time. The results of the study are expected to be extended to stroke patients in the future by utilising the potential for adapting the training difficulty according to the patient's muscular state, and also to have a large number repetitions in a robot-assisted training environment.

Fig 1. Context of the study.

HapticMaster robot was configured to adapt its environment based on the detected muscle fatigue. EMG measured from upper limb muscles was used to detect the fatigue. The adaptive strength training algorithm forms a closed loop, as shown. Kinematic measurements, detected fatigue, reported fatigue, and demographic measurements were also analysed off-line.

Fig 2. Experimental setup.

The experimental setup included HapticMaster robot, visual guidance and animated background. The front-end of the rowing boat was shown on the LCD display in front of the participant. The rowing environment was embedded with audio cues and haptic sensation of underwater viscosity.

Fig 3. EMG signal acquisition setup.

The EMG acquisition device with g.USBamp amplifier, three bipolar electrodes and a ground electrode.

Fig 4. EMG electrode locations.

EMG electrodes are connected to three upper limb muscle locations (Biceps Brachii (BB), Anterior Deltoid (DLTF), and Middle Deltoid (DLTM)). The ground electrode was connected to a bony area near the elbow.

Fig 5. Experiment protocol.

Description of the different stages of the experiment protocol.

Fig 6. Simulink model for EMG data acquisition.

The model has an on-line signal processing algorithm that performs fatigue detection for each muscle simultaneously. The detected fatigue state was communicated to the HapticMaster control algorithm.

Fig 7. Algorithm for robotic adaptation.

The flow chart of the adaptation algorithm for Intervention group participants.

Fig 8. Progress of EMG median frequency.

Median frequency in a typical Intervention group participant (Subject 20), who received adaptive robotic assistance based on the detected muscle fatigue using EMG features. The doted regions represent a significant decrease in median frequency, which resulted in the detection of fatigue.

Fig 9. Status of fatigue flags.

The fatigue flags in a typical Intervention group participant (Subject 20), who received adaptive robotic assistance based on the detected muscle fatigue. The doted regions represent the detection of fatigue in the DLTM muscle, which decided the final state of fatigue.

Fig 10. Task difficulty in Control-1 group.

The progress of task difficulty in Control-1 participants, who received 30 seconds break period after each trial of 1-minute duration before the MVC+ increment. This group did not receive any robotic adaptation based on muscle fatigue.

Fig 11. Task difficulty in Control-2 group.

The progress of task difficulty in Control-2 group participants, who received a manual robotic adaptation based on the subjective fatigue reported.

Fig 12. Task difficulty in Intervention group.

The progress of task difficulty in Intervention group participants who received an automatic robotic adaptation based on the detected fatigue using EMG features.

Fig 13. Box plots for task duration.

Box plots showing the duration of the experiment in the three groups of participants.

Fig 14. Box plots for number of repetitions.

Box plots showing the number of repetitions of the rowing task in the three groups of participants. The Intervention group could do more task repetitions due to the auto-adaptation of the task difficulty during the progressively challenging exercise.

Fig 15. Box plots for speed of repetitions.

Box plots showing the rate of task repetitions (repetitions/minute) of the rowing task in the three groups of participants.

Fig 16. Box plots for time-to-fatigue.

Box plots showing the time taken to reach the first reported state of fatigue in the three groups of participants.