Deep-Learning to Map a Benchmark Dataset of Non-amputee Ambulation for
Controlling an Open Source Bionic Leg
Abstract
Powered lower limb prosthetic devices may be becoming a promising option
for amputation patients. Although various methods have been proposed to
produce gait trajectories similar to non-disabled individuals,
implementing these control methods is still challenging. It remains
unclear whether these methods provide appropriate, safe, and intuitive
locomotion as users intend. This paper proposes a direct mapping of
voluntary movement of a residual limb (i.e., thigh) to the desired
movement of amputated limbs (i.e., knee and ankle) to control the
prosthetic legs. The proposed model was learned with the gait trajectory
of intact limb individuals from a publicly available biomechanics
dataset and applied to control the prosthetic leg without post-tuning
the network. Thus, the proposed method does not require the training
time with amputation patients nor configuration time to use while
providing a closely resembling gait trajectory of the intact limb. As
preliminary testing, three able-bodied subjects participated in bypass
tests. The proposed model accomplished intuitive and reliable
level-ground walking at three different step lengths: self-selected,
long- and short-steps. The results indicate that intact benchmark data
can be directly used to train the model to control prosthetic legs.