This study determined whether the kinematics of lower limb trajectories during walking could be extrapolated using long short-term memory (LSTM) neural networks. It was hypothesised that LSTM auto encoders could reliably forecast multiple time-step trajectories of the lower limb kinematics, specifically linear acceleration (LA) and angular velocity (AV). Using 3D motion capture, lower limb position-time coordinates were sampled (100 Hz) from six male participants (age 22 ± 2 years, height 1.77 ± 0.02 m, body mass 82 ± 4 kg) who walked for 10 min at 5 km/h on a 0% gradient motordriven treadmill. These data were fed into an LSTM model with a sliding window of four kinematic variables with 25 samples or time steps: LA and AV for thigh and shank. The LSTM was tested to forecast five samples (i.e., time steps) of the four kinematic input variables. To attain generalisation, the model was trained on a dataset of 2,665 strides from five participants and evaluated on a test set of 1 stride from a sixth participant. The LSTM model learned the lower limb kinematic trajectories using the training samples and tested for generalisation across participants. The forecasting horizon suggested higher model reliability in predicting earlier future trajectories. The mean absolute error (MAE) was evaluated on each variable across the single tested stride, and for the fivesample forecast, it obtained 0.047 m/s 2 thigh LA, 0.047 m/s 2 shank LA, 0.028 deg/s thigh AV and 0.024 deg/s shank AV. All predicted trajectories were highly correlated with the measured trajectories, with correlation coefficients greater than 0.98. The motion prediction model may have a wide range of applications, such as mitigating the risk of falls or balance loss and improving the human-machine interface for wearable assistive devices.

Enhancing the capabilities of the dismounted combatant has been an enduring goal of international military research communities. Emerging developments in exoskeleton technology offers the potential to augment the dismounted combatant's capabilities. However, the ability to determine the value proposition of an exoskeleton in a military context is difficult due to the variety of methods and metrics used to evaluate previous devices. The aim of this paper was to present a standard framework for the evaluation and assessment of exoskeletons for use in the military. A structured and systematic methodology was developed from the end-user perspective and progresses from controlled laboratory conditions (Stage A), to simulated movements specific to the dismounted combatant (Stage B), and real-world military specific tasks (Stage C). A standard set of objective and subjective metrics were described to ensure a holistic assessment on the human response to wearing the exoskeleton and the device's mechanical performance during each stage. A standardised methodology will ensure further advancement of exoskeleton technology and support improved international collaboration across research and industry groups. In doing so, this better enables international military groups to evaluate a system's potential, with the hope of accelerating the maturity and ultimately the fielding of devices to augment the dismounted close combatant and small team capability.

Objective The aim of this review was to determine how exoskeletons could assist Australian Defence Force personnel with manual handling tasks. Background Musculoskeletal injuries due to manual handling are physically damaging to personnel and financially costly to the Australian Defence Force. Exoskeletons may minimize injury risk by supporting, augmenting, and/or amplifying the user’s physical abilities. Exoskeletons are therefore of interest in determining how they could support the unique needs of military manual handling personnel. Method Industrial and military exoskeleton studies from 1990 to 2019 were identified in the literature. This included 67 unique exoskeletons, for which Information about their current state of development was tabulated. Results Exoskeleton support of manual handling tasks is largely through squat/deadlift (lower limb) systems (64%), with the proposed use case for these being load carrying (42%) and 78% of exoskeletons being active. Human–exoskeleton analysis was the most prevalent form of evaluation (68%) with reported reductions in back muscle activation of 15%–54%. Conclusion The high frequency of citations of exoskeletons targeting load carrying reflects the need for devices that can support manual handling workers. Exoskeleton evaluation procedures varied across studies making comparisons difficult. The unique considerations for military applications, such as heavy external loads and load asymmetry, suggest that a significant adaptation to current technology or customized military-specific devices would be required for the introduction of exoskeletons into a military setting. Application Exoskeletons in the literature and their potential to be adapted for application to military manual handling tasks are presented.

No abstract

The forecasting of lower limb trajectories can improve the operation of assistive devices and minimise the risk of tripping and balance loss. The aim of this work was to examine four Long Short Term Memory (LSTM) neural network architectures (Vanilla, Stacked, Bidirectional and Autoencoders) in predicting the future trajectories of lower limb kinematics, i.e. Angular Velocity (AV) and Linear Acceleration (LA). Kinematics data of foot, shank and thigh (LA and AV) were collected from 13 male and 3 female participants (28 ± 4 years old, 1.72 ± 0.07 m in height, 66 ± 10 kg in mass) who walked for 10 minutes at preferred walking speed (4.34 ± 0.43 km.h-1) and at an imposed speed (5km.h-1, 15.4% ± 7.6% faster) on a 0% gradient treadmill. The sliding window technique was adopted for training and testing the LSTM models with total kinematics time-series data of 10,500 strides. Results based on leave-one-out cross validation, suggested that the LSTM autoencoders is the top predictor of the lower limb kinematics trajectories (i.e. up to 0.1s). The normalised mean squared error was evaluated on trajectory predictions at each time-step and it obtained 2.82–5.31% for the LSTM autoencoders. The ability to predict future lower limb motions may have a wide range of applications including the design and control of bionics allowing improved human-machine interface and mitigating the risk of falls and balance loss.