Team Member
- Jaehyun Bae
Motivation
DARPA Warrior Web Program
This project is motivated by DARPA Warrior Web Program.
- There has been a lot of research into exoskeletons over the years to alleviate heavy loads that soldiers should burden, but strapping a person into a robotic outfit just isn't practical in a combat zone yet.
- DARPA's Warrior Web program aims to build a lightweight suit that improves a soldier's endurance and overall effectiveness, while preventing injuries.
- The main goals by developing the warrior web are
- To prevent and reduce musculoskeletal injuries.
- To augment positive work done by the muscles and reduce the physical burden
- You can see the details on http://www.darpa.mil/Our_Work/DSO/Programs/Warrior_Web.aspx
Harvard Exosuit
In order to develop an under-suit that doesn’t interrupt wearer’s free movement, researchers are trying to make it soft and deformable, but still capable of applying force to body joints. Among several different projects in the Warrior Web program, I focused on Harvard exosuit.
- Harvard Biodesign group, one of a project groups in this program, is trying to make their warrior web suit soft and light, and they call their suit harvard exosuit.
- Experimental data proved that it can help loaded walking by reducing metabolic costs.
Challenges
As it is a new approach to assist human gait with deformable structure, there are many challenges while developing the exosuit. The examples of the challenges are
- It is difficult to analyze the effectivness of exist.
- It is difficult to find the optimal input force for actuators to reduce the metabolic cost
- It is difficult to identify the effect of change of design parameters
The reasons for the challenges are
- The under-suit may be soft and deformable
- Hard to identify how external actuation assists loaded gait.
- Experimental metabolic cost data is inconsistent case by case
Project Goals
Through this project, I tried to resolve the challenges in developing exosuit with Opensim simulation. Simulation can help developing soft wearable exosuit as it can gives an intuition on how exosuit help muscles, and what are the key features that one should care about when developing the suit.
I hope this project will construct a systematic way of analyzing and designing soft wearable device. The goal that I set as a starting point are
- Evaluate the effectiveness of wearing active actuator on metabolic cost reduction during loaded walking.
- Explain how exosuit can help loaded gait.
- Verify the impact of changes in design parameters.
- Find optimal control inputs for exosuit actuators.
- Evaluate the effectiveness of Biarticular actuator
Strategy
Experimental data
Two different data were collected from a same subject.
- One gait cycle of loaded walking (From left toe off to next left toe off)
- One gait cycle of unloaded walking (From left toe off to next left toe off)
For both data, walking speed is identical, and mass of the load for loaded walking was 38kg.
- Data type
- Marker position data
- Ground reaction force data
- Subject
- mass: 61.3kg
- Sex: male
Modeling
To simulate an exosuit wearer, it is most important to create a model which can replicate a real subject as possible as we can. In this project, as the experimental data was acquired from a subject walking freely, it is not possible to make a realistic exosuit wearer. However, to make my model dynamically consistent to the experimental data, my model was gone through the basic steps of modeling procedure in Opensim, and then added actuators and metabolic cost probes.
How to model a subject wearing active actuator
The diagram describes what procedure my model had gone through.
- The first three steps are basic modeling procedure in Opensim to make a model dynamically and kinematically consistent to the experimental data. For more information, refer to
- After RRA is done to the model, probes for calculating metabolic cost were added to the model. . For more information about how to add the probes, refer to
Simulation-Based Design to Reduce Metabolic Cost
- And then, I added active actuators to the model. Here, I used PathActuator class to simulate active actuators of exosuit, as they are cable driven actuator. If you are interested in how PathActuator works, refer to
Sample models
Here are the figures of sample simulation models. I created several different types of models for both loaded gait and unloaded gait.
Loaded gait model
Three different types of models were created for loaded gait simulation
- A model without actuator
- A model with path actuators supporting plantar flexion
- A model with path actuators supporting hip extension
The path actuator supporting plantar flexion is attached to heel and tibia, and the pathactuator supporting hip extension is attached to backpack and femur. Loaded mass was added to torso for simplicity.
Unloaded gait model
Same types of models were created for unloaded gait simulation. The main difference between loaded gait model and unloaded gait model is the mass of torso, and transparancy of backpack.
Optimization methodology
The idea to optimize the control input force for the actuator is to take advantage of CMC tool. The main reason we use CMC in OpenSim is to find a most suitable excitations for muscles to create body movement while minimizing activation. To see how it works, refer to
In this project, I make use of different use of the objective function in CMC in order to find optimal input force.
- CMC procedure contains static optimization process, and it tries to minimize the cost function J which can be represented as
- When we add active actuators on OpenSim Model, the activation term in cost function becomes
- Where X_muscle is muscle control and X_actuator is actuator control. X_actuator is part of activation state, and it is also adjusted after the optimization process.
- Now, if we diminish the influnece of X_actuator on J, and run CMC, the optimizer tries to find X_actuator in order to minimize muscle activation.
- We know that minimizing muscle activation correponds to minimizing metabolic cost, so we can we can say that the actuator input force resulted from CMC after diminishing the influence of X_actuator is the optimal actuator input for most efficient metabolic reduction.
- Muscle force is constructed by the equation
- And if we assign large value of maximum force to each actuator, then actuator control x_actuator decreases, so that the influence of actuator to J is decreases.
- Using this methodology, I could find an optimal input for each actuator, and also see the metabolic cost reduction after active actuators are added to a model.
Result & Discussion
Metabolic cost change when active actuators are added to model
First, I investigated how much metabolic cost is reduced when active actuators are added to a model, and optimal input force is applied to actuators. I did simulation for both loaded and unloaded walking cases, and I compared the influence of hip actuator and ankle actuator to metabolic cost reduction.
I assigned 10,000N to maximum active actuator force in order to find optimal control input force through CMC.
| Loaded walking | Unloaded walking |
|---|---|
 |  |
- Metabolic cost reduction when active actuators are added to loaded gait model
- Ankle actuator: 10.35%
- Hip actuator: 6.62%
- Metabolic cost reduction when active actuators are added to unloaded gait model
- Ankle actuator: 10.62%
- Hip actuator: 1.04%
- Things to notice
- The metabolic cost is much lower during unloaded walking than loaded walking. Loaded walking costs only 75% metabolic energy compared to loaded walking.
- Ankle actuator works better to reduce metabolic cost than hip actuator when we can apply optimal input force.
- Hip actuator is not assistive to unloaded gait.
Therefore, we can say that ankle actuator helps metabolic cost reduction better than hip actuator if we have an optimal actuator which has no maximum force limitation.
Optimal actuator input
| Loaded walking | Unloaded walking | |
|---|---|---|
| Ankle actuator |  |  |
| Hip actuator |  |  |
In hip actuator case, the optimal input force is very complex, and I could not find any intuition from it. My initial guess on the optimal input force of hip actuator was to follow the hip flexion angle change, but the result doesn’t follow it at all. The complexity may be due to the optimization procedure in CMC, and I will try to figure out the reason in a future. However, the good thing about hip actuator is that it doesn’t require large amount of force to reduce the metabolic cost. The maximum force in this optimal input is about 400N, which is significantly lower than ankle actuator input, and it is achievable.
Unloaded walking
This is the optimal input force result in unloaded case. Still, you can see that ankle actuator input profile is clean and goes well with our intuition, but hip actuator input profile isn’t.
Analysis of optimal input force for ankle actuator
This result shows how muscle force changes when a model has ankle actuators. The graphs show plantar flexor muscle forces, and first row is muscle forces of a baseline model, and the second row is the muscle forces of a model with ankle actuator.
The red line is The muscle forces of gastrocnemius, and it barely change when ankle actuators are added. However, other muscle forces, which are from uniarticular muscles, are significantly decreased. Therefore, we can say that ankle actuator assists uniarticular muscles during loaded walking
If we draw the sum of baseline uniarticular forces and active actuator input force together, we can see that the active actuator force follows baseline uniarticular muscle forces. The redline here is sum of baseline uniarticular forces and blue line is active actuator input force. This force signal is clear and easy to implement real world. However, the maximum actuation force is about 2500 N, which is too high, so we need to deal with it if we want to use this profile.
Best realistic actuation input force for ankle actuator
As the optimal input force for ankle actuator is not achievable, I tried to find a realistic ankle actuator force which has its maximum force of 400N. My initial guess was to saturate the optimal input force that I found earlier at 400N. Therefore, I cut the optimal input force at 400N, and create new input force. I added this force profile to CMC tool as a control constraints, and run CMC again.
I compared the saturated optimal input to a new CMC results which was acquired with 4000N maximum actuation force and bounded control input between 0 and .1. The new CMC results also has maximum force of 400N as the control input is bounded, and it gives better input force in terms of metabolic cost reduction than a result of CMC which was acquired with an actuator with 400N actuation and conventional control input.
In these graphs, you can see a similaritly between the saturated optimal input and a results of new CMC procedure.
| Loaded walking | Unloaded walking |
|---|---|
 |  |
Discussion
ModelBiarticular actuator
Now that we know both ankle actuator hip actuator works well to reduce metabolic cost during loaded walking, the natural progress is to create biarticular actuator which can affect both ankle plantar flexion and hip extension. In order to reduce the number of actuator, I created biarticular actuator with 1 DOF, and see how much it reduces metabolic cost, and what it’s optimal input force is.
Simulation result
Loaded walking- rate of metabolic reduction
Control input is noisy, which makes it hard to realize
Biarticular actuator is not as effective as uni-articular actuators in terms of metabolic cost reduction.
Conclusion
Featured result
Limitations
Source code
You can find the model that I used in htt~~~~