Data-Driven Gait Segmentation for Walking Assistance in a Lower-Limb Assistive Device

  title={Data-Driven Gait Segmentation for Walking Assistance in a Lower-Limb Assistive Device},
  author={Aleksandra Kalinowska and Thomas A. Berrueta and Adam Zoss and Todd D. Murphey},
  journal={2019 International Conference on Robotics and Automation (ICRA)},
Hybrid systems, such as bipedal walkers, are challenging to control because of discontinuities in their nonlinear dynamics. Little can be predicted about the systems’ evolution without modeling the guard conditions that govern transitions between hybrid modes, so even systems with reliable state sensing can be difficult to control. We propose an algorithm that allows for determining the hybrid mode of a system in real-time using data-driven analysis. The algorithm is used with data-driven… Expand
Application of Interacting Models to Estimate the Gait Speed of an Exoskeleton User
An interacting multi-model (IMM) estimation scheme is presented to address state estimation for lower-extremity exoskeletons and to handle their hybrid dynamics, which includes new approaches that enable it to estimate states of hybrid systems with dynamics that are unique to each phase. Expand
Lower limb rehabilitation exoskeleton robot: A review
Lower limb rehabilitation exoskeleton robots (LLRERs) play a positive role in lower limb rehabilitation and assistance for patients with lower limb disorders, and they are helpful to improveExpand
Accelerometer-Based Gait Segmentation: Simultaneously User and Adversary Identification
A new gait segmentation method based on accelerometer data and a new distance function between two time series are introduced, showing novel and effectiveness in simultaneously identifying user and adversary. Expand
The Effect of Sensor Fusion on Data-Driven Learning of Koopman Operators
It is shown that OC-KOs are effective for sensor fusion by proving that when learning a Koopman operator, output measurement functions serve to constrain the space of potential Koop man observables and their eigenfunctions. Expand


Identifying Gait Phases from Joint Kinematics during Walking with Switched Linear Dynamical Systems*
  • Luke Drnach, Irfan Essa, L. Ting
  • Computer Science
  • 2018 7th IEEE International Conference on Biomedical Robotics and Biomechatronics (Biorob)
  • 2018
This work used Switched Linear Dynamical Systems to model joint angle kinematic data from healthy individuals walking on a treadmill with normal gaits and with gaits perturbed by electrical stimulation and found that SLDS models accounted for over 88% of the variation in each joint angle. Expand
Model-Based Control Using Koopman Operators
It is illustrated how the Koopman operator can be used to obtain a linearizable data-driven model for an unknown dynamical process that is useful for model-based control synthesis. Expand
An adaptive gait learning strategy for lower limb exoskeleton robot
The lower limb joint angles were used as the gait feature, the joint angle of lower limb are important gait kinematics parameter, and the joint angles and the pressure distribution of the foot were got through the motion information acquisition system and then the effective features of the signals were extracted. Expand
Online User Assessment for Minimal Intervention During Task-Based Robotic Assistance
The mode insertion gradient (MIG) is used as a filtering criterion that instantaneously assesses the impact of user actions on a dynamic system over a time window into the future and exhibits a low, but significant, negative correlation between skill level and rate of controller intervention during assistance. Expand
A practical guide to gait analysis.
Motion analysis, in particular gait analysis, provides objective preoperative and postoperative data for outcome assessment and has contributed to the development of orthotics and new surgical techniques. Expand
Gaussian Processes for Data-Efficient Learning in Robotics and Control
This paper learns a probabilistic, non-parametric Gaussian process transition model of the system and applies it to autonomous learning in real robot and control tasks, achieving an unprecedented speed of learning. Expand
Sequential Action Control: Closed-Form Optimal Control for Nonlinear and Nonsmooth Systems
A new model-based algorithm that computes predictive optimal controls online and in a closed loop for traditionally challenging nonlinear systems and can avoid local minima and outperform nonlinear optimal controllers and recent case-specific methods in terms of tracking performance and at speeds that are orders of magnitude faster than traditionally achievable ones. Expand
Model learning for robot control: a survey
This paper surveys the progress in model learning with a strong focus on robot control on a kinematic as well as dynamical level and deduces future directions of real-time learning algorithms. Expand
ED-FNN: A New Deep Learning Algorithm to Detect Percentage of the Gait Cycle for Powered Prostheses
A novel gait percent detection algorithm that can predict a full gait cycle discretised within a 1% interval is proposed and is called an exponentially delayed fully connected fully connected neural network (ED-FNN). Expand
Mechanically induced stumbling during human treadmill walking
Both the EMG and kinesiologic responses were reproducible when perturbations were presented in the same part of the swing phase of different step cycles. Expand