Project Design and Management

Old Dominion University

Project Design and Management

MAE 434W

Professor S. Ringleb

Summer 2016

Instrumented Walker

Jairus Potts

Skyler Bullington

Chris Barnett

Larson Stacy

Rockwell Shields

Tommy Frankenberger

Table of Contents

Abstract…………………………………………………………………………………………..

Introduction……………………………………………………………………………...... 1

Methods………………………………………………………………………………………….2

Completed…………………………………………………………………………………..……2

Purposed…………………………………………………………………………………….…...3

Preliminary Results……..………………………………………………………….…...... 4

Discussion………………………………………………………………………………………..5

References ……………………………………………………………………………...... 9

List of Figures

Figure 1: Initial Calibration Testing...... 5

Figure 2: Load Cell …………………………………………………………………………….6

Figure 3: Load Cell Placement………………………………………………………………...9

Figure 4: Initial Load Cell Testing…………………………………………………………….10

Appendices

I. Gantt Chart ……………………………………………………………….….…..……...... 12

II. Budget...……………………………………………….….………….……………………..13

Abstract

Walkers are widely used for their lower extremity support and ease of use, making them ideal for analyzing the forces being placed on them. Although walkers can assist in recovery, they can also be used incorrectly in a way that lets the patient rehabilitate without recovering muscle mass and mobility while putting excessive strain on the upper body [1]. Collection of data from load cells positioned on the walker will allow for clinicians to assess how often and how much the walker is relied. In order to collect this data, single axis load cells are going to be attached between the 4 legs and feet pads of the walker. The location for these cells was determined by the necessity of an attached forearm rest for patients with insufficient upper limb strength to hold themselves up, therefore accounting for any hand placement. These sensors will route to a data distribution system on the walker, and be transmitted to a remote desktop. This data will sync to Motion Monitor in order for a complete analysis of the forces being applied to the walker.

Introduction

Spinal cord injuries have the potential to impair, restrict, or eliminate movement capabilities of an individual [2]. In order to combat these side effects of such injuries, clinicians help patients with treatment techniques such as surgery and physical therapy. Many of these treatments involve rehabilitation tools, such as walkers, exoskeletons, and crutches [2]. However, it is sometimes unclear as to the effectiveness of these treatments.

An instrumented walker can provide a tool to assess loads that are applied by a patient who suffers from reduced mobility. Longitudinal assessment of the mobility of patients with spinal cord injury will provide clinicians with continuous measures of a person’s functional ability and activity levels, and can help evaluate a person’s health over a long period of time [3]. Despite thousands of patients with injuries or diseases requiring the use of a walker, less than a dozen walkers have been developed to record how much weight the patients are putting on them [4]. Determining the amount of weight that a patient applies to the walker is useful to a treating physician as it allows them to measure recovery progress and evaluate the effectiveness of recovery techniques, and the same can be said for crutches.

In the past, clinical researchers have employed instrumented wired walkers to help assess upper limb weight bearing. These walkers have been able to reliably report applied forces, but have had stringent requirements on hand placements and have been encumbered by wired designs, limiting range and flexibility of operation [5]. In one study that investigated methods for walker instrumentation, six degrees of freedom load cells were used to replace the handles of a walker which allowed for recording of data but restricted hand placement when the walker was operated, limiting the range of patients who could operate the walker [6]. In another previous instrumentation, pressure sensors were placed in each handle of a walker and connected to a wireless transmitter that transmitted sensor data to a computer to calculate weight distributed through each side of the walker [5].

These limitations on hand placement and range of operation due to extraneous wiring directly limit the effectiveness of these rehabilitation techniques. The purpose of this project is to develop, build, and test a instrumented walker that is capable of reporting weight regardless of hand placement, and to develop a design for a set of crutches for the same purpose.

Figure 1. Initial Calibration

Methods

Completed Methods:

Figure 2. Selected load cell

Preliminary research was performed on type, accuracy, limitations, and installation requirements of sensors used in multi-dimensional force analysis. The decision was made to move forward with single axis load cells, providing a removable sensor solution outlined in the design requirements. A medical grade walker composed of a lightweight aluminum alloy, and steel crossbeam by Medline (Medline Industries, Mundelein, IL) and a set of Lofstrand crutches were provided for instrumentation by the Mayo Clinic (Rochester, MN).

A presentation of preliminary design concepts were presented to the Mayo Clinic for review and to determine budgetary concerns. The cost analysis in the presentation was performed on current consumer based microcontrollers. The analysis concluded with viable options for configurations, analog channels and powerrequirements for multiple sensors. Testing was conducted to determine methods for amplifying sensor signals, which highlighted that a custom designed circuit is required for improved signal accuracy.

Calibration algorithms for use with microcontrollers were developed in the C programming language, then interfaced and tested with the Arduino Uno (Arduino LLC, USA) and a serial console application for command line interface. The walker undergoing simple load cell testing can be seen in Figure 1. The algorithms were developed for various wiring configurations and amplifications. The accuracy, calibration curve and sensitivity of the load cells can be tested on all future components with only minor changes. Given the design requirements, the desired method for reading load cell signals was determined to be achieved through individual excitation and signal outputs.

To calibrate the load cells a simplified approach was taken. Research was conducted and an Integrated Circuit was identified for use with load cell applications. The ADS1232(Texas Instruments, Dallas Texas) is a precision Analog to Digital Converter (ADC) IC with 24 bit resolution and an onboard programmable gain amplifier. The ADS1232REF reference board in combination with corresponding evaluation software allowed for a faster calibration and an accurate bench testing process. These components provide greater accuracy of the load cell signals when excited to +5V, which is half of the benchmark excitation, providing a lower voltage for a battery power supply. The reference board also provides the option to use external power supplies to collect and quantify data due to noise from the power supply to eliminate erroneous levels from load cell signals.

Custom power requirements were needed to provide the smaller excitation to the load cells and power the microcontroller from the same battery. The LMR23825 step down converter was chosen for its low power consumption and power down features. Utilizing a 12 Volt battery, the converter can apply +5 Volts and micro Amps efficiently to the four load cells individually without sacrificing battery life or adding excess noise. The custom circuits have been redesigned to eliminate excess noise with Surface Mounted Devices, and have been ordered from a manufacturer.

Custom attachments for load cells with press fit tolerances, that will insert in the bottom of the legs of the walker have been designed and machined. These attachments will allow the load cell to screw into the leg eliminating the initial friction sleeve design. The 3D model of the walker in SolidWorks has been completed. All walker components have been modeled for a complete drawing assembly.

Proposed Methods:

Four individual single axis load cells with a maximum weight bearing load of 100 lbf, overload factor of 2.5, and +- 1 lbf accuracy will be placed between the 4 legs and feet of the walker (Figure 2). The load cells will be screwed into the leg of the walker directly to allow the sensor to achieve the full extent of its ability yet remain compact and uniform on the walker (Figure 3). Wiring from the load cells and circuit board will be routed to be unobtrusive.

The data acquisition system will be comprised of two micro controllers. One integrated on the instrumented walker and connected to the load cells through the external ADC. The digital signal is transferred to the receiving unit through a wired umbilical until further testing can be conducted through wireless transmission. The receiving unit translates the digital signal over serial USB interface to the remote PC for software integration. The sensor signals will be sampled at 360Hz on the walker and transmitted between the two data acquisition cards [4]. Initial experimentation will begin with integration of LabView (National Instruments, Austin, Texas) software due to its Arduino integration, in order to obtain the best method for serial communication in desired software.

A custom program in The Motion Monitor (Innovative Sports Training, Chicago IL) will receive the data for real time comparison of total data collection. Algorithms to adjust sensor signals, serial communication scripts, or modifying equations in Motion Monitor software will be written to fulfill the needs of the Mayo Clinic. Testing will be performed to ensure the software is able to report weight through a combination of readings from the legs by placing known weight at various locations on the handles of the walker. Dynamic testing will also be conducted with a loaded walker in moving scenarios. It will be important to verify that load placement has no effect on the load output readings and to acknowledge any further calibration.

Figure 3: Load Cell Placement

Preliminary Results:

Although data collection has yet to begin the group has a final design in which hand placements and weight limitations were both taken into consideration. The established design requirements depict that hand placement restrictions have to be taken into consideration along with the weight restrictions. The single axis sensor selection provides the only solution to design criteria. Placing load cells at the feet of the walker will allow for the load being applied to the z-axis to be analyzed. With the cell threaded into place, this will eliminate the possibility of discrepancies with friction sleeve mounts initially conceptualized.

Initial load cell testing with the ADS1232REF evaluation board has shown precise measurements will be achievable from the +5 Volt excitation range (Figure 4). The platform determined that an individual power circuit was needed to achieve the best resolution for amplification of various load cell power requirements. Simple amplification circuits skewed signal data due to errors in the amplification configuration and IC efficiencies not meeting criteria. This showed that multiple load cells, in a full or half bridge configuration, will need to be individually amplified to achieve the manufactured precision of the load cells. To overcome these amplification errors, a variable gain level was identified as a requirement for the power regulation. Further testing is still needed to address power consumption and power supply requirements with the full system.

Figure 4. Load Cell Initial Testing 0-60 lbs

Discussion

The purpose of this project was to design a system that can be used to instrument a pre-existing walker that is capable of recording weight transmitted through the walker. A design for a single axis measurement system was proposed and designed. The design consideration for implementing a single axis system resides from research and consulting an experienced electronics technician.

Although the current design is limited in the aspect that it is a measurement device designed for a single axis, improvements were made in the form of removability of instrumented circuit board and load cells, and flexibility in the hand placements. Additionally with a larger budget more precise sensors can be implemented to improve readings. The next step is to move forward with fabrication of a prototype. The initial configuration will be wired for system optimization. Upon successful completion and testing of the single axis system, the system will then be modified and designed to crutches, with a preliminary design relying on wireless configuration.

Appendices

Gantt Chart

References

[1] Matteo Lancini et al, “Instrumented Crutches to Measure the Internal Forces Acting on Upper Limbs in Powered Exoskeleton Users,” Advances in Sensors and Interfaces, 6th International Workshop, Gallipoli, Italy, 2015 IEEE.

[2] T. Wang, Jean-Pierre, and Merlet, "Walking analysis of young-elderly people by using an intelligent walker ANG," pp. 1-10, 2004.

[3] M. Alwan, G. Wasson, P. Sheth, A. Ledoux, and C. Huang, "Basic walker-assisted gait characteristics derived from forces and moments exerted on the walker’s handles: Results on normal subjects," vol. 1, pp. 1-4, 2007.

[4] A. Fast, F. S. Wang, R. S. Adrezin, M. A. Cordaro, J. Ramis, and J. Sosner, "The instrumented walker: usage patterns and forces," Arch Phys Med Rehabil, vol. 76, pp. 484-91, May 1995

[5] R. Gerena, P. VanDeventer, A. Vistamehr, C. Conroy, P. Freeborn, H. Govin, et al., "Wireless instrumented walker for remote rehabilitation monitoring," in 2015 IEEE Virtual Conference on Applications of Commercial Sensors (VCACS), pp. 1-7.

[6] M. Alwan, G. Wasson, P. Sheth, A. Ledoux, and C. Huang, "Passive derivation of basic walker-assisted gait characteristics from measured forces and moments," in Engineering in Medicine and Biology Society, 2004. IEMBS '04. 26th Annual International Conference of the IEEE, pp. 2691-2694.

[7] Finley, M. and Oliveira, A. S. (2011). “Upper extremity joint stresses during

walker-assisted ambulation in post-surgical patients.” Brazilian Journal Of Physical Therapy / Revista Brasileira De Fisioterapia, 15(4), 332-337.

[8]Ming, D. et al. (2009). “Measurement of upper extremity joint moments in walker- assisted gait.” IET Science, Measurement & Technology, 3(5), 343-353.

[9]R. A. Bachschmidt, G. F. Harris, and G. G. Simoneau, "Walker-assisted gait in rehabilitation: a study of biomechanics and instrumentation," Neural Systems and Rehabilitation Engineering, IEEE Transactions on, vol. 9, pp. 96-105, 2001.