Understanding LSTM Network Behaviour of IMU-Based Locomotion Mode Recognition for Applications in Prostheses and Wearables

Abstract

Human Locomotion Mode Recognition (LMR) has the potential to be used as a control mechanism for lower-limb active prostheses. Active prostheses can assist and restore a more natural gait for amputees, but as a medical device it must minimize user risks, such as falls and trips. As such, any control system must have high accuracy and robustness, with a detailed understanding of its internal operation. Long Short-Term Memory (LSTM) machine-learning networks can perform LMR with high accuracy levels. However, the internal behavior during classification is unknown, and they struggle to generalize when presented with novel users. The target problem addressed in this paper is understanding the LSTM classification behavior for LMR. A dataset of six locomotive activities (walking, stopped, stairs and ramps) from 22 non-amputee subjects is collected, capturing both steady-state and transitions between activities in natural environments. Non-amputees are used as a substitute for amputees to provide a larger dataset. The dataset is used to analyze the internal behavior of a reduced complexity LSTM network. This analysis identifies that the model primarily classifies activity type based on data around early stance. Evaluation of generalization for unseen subjects reveals low sensitivity to hyper-parameters and over-fitting to individuals’ gait traits. Investigating the differences between individual subjects showed that gait variations between users primarily occur in early stance, potentially explaining the poor generalization. Adjustment of hyper-parameters alone could not solve this, demonstrating the need for individual personalization of models. The main achievements of the paper are (i) the better understanding of LSTM for LMR, (ii) demonstration of its low sensitivity to learning hyper-parameters when evaluating novel user generalization, and (iii) demonstration of the need for personalization of ML models to achieve acceptable accuracy.

Artificial Intelligence-Based Wearable Robotic Exoskeletons for Upper Limb Rehabilitation: A Review

“…Additionally, it is noteworthy the use of AI-based signal processing or control techniques that may bring with it some design limitations, as current AI systems may still suffer from poor explainability [100], explicit or implicit bias, or other problems related to the machine learning models [101,102], factors that may or may not manifest in a predictable way when it comes to the interaction between a human user or patient and the rehabilitation system. In view of the above, it is advisable to invest further efforts and developments in explainable AI systems that allow for a complete understanding of learning models when they are involved in critical systems such as robotic rehabilitation devices.…”

Locomotion Mode Recognition with Inertial Signals for Hip Joint Exoskeleton

“…The evaluation method also influenced the numeric recognition results. For instance, the recent studies on IMU-based locomotion mode recognition achieved >95% average recognition accuracies with intrasubject crossvalidation [21,22]. The sensor setups are quite different from ours.…”

“…The sensor setups are quite different from ours. The study mounted an IMU board on the amputated foot for terrain identification [21], while the study fixed the IMU boards on the shanks, the waist, and wrists [22].…”

A Novel Cryptocurrency Price Prediction Model Using GRU, LSTM and bi-LSTM Machine Learning Algorithms

“…The LSTM is an RNN-style architecture with gates that govern the flow of information between cells. The input and forget gate structures can modify information traveling along the cell state, with the ultimate output being a filtered version of the cell state based on context from the inputs [44]. The LSTM design has been criticized for being ad hoc and for having a large number of components whose purpose is not immediately clear.…”