P15001: Soft Ankle-Foot Orthotic

Systems Design

Table of Contents

The specific goal of our team is to incorporate previous work done, including using a McKibben muscle and a terrain sensing system into an untethered AFO. It also should have an aesthetically pleasing flexible exoskeleton made from allergy conscious materials which comfortably fits into a user’s existing footwear. The exoskeleton need to be integrated with the actuation device, sensing system, and microcontroller. The AFO must also must be capable of applying torque to and rotating the user’s foot and should be designed to endure an entire day use untethered. The sensors and microcontroller system should incorporate the existing terrain sensing system as well as implementing more suitable heel strike sensing. The resulting design and prototype must follow the safety standards set forth by the Institutional Review Board as well as the ASME Boiler and Pressure Vessel Code.

Functional Decomposition

Return To Top

Engineering Requirements Mapping

Return To Top

Morphological Chart

Return To Top

Selection Criteria

Return To Top

Pugh Chart

Return To Top

System Design

Chosen Design

Return To Top


System Concept #2 System Concept #3 System Concept #4

Return To Top

Research and Benchmarking

Research Articles

The following is a paper about a Powered Ankle-Foot Prosthesis. It was published in the IEEE Robotics & Automation magazine in 2008. This paper was of benchmarking interest to our group because it contains quantitative information about human ankle kinematics. The kinematic values presented in this paper were used in our "Force to Lift Foot" feasibility calculations.

Powered Ankle-Foot Prosthesis

Feasibility Testing

Return To Top

Compressed Air (CAIR) Pure Calculations

Using past projects' research (See P12029: ROBO Ant), untethered compressed air feasibility calculations were performed. These calculations were based on a single air muscle in order to ascertain the number of times it could be flexed given a standard paintball CAIR reservoir (V=48in3, 3000psi). One "step" corresponds to a single flex of the air muscle. The optimization plots below demonstrate the relationship between pressure and muscle volume to tank life. According to a NY Times report the average American takes less than 6,000 steps per day; this corresponds to 3,000 steps for each leg. This value was set as the goal of the feasibility calculations. In the figure below, the two optimization parameters were combined and the results show concept feasibility. Moving forward, there are also higher pressure storage tanks available such as the Guerrilla Air Carbon Fiber Tri Label Compressed Air Tank

These carbon fiber tanks are more expensive and also pose regulation concerns due to higher operating pressure. The befits of using a V=48in3, 4500psi storage tank can be seen in the image below.

Both storage vessels demonstrate untethered feasibility for an entire day.

Additional details, including assumptions can be found in the Complete CAIR Pure Calculations Report

Return To Top

Heelstrike Force and Sensing

In order to understand what type of sensor that needs to be used, the force of the foot while walking needs to be understood. During normal locomotion, the pressure of the foot is different in different locations. The University of Hong Kong did a study about a pressure on the foot. The following two figures display this pressure: Peak pressure wire frame diagram for a subject during walking Peak plantar pressures (kPa) from 111 adults and 125 children during walking
From this information, it is clear that the placement of the sensor would have the least force if it is closer to the center of the foot. This benefit needs to be weighted with when that force will be applied (the sensing needs to happen when the heel strike occurs). Complete Heelstrike Force and Sensing Report

Return To Top

Distance Sensor Speed

In order to insure that the type of sensor that is currently used in the gait monitoring system is operating fast enough, some worst case calculation should be completed. These calculations are assuming that an ATmega328 microcontroller is used as well as a GP2Y0A0YK IR long distance measuring sensor. Also assumed is that an analog to digital conversion is being used (this would of course be slower than a digital read or an interrupt. These methods may be used in the future but they would be faster, making an ADC read the worst case). Information from distance sensor (The following timing chart was taken from the GP2Y0A0YK datasheet):

From this information, worst case to first output would be 53ms. This is assuming that operation has stabilized (meaning that power has been applied at least 100ms previously).
From the microcontroller datasheet, ADC characteristics section, conversion time at maximum would be 260 us. For the post processing data, the assembly code FMULS (fractional multiply signed) takes two clock cycles and a compare takes two. Assuming that we are working in C and not assembly, the overhead of the code needs to be taken into account. For direct porting verses a digital write there are 50 cycles to 2. Therefore it is possible to assume that post processing would take 100 cycles. For worst case that number can be doubled. Therefore 200 cycles or 2500 ns for a 16MHz clock.
With all this information. It is reasonable to say that reaction to an object should be well under 100ms or 1/10 of a second. More time may have to be added in the logging is added to the system, but 100ms is a reasonable time frame for pure measurement.

Complete Distance Sensor Speed Report

Return To Top

Power Sources

Power Source Selection Proposal

Return To Top

Force to Lift Foot

This is a table of the data collected during the force feasibility tests. Force Feasibility Testing Proposal
Lift Foot Testing Video
Feasibility Testing Foot Lift Setup 2
Feasibility Testing Foot Lift Setup 3

Return To Top

Budget Feasibility

This is a very high level budget proposal. It is important to note that only $400 is used out of the $500. This is left for unforeseen circumstances.
Return To Top

Weight Feasibility

This is a very high level weight feasibility test. Many of the values are conservative, approximate weights, and some of the weights were taken from previous parts used on other projects.
Return To Top

Follow-up Interviews

Interview Date Interview Subject Interview Questions Interview Notes
September 24, 2014 Dr. Elizabeth DeBartolo September 24 Questions September 24 Meeting Notes
September 29, 2014 Nazareth Clinic Clients N/A Client Interview
Client Observations

Return To Top


Regulator Resources

Past Projects

Return To Top

Performance Limits

Return To Top

Risk Assessment

Return To Top

System Design Review

The following presentation was given on October 2: Presentation
The feedback that was gathered from this review lead to the following action items.

Return To Top

Home Planning & Execution Problem Definition Systems Design Subsystems Design Detailed Design Gate Review
Build Preparation Build & Test Integrate & Assemble System Validation System Verification Final Review