Unlike traditional hierarchical controllers for robotic leg prostheses and exoskeletons, continuous systems could allow persons with mobility impairments to walk more naturally in real-world environments without requiring high-level switching between locomotion modes. To support these next-generation controllers, we developed a new system called KIFNet (Kinematics and Image Fusing Network) that uses lightweight and efficient deep learning models to continuously predict the leg kinematics during walking. We tested different sensor fusion methods to combine kinematics data from inertial sensors and computer vision data from smart glasses and found that adaptive instance normalization achieved the lowest RMSE predictions for knee and ankle joint kinematics. We also deployed our model on an embedded device. Without inference optimization, our model was 20 times faster than the previous state-of-the-art and achieved 20% higher prediction accuracies, and during some locomotor activities like stair descent, decreased RMSE up to 300%. With inference optimization, our best model achieved 125 FPS on an NVIDIA Jetson Nano. These results demonstrate the potential to build fast and accurate deep learning models for continuous prediction of leg kinematics during walking based on sensor fusion and embedded computing, therein providing a foundation for real-time continuous controllers for robotic leg prostheses and exoskeletons.

Convolutional neural networks trained using supervised learning can improve visual perception for human-robot walking. These advances have been possible due to largescale datasets like ExoNet and StairNet - the largest open-source image datasets of real-world walking environments. However, these datasets require vast amounts of manually annotated data, the development of which is time consuming and labor intensive. Here we present a novel semi-supervised learning system (ExoNet-SSL) that uses over 1.2 million unlabelled images from ExoNet to improve training efficiency. We developed a deep learning model based on mobile vision transformers and trained the model using semi-supervised learning for image classification. Compared to standard supervised learning (98.4%), our ExoNet-SSL system was able to maintain high prediction accuracy (98.8%) when tested on previously unseen environments, while requiring 35% fewer labelled images during training. These results show that semi-supervised learning can improve training efficiency by leveraging large amounts of unlabelled data and minimize the size requirements for manually annotated images. Future research$\text{will}$focus on model deployment for onboard real-time inference and control of human-robot walking.