Canine Exoskeleton. Mid-project Report. Fall Semester Full Report-

Transcription

1 Canine Exoskeleton Mid-project Report Fall Semester Full Report- By Jordan Bernhardt Lindy Gillette Colleen Jones Kat Killingsworth Ryan Leuenberger Department of Electrical and Computer Engineering Colorado State University Fort Collins, Colorado Project advisor: Dr. Anura Jayasumana (revisions made from Dr. Jayasumana s comments but this version has not been approved yet) Approved by:

2 ABSTRACT As many large breeds of dogs reach old age, the chances of developing hind limb disabilities due to a variety of diseases and other causes increase. Degenerative myelopathy (DM) is a common example of a disease that affects many breeds of dogs as they get older and can impair their hind limb function. Degenerative myelopathy is the degeneration of the spinal cord which can cause paralysis over time. Symptoms in dogs with DM start with hind limb weakness and incoordination, as well as muscles starting to atrophy. Over time, paralysis starts to set in and can completely paralyze the hind limbs. Because DM is a degenerative disease, there are no long-term treatments for dogs diagnosed with DM. Treatment to manage pain and symptoms is common, but there is no long term cure. Exercise in the beginning stages of DM can help delay muscle atrophy, as well as maintaining a healthy diet. Excessive weight gain places more stress on the hind limbs, causing more pain and less mobility. The current braces on the market for canines affected with DM require the canines owner to assist the dog the entire time that the dog is walking (ie hold the brace to lift the dog s hips). This is cumbersome, heavy and may not be reasonable for many pet owners, especially those with large dogs. The Canine Exoskeleton Mobility System aims to provide an entirely mechanical system that will provide the dog with more autonomous and independent limb movement. This increases the quality of life for these dogs and their owners. The Canine Exoskeleton is a device that uses electromechanical systems (i.e. motors and control circuitry) acting on a brace that will provide support to the rear legs of a canine. This year s team will focus on developing a second generation canine exoskeleton. Some differences between the first and second generation will be increased range of motion, decreased input to output delay, and additional comfort to the dog. In order for a functional device to improve a canine s life, the device will need to provide a full range of motion, such as sitting and walking in a nonlinear fashion.the first generation of the prototype was too slow to realistically mimic the movement of a dog, so improvements in speed are a top priority for the design team. The project team has come up with a plan to make these improvements and hopes to work with both healthy and disabled dogs this year to continue to reach a final second generation model. Since the end of August 2016, we have accomplished getting a revised protocol approved for live animal testing, building a proof of concept for improving the mechanical speed of the exoskeleton, and accruing funding from various companies and private individuals. The improvements over the first generation have shown promise, though more data needs to be collected to define specifications and required. Over the next few months, we hope to write the appropriate sensing code and build a custom brace that will be able to be used on a partially paralyzed dog. 1

3 Table of Contents ABSTRACT 1 Table of Contents 2 List of Tables and Figures 5 Chapter 1: Introduction 7 Chapter 2: Summary of Previous Work Items on the Market 10 Figure 1: Walkin Wheels Dog Wheelchair [2016] 11 Figure 2: Help Em Up Dog Harness [2016] Human Exoskeletons 12 Figure 3: Ekso GT Rehabilitation Exoskeleton [Ekso 2015] Generation 1 Exoskeleton 13 Chapter 3: Ethics 15 Chapter 4: Project Plan Summary Customer Requirements and Specifications 16 Table 1: Customer Requirement and Engineering Specifications Constraints 17 Figure 4: Goals and Constraints for this year s team 18 Figure 5: Risk analysis for design improvements 18 Chapter 5: Hardware Previous Design 19 Figure 6: Generation 1 prototype 19 Figure 7: Generation 2 prototype Linkage System 20 2

4 Figure 8: Linkage Path of Motion using Linkages 20 Figure 9: Hip Angle 20 Figure 10: The New Linkage System Throughout the Gait Cycle Pulley System 21 Figure 11: Pulley system 22 Figure 12: Piston type gas spring [7] Dog Pack and Electrical systems 23 Figure 13: Ruffwear Approach Dog Pack 24 Figure 14: Schematic design for the exoskeleton (excluding sensing system) 24 Chapter 6: Software Systems Sensing system 26 Chapter 7: Logistics VTH Collaboration IACUC RICRO Orthopets Collaboration 28 Figure 15: Dog wearing an Orthopets Brace [Orthopets 2015] 29 Chapter 8: Fundraising 30 Chapter 9: Testing 31 Chapter 10: Conclusion and Future Work Uncertainties Springs Coding Speed Dogs Reaction Plans for Second Semester Finding Dogs with Partially Paralyzed Limbs 33 3

5 Customized Canine Exoskeletons Final System for the Second Semester Plans for Future Years Final Thoughts 33 References 35 Appendix A - Abbreviations 36 Appendix B - Budget 36 Figure 16: Budget 36 Table 2: List of costs incurred and deposits made into the team s account. 37 Appendix C - Project Plan Evolution 38 Table 3: Timeline for the holiday break 38 Table 4: Timeline for the second semester till E-Days 39 Project Timeline 10/28/16 40 Table 5: Timeline for the second semester till E-Days. 40 Figure 17: Original Gantt Chart for Appendix D - Project Code Files 42 Appendix E - IACUC and RICRO Certifications 46 Appendix F - Donation Letter 48 Appendix G - Device Testing Verification C 51 4

6 List of Tables and Figures Figure 1: Walkin Wheels Dog Wheelchair [2016] 11 Figure 2: Help Em Up Dog Harness [2016] 12 Figure 3: Ekso GT Rehabilitation Exoskeleton [Ekso 2015] 13 Table 1: Customer Requirement and Engineering Specifications 17 Figure 4: Goals and Constraints for this year s team 18 Figure 5: Risk analysis for design improvements 18 Figure 6: Generation 1 prototype 19 Figure 7: Generation 2 prototype 19 Figure 8: Linkage Path of Motion using Linkages 20 Figure 9: Hip Angle 20 Figure 10: The New Linkage System Throughout the Gait Cycle 21 Figure 11: Pulley system 22 Figure 12: Piston type gas spring [7] 23 Figure 13: Ruffwear Approach Dog Pack 24 Figure 14: Schematic design for the exoskeleton (excluding sensing system) 24 Figure 15: Dog wearing an Orthopets Brace [Orthopets 2015] 29 Figure 16: Budget 36 Table 2: List of costs incurred and deposits made into the team s account. 37 Table 3: Timeline for the holiday break 38 Table 4: Timeline for the second semester till E-Days 39 Table 5: Timeline for the second semester till E-Days. 40 Figure 17: Original Gantt Chart for

7 6

8 Chapter 1: Introduction The development of canine and human exoskeletons has increased over the last decade. The goal of the canine exoskeleton developed in part with Orthopets aims to aid in the rehabilitation of canines with partially paralyzed hind limbs. Current harnesses and slings require the owner to assist and lift up the dog in order to raise the hind limbs off the ground. The goal with this team s canine exoskeleton design is for the dog to be able to initiate movement on its own and walk with assistance from the motorized brace. Developing canine exoskeletons and testing the device on a partially paralyzed dog as a proxy prior to conducting similar research on humans can lead to significant advances in this industry. The demand for human exoskeletons, whether it be for hands or legs, has never been higher with cases of muscle weakness, stroke effects, paralysis, muscle diseases, etc. Bos et al. 2016) In order for patients to rehabilitate or perform daily tasks, current methods limit their abilities and progress. One company, Ekso Bionics, has developed the first and only FDA cleared exoskeleton for use with stroke and spinal cord injuries from L5 to C7. (Ekso 2016) Thousands of people can be positively impacted by this new exoskeleton. According to Neil Bowler from BBC News, Robotic or mechanical exoskeletons could offer humans the kind of protection, support and strength they afford in nature. (Bowler 2014) Through the development of the canine exoskeleton, more research can be conducted in the future than can lead to different types of exoskeletons. Currently, there are various harnesses and slings that offer dogs assisted mobility through the use of the owner lifting the dog s hind limbs off the ground. Other products on the market include dog wheelchairs with a sling, four wheel supports, and wheelchairs without a sling. The difference with the canine exoskeleton and slings/wheelchairs is the need for human assistance and limited mobility. Wheelchairs are not able to take sharp turns or go through narrow spaces. As well as limited mobility, the wheelchairs do not allow for the dog to sit or lay down while connected to the device. Lifting the hind limbs off the ground with the sling places unwanted pressure on the abdominal cavity of a canine, posing more risk of injury to the dog. The canine exoskeleton aims to provide more mobility with turning, allowing the canine to sit and lay down with assistance, and be able to walk on its own. Not only will the exoskeleton help with rehabilitation, but would provide at-home relief. Not only can human and canine exoskeletons offer rehabilitation and therapeutic options, but can help with emergency rescue situations and supply the need for increased power output. The Hulc, a human exoskeleton designed by Lockheed Martin, aims to provide more power and weight support for soldiers in the field. (Bowler 2014) Another lower limb exoskeleton from Indego claims to enable[ing] people with spinal cord injuries to walk and participate in over-ground gait training. (Indego 2016) The Indego device allows the patient to walk assisted with canes and have no bulky battery pack attached to the back of the patient. It is lightweight and no wires 7

9 are interfering with the appearance and functionality of the device. To improve the canine exoskeleton, a smaller pack that houses the motors and wires needs to be developed. Another option to consider is having a way to dial back the amount of power outputted to the exoskeleton for rehabilitation purposes so the patient can learn to walk on his/her own. The team s goal is to develop a second generation design of the canine exoskeleton. The team will use the prototype that was developed by last year, which is the first generation prototype as their starting point, and make adjustments where they see fit to create a new and improved design. Some areas of improvement from the second generation prototype are the speed, the integration of the sensing and motor devices, and the weight of all the components. At the end of the first semester, the team has created a working prototype of the pulley and spring system. The wireless gyroscopes are in the process of being coded to transmit walking data to the arduinos. To successfully develop a canine exoskeleton, it is important to keep in consideration the federal laws and regulations put in place. The Occupational Safety and Health Administration (OSHA), aims to ensure the safety of workplaces. It is important for the senior design team to conduct research in a safe environment so they do not hurt themselves or the canine subjects. By following the guidelines of CSU s Research Integrity and Compliance Review Office (RICRO), the protection of animal subjects is held to the highest of standards. Another standard the senior design team must be held to is the Institutional Animal Care and Use Committee (IACUC). IACUC is responsible for approving research protocols with the use of animals to maintain the ethical treatment of animal subjects. When testing the canine exoskeleton on healthy and disabled canines, it is important to make sure that the safety of the animal is of the utmost concern. The remaining of the report outlines the work that has been achieved this semester and what is to follow in the coming semester. Chapter 2 will review the first generation of the project as well as the the market and products on the market right now. Ethics is a huge part of the project, as the team will be dealing with live animals. Chapter 3 will address the ethics involved and the protocols in place for ethical checks. Chapter 4 will outline the goals the team has for the second generation exoskeleton. This will include a list of constraints from the customer, and the constraints the group has to work within. Detailed design review for each new system will be outlined in Chapter 5. This will include all of the new designs that will be implemented in the second generation. The first generation will be outlined followed by the new linkage system that has been implemented, the spring and pulley design change, and finally the dog pack that will include the electronics and drivers of the system. The software systems of the device will be explained in Chapter 6. Chapter 7 will outline the logistics that the team has to go through, the certifications that have been obtained and the negotiation that has been made with Orthopets to receive help with the manufacture of braces. Fundraising has taken up a lot of effort this first semester to build a budget to make the required changes, Chapter 8 will explain what the team 8

10 has done to raise funds. Chapter 9 will explain the testing protocol that is in place for the new generation. The final Chapter, Chapter 10, will conclude the report with uncertainties that remain, the plans in place for the second semester, as well as future work. The report will be concluded with the group's final thoughts on the project after a semester of work. 9

11 Chapter 2: Summary of Previous Work There are devices on the market currently that aim to fix the problem of hind leg paralysis. These current devices will be explained in this chapter. This chapter will also explain how some aspects of the human exoskeletons that are being created could be applied to the canine exoskeleton. Finally this chapter will review the work done to complete the first generation of the project This multiyear project was started last year by a different senior design team. They set a solid foundation for this project to be built off of by creating a proof of concept prototype. The prototype consisted of a motorized brace system moving a cart to simulate a dog walking and a sensing system to gather data about canine gait. The two systems worked independently of each other, but were planned to be integrated together. The sensing system would provide real time gait data to the motorized system and would move according to that gait. Last year s report as well as their website are available below in the references section. 2.1 Items on the Market The current solutions to hindlimb disabled canines are limited in the options available as well as their functionality. The two most common solutions are dog wheelchairs and dog harnesses. Both of these device work to help the canine achieve more mobility, however they both lack in certain important areas as well. One of the most common current treatment methods for hind limb deficiency in canines is dog wheelchairs. An example for these wheelchairs is the Walkin Wheels Dog Wheelchair seen in Figure 1. These wheelchairs are most commonly used for the rear limbs, but are available for the front limbs as well. Wheelchairs accomplish their goal of increasing the mobility of canines with limited capabilities in their legs, however they do have their limitations as well. Wheelchairs don t promote proper gait of the dog, since the dog may simply just get accustomed to the wheelchair supporting the rear portion of its body while it only uses its front legs to pull itself forward. This is problematic because this change in use of the dog s legs compared to normal may induce more stress on the dog s functional legs than normal which can lead to an increased chance of injury as well as muscular dystrophy and bone loss in the rear legs from disuse. Wheelchairs also don t offer much mobility vertically. If the dog wants to climb stairs for instance, it would have to drag its body up using its front legs, and in this case the wheelchair would only be hindering the dog by adding additional weight. 10

12 Figure 1: Walkin Wheels Dog Wheelchair [2016] The other solution to canine hindlimb disabilities on the market right now is dog harnesses such as the Help Em Up dog harness as seen in figure 2. These harnesses allow the dog to put a variable amount of weight on their hindlimbs, eliminating the problem of muscular dystrophy and bone loss that the wheelchairs have. The main issue with the harnesses is that they require help from the owner at all times since the owner has to hold up the dog with the harness. This is impractical and may also tire out the owner from having to assist the dog at all times, especially since large breeds of dogs are more prone to hindlimb disabilities. 11

13 Figure 2: Help Em Up Dog Harness [2016] 2.2 Human Exoskeletons Even though no exoskeleton suit exists for dogs at the moment, various companies such as Ekso Bionics have created an exoskeleton suit for humans as seen in figure 4. Ekso s device uses actuators at the hip and knee joint to assist in the movement of the patient s legs similar to the canine exoskeleton device, however the controls for their device are manual and not autonomic like the canine exoskeleton device hopes to be. Despite the difference in intended patient, certain design elements from the Ekso device may prove useful for this project. Additionally, lessons learned from creating this canine exoskeleton could benefit human exoskeletons as well. 12

14 2.3 Generation 1 Exoskeleton Figure 3: Ekso GT Rehabilitation Exoskeleton [Ekso 2015] The motorized system consists of 3 motors for each leg actuating each major joint in the canine hindlimb: the hip, stifle (or knee), and hock (or ankle) joints. In order to move the hip joint, last year s team built a linkage system actuated by a stepper motor. The linkage system allowed proper range of motion as well as path of motion for the canine s thigh. In order to move the the stifle and hock joints, they used a linear actuators. The linear actuators allowed both extension and flexion with the same motor. All three of these motors were attached to a brace that would go on the canine hind limb. For the prototype they simulated the hindlimb with a hard foam model based off of a 3D model of the test dog being used. This foam model, brace and motor system was then attached to a metal cart that the team built to simulate the approximate weight of a canine and to allow the system to move. Although last year s team accomplished a lot the mechanical system had some major areas for improvement, with the main area for improvement being the speed of the device. The linkage system used a rocker instead of a crank meaning that the stepper motor had to switch directions to accomplish the desired motion. This meant that the stepper motor had to stop and start each time the rocker switched direction, therefore slowing the overall speed of the device down. In addition the linear actuators controlling the stifle and hock movement were also slowing the device down by nature of how linear actuators function. The sensing system consisted of pressure sensors and gyroscopes attached to a separate brace custom made for the canine being tested on. The sensing system was tested on a live, healthy black labrador via a custom made brace that was form fitted to the canine. The pressure sensors were located on the bottom of the paw as well as at key locations along the hindlimb. The sensors on the paw allowed data to be gathered about where the canine is in the stance phase of 13

15 its gait. The sensors on the the hindlimb allowed data to be collected about the canine s kinematics. The gyroscopes were attached to the canine s front leg and allowed data to be collected about when the canine wants to move as well as kinematic data. This is possible because canines walk with opposite limbs moving simultaneously, meaning that as the right front limb moves, the back left limb moves at the same time. The combination of the pressure sensors with the gyroscopes allowed last year s team to gather all of the necessary kinematic and kinetic data to send to the motorized system to allow motion with the proper gait to occur. 14

16 Chapter 3: Ethics Throughout the product development and production phases of getting a product to market, there are multiple things to consider ethically. Speaking in a broad context, there are various ethical concerns to consider. When designing a mobility exoskeleton for canines, it is of the utmost importance to keep the safety of the canine in consideration. As an engineering on the project, it is especially important to have the correct certifications to work with live animals so they are treated in an ethical manner. With a possible motor and battery pack on the back of the exoskeleton, there cannot be any hot spots that will burn the canine. Additionally, the braces attached to the hind legs cannot cause raw spots from rubbing or prevent normal movement of walking. Once the exoskeleton model has been developed for a healthy dog, selecting canines that have Degenerative Myelopathy needs to be conducted ethically to insure the condition is not exacerbated and further injure the canine. The goal is to improve mobility, not diminish. Canines are treated as family members all over the world. There is a global need for this device because degenerative myelopathy can affect large and giant canine breeds. There are different social classes in the world and it affects the availability of care to canines. In third world countries, dogs tend to become stray and have no one look after them. Ethically speaking, it would be ideal for the design to be made with materials that can be sourced inexpensively so the market for the product is able to grow. During the 21st century, it has become extremely important for companies to make sure their design and production practices are environmentally friendly. Ethically speaking, dumping batteries into landfills is not eco-friendly. Once the batteries become old or are punctured, they can release toxins into the ground. If every canine exoskeleton uses disposable batteries, multiplied by the number of canines using the mobility system, and by the frequency of use, there are millions of batteries that could end up in landfills. Not only are there environmental ethical concerns to consider, but societally speaking, it is important for the canine exoskeleton system to be available to dog owners. For humans, health insurance covers certain procedures and braces, so what would be preventing pet insurance to cover the canine exoskeleton. Constructing a customized canine exoskeleton would cost a certain amount of money and it would be an ethical question as to where to draw a line as to how much would have to come out of the buyer s pocket versus pro bono or insurance. Making canine exoskeleton mobility systems available to all societal classes would be difficult due to the fact that there needs to be a source of income to make the braces but donations, sponsorships, and crowd funding can play a huge role in making this device available to everyone. 15

17 Chapter 4: Project Plan Summary The previous year s team was able to complete a proof of concept of the device. They created two different systems, a sensing input as well as an output system. They put the sensing system on a living healthy dog for testing but were unable to begin testing on a paralysed dog. This year s team plans to advance last year s design by making a second generation prototype and to begin testing on a paralysed dog. As a team we plan to integrate the two separate systems. This will involve gyroscopes being placed on the front legs of the dog. The signals received from the gyroscopes will then cause the hind limb braces to move accordingly. As a team we also aim to increase the speed of the device. Last years team has the hind legs moving but they move at a much slower speed than what is realistic for a dog. The third main objective for this team is to limit the weight and size of the device in its entirety. The dog will have the all the electrical and power elements carried on its back with the braces attached to the hind legs. The weight of the device needs to be minimized to prevent any future injury to the dog as well as to make it feasible for the dog to carry the weight. With all the changes we plan to make this year the end goal is to have a working device on a partially paralysed dog. We believe with the goals we have set we will be able to begin testing and begin helping partial paralysed dogs walk again. 4.1 Customer Requirements and Specifications Each dog owner has specific day-to-day activities. Building the Canine Exoskeleton has to accommodate a wide range of these activities, such as having a battery pack that does not need to be recharged during the day. There could be times where the dog owner won t have the ability to recharge the battery pack, thus making the exoskeleton inoperable. In addition to the battery pack, the exoskeleton has to be customizable for dogs of different sizes. The braces, mechanical system, and size of the dog pack and its contents would need to be able to fit small dogs such as Border Collies to Great Dane size dogs. An emergency stop function needs to be available in case something were to happen that could injure the dog. The emergency stop would act as a neutral position as to where the mechanical system can move freely without injuring the dog. 16

18 Customer Requirements Engineering Specifications Desired Change Target Value Cheap Monetary Cost of system ($) down $1500 Safe Probability of falling (out of 10 trials) down 1:10 Reliable Joint angle variation (degrees) down 30 Probability of device sensing dog s intended movement (out of 10 trials) up 6:10 Provides rehabilitation Rehabilitation time (days) down 90 days Provides mobility Distance dog can walk (meters) up 30 m Strong Force the system can support (lb) up 100 lb Speed Moves the dog at lifelike speed (mph) up 1.5 MPH Move in real time Delay time between sensing and motor system (seconds) down.001 s Comfortable Number of pressure points/stress concentrations down 1 Battery Life Time the battery provides constant power (hrs) up 5 hrs Weight The weight of the dog pack (lbs) down 20 lb 4.2 Constraints Table 1: Customer Requirement and Engineering Specifications There are multiple design constraints that have to be taken into account, with respect to the dog owner, the dog, and the team. The design of the braces and the dog pack that contains the motors, batteries and other components must be comfortable to the dog. The weight limit for the dog pack is about 25% of the dog s weight. Taking this into consideration, the weight of all of the components needs to be closely watched because dogs are of all different sizes, constricting the overall capacity of the dog pack. The braces attached to the hind limbs cannot pinch anywhere. The braces cannot be cumbersome so as the movement of the dog is not hindered. With respect to the team constraints, money is a large factor. Each canine exoskeleton system requires new 17

19 braces, a dog pack, motors and circuitry which can become extremely costly. Optimizing the cost of each system is essential. In order to maximize the efficiency of the system, the speed of the microprocessors needs to be ideal. This means that there is no lag time between the sensing systems and motorized systems. Figure 4: Goals and Constraints for this year s team Figure 5: Risk analysis for design improvements 18

20 Chapter 5: Hardware 5.1 Previous Design The previous design team created a two part system that consisted of a mechanical cart and sensing system as previously stated in Chapter 2. Below shows the previous mechanical prototype system for reference. Figure 6: Generation 1 prototype Figure 7: Generation 2 prototype This year s second generation device will be built off of the foundation of the first generation device, but with our new design systems implemented as described in this chapter. As seen in the 19

21 figures above, the second generation prototype will integrate the two systems from the first generation into one system. 5.2 Linkage System In order to provide motion for the hip joint of the canine, a linkage system was fabricated similar to last year s design with some slight improvements. The program, Linkages, was used to create this linkage system as seen in figures 7 and 8. Figure 7 shows the general path of motion that the new linkage system will take, and figure 8 shows the exact range of motion that the thigh of the canine will have while walking with the maximum and minimum hip angles reference to ground on the right side of the figure. Figure 8: Linkage Path of Motion using Linkages Figure 9: Hip Angle The linkage system was changed from using a rocker as the driving link to using a crank as the driving link. This change will allow the system to achieve the desired motion at a faster speed due to the motor not having to stop and reverse direction each time the rocker link changed direction while maintaining the required torque of 7.8 Nm to rotate the hip joint. This design change in addition to the changes being made to the stifle and hock actuators should solve one of 20

22 the major problems of last year s prototype, which was the speed of the device. The new linkage system can be seen in figure 9. Start of Stance Mid Stance Swing 5.3 Pulley System Figure 10: The New Linkage System Throughout the Gait Cycle One improvement to the previous design was to create a new system to control the linear movement needed to induce flexion in the stifle and hock joints. The previous system used linear actuators to control this motion. The problem with these is that the maximum speed is slow and do not allow motion unless powered, which could be a potential hazard if the dog resists the locked position. The new design to be implemented will consist of a brushless DC motor with a pulley and string which will be able to actuate the joint in the same motion by winding the string around a pulley attached to the motor. In addition, the new linear actuation will consist of a spring loaded piston in order to return the joint to the extended position along with providing a dampening effect for the force caused by the impact with the ground of the canine s leg. The spring piston will have an adjustment so that the resistance that the spring provides is selectable depending on the weight and needs of the individual canine. The figure below shows the proof of concept for the pulley system. 21

23 Figure 11: Pulley system The figure above shows the pulley system attached to the current prototype brace. The blocks of wood that were used to create the pulley system proof of concept were superimposed laterally to the brace. The red arrow indicates the brushless DC motor with the pulley attached. The string, indicated by the green arrow, is attached between the pulley and the calf of the dog brace. The blue arrow indicates the electronic speed controller (ESC) that is used to control the motor. Below in Figure 12, as gas spring used for closing a storm door is similar to the design that we intend to use for our spring system. Though the mechanism of this spring is different from what our intended design will be, the motion of the spring is in line with our vision. Our free moving piston spring design will use a mechanical spring instead of gas to create the rebounding effect for our exoskeleton. Our design will utilize an adjustable collar that will be able to adjust the tension that the spring provides. This adjustment will allow for a wider range of dogs to be assisted by our exoskeleton. By adjusting the spring, we will be able to both meet a wider range of weight specifications of a dog as well as provide a varying range of assistance to different dogs depending on their specific disability and degree of paralysis in the legs. 22

24 Figure 12: Piston type gas spring [7] In order to support the canine while the motors are inactive, the springs will be able to support a minimum of 25 lbs per spring. These springs accompanied by the piston will act as the extension force for the stifle and hock joints, while the motor will act as the flexion force. From biomechanical analysis, the motor will have a torque requirement of 32.5 Nm in order provide enough force to flex the stifle and hock joint. 5.4 Dog Pack and Electrical systems The electrical systems consisting Arduinos and motor driver will be housed in a dog pack that was donated by Ruffwear, a manufacturer of canine outerwear. The pack will be modified as necessary to accommodate the stepper motors and motor drivers that control the hip joint as well as the wires running from the front and back of the exoskeleton. The batteries along with all of the electrical systems should not exceed the rated weight provided by Ruffwear of 20% of the canine s body weight. The electrical system must be protected from damage that may occur which includes water damage, abrasion, and kinking. The design will implement a length of tubing to allow for the wiring needed to be run through in order to prevent unforeseen damage. The entirety of the electrical system to be contained in the dog pack will include: 2 lithium batteries, 2 stepper motors, 2 stepper motor drivers, and 1-2 Arduinos. Additional processing may be needed along with receivers for the sensing system. Below shows a picture of the Approach Pack that was donated to our design team by Ruffwear. 23

25 Figure 13: Ruffwear Approach Dog Pack Figure 14: Schematic design for the exoskeleton (excluding sensing system) Figure 14 shows the electrical systems that will be implemented into the exoskeleton. The power source will consist of the 2 lithium polymer batteries that can be recharged. The main control system will consist of the arduino which is connected to both stepper motor drivers and the 4 24

26 ESCs that drive the brushless DC motors. The motors will be controlled with the Arduino digital pins to send the proper PWM signals to each of the motor drivers. The sensing system consisting of the pressure sensors and the wireless gyroscopes are not shown, but will also provide feedback through the Arduino via the analog and digital pins. 25

27 Chapter 6: Software Systems 6.1 Sensing system The sensing system was originally a series of seven pressure sensors on each back leg and a gyroscope on each front leg. The series of sensors on the back legs were meant to read movement of the upper leg and the sensors were to sense proper gait. The problem with this is that, after talking to the VTH, we cannot assume the back legs will have the neurological function to move enough to trigger the process of movement. The assumption that is being used now is the dog has no neurological function in the hind limbs so the front limbs, which are assumed to be functional, must initiate movement. Gyroscopes are used to make this initiation of movement possible. Gyroscopes have multiple degrees of freedom, the ones being used have six, and it displays acceleration of three axis and angular velocity in three axis. The intent is to use the gyroscopes to initiate the back leg motion while the front leg moves at the same time. The plan is to use them to initiate this by generating that acceleration and angular velocity and sending that through a board and to a radio frequency transmitter. The receiver that is attached to the main board and it will read in the data and be able to use the data to make decisions on how to move. The sensors are the other part of the sensing system. The old team had a set of seven per leg in the Generation 1 design but the plan is to limit them down to four because some of them were reading in data for a healthy dog. That data from those sensors can no longer be assumed as true because, as the VTH has made clear, having proper back leg motion can t be assumed. We are actually advised in the opposite direction, we are to assume that the animal has no function of the hind limbs at all. Since that is the case, we took out the three sensors that did that but kept the ones that will ensure proper gait because, if proper gait is not happening, modifications in the main board will need to be made. Currently, these are only for testing and at some point maybe coded into a feedback loop that also helps this dog move more naturally. 6.2 Main Computer The main computer holds the main code that runs everything attached to the mechanism. It will take in the data from the receivers of the gyroscope circuits then take that in and run it through the algorithms to see if the dog is trying to move or if it is just fidgeting. If the algorithm finds that the dog is moving, then it sends the gyro numbers through an equation to tell it how fast it is moving, which then, with other equations that have been created, tells the motors on the desired leg to move for a set amount of milliseconds then stop turning. Our code will also have a 26

28 fail-safe mechanism that if a button is pressed, we can bring the dog down to a safe sitting or lying position. 27

29 Chapter 7: Logistics 7.1 VTH Collaboration Members from the VTH are working in collaboration with the project. These individuals include, Dr. Dean Hendrickson, Dr. Rebecca Packer, Dr. Nic Lambrechts, and physical therapist Sasha Foster. Together, these individuals help with overseeing animal care, as well providing neurological and orthopedic insight to different design iterations. Along with information about dog breeds and walking mechanics, these individuals also helped spread the word in looking for a disabled canine to complete the project on IACUC The Institutional Animal Care and Use Committee (IACUC) provides certification for ethical use of animals in research and testing throughout all of CSU. Each team member had to complete training by reading passages and answering questions to receive the certification. The project plan also had to be approved by IACUC before any interactions with a canine were allowed. Please reference Appendix E RICRO The Research Integrity & Compliance Review Office supplies certifications in ethics and compliance. Although this certification was not required, each team member completed it because this project does have large ethics background. The team wanted to be as educated and certified as possible to help IACUC approval as well as to give owners peace of mind their dog s well-being was of utmost importance. Please refer to Appendix E for each team member's RICRO certification. 7.2 Orthopets Collaboration A representative from Orthopets, Martin Kaufman the founder of Orthopets, has been in close collaboration with this project since the beginning. Martin has a wealth of knowledge in canine kinetics, kinematics and physiology the combination of this knowledge makes him a very valuable asset in designing the exoskeleton with the canine in mind. New design changes are run by Martin so he can help with working out the tweaks and making fine adjustments to make the changes most compatible for the canine. In addition to helping out the team with this knowledge and experience, he has also helped the team with fabrication of the device itself. The current sensing brace and prototyp brace from last 28

30 year were both fabricated through Orthopets with the help of Martin. Orthopets plans to help with the fabrication of any future devices as well. Figure 15: Dog wearing an Orthopets Brace [Orthopets 2015] 29

31 Chapter 8: Fundraising This project does not have an industry sponsor and therefore does not have a large budget. One of the many challenges last semester faced was fundraising. As improvements needed to be made the lack of a large budget makes it very difficult. A lot of the items required for purchase are pricey. Knowing this from the beginning, fundraising has been a large focus for the first semester. Many funding opportunities have been taken advantage of, such as Keysight sponsorship, CSU Charge, as well as reaching out to personal connections. Keysight held a sponsorship competition for $1000. The first stage of the sponsorship was a written proposal. The team completed this and made it onto the next round. Round two in competition for the sponsorship was a 30 minute presentation to a representative, Dan Ferguson. There were a total of four groups to make it to the second stage of the sponsorship. The Canine Exoskeleton team was the finalist and was awarded the $ donation. Along with seeking out sponsorships from companies the team also created a funding page. The team has created a campaign with CSU Charge! in an attempt to receive more money from donors. This campaign took multiple months to set up. The team has reached out to social media platforms with a wide range of followers to increase the audience exposed to the project. The campaign extends until after the semester is over so it is unclear how much income this fundraising method will bring in. Regardless of the amount of monetary value obtained the campaign has been exposed on social media by a variety of people. This has given the project a platform to spread the word about what students at CSU are working towards. 30

32 Chapter 9: Testing A step by step detailed testing document is located in Appendix G. The device in its entirety will be tested before it is attached to a live canine using the cart, then further testing will be completed with the device on the canine. The cart manufactured last year will be used this year as the testing fixture. The braces from the disabled dog will be attached to this cart, the input data will come from gyroscopes attached to the front legs of a healthy canine. The final device will be tested to see if the hind braces move in real time and synchronized with the front legs of the real canine. The new linkage system needs to be tested with the current cart to see if the crank design, instead of the rocker design, creates less of a time pause in the dog s movement. The new batteries need to be tested to see how long the lifespan is. The new design of a brushless motor and spring will need to be tested. The spring will need to be tested to assure that it can hold the weight of a canine in a standing position as well as be able to compress with the tension provided by the string attached to the motor. The motor needs to be tested to assure that it produces enough torque to overcome the spring constant while still being fast enough to move at the speed of a typical canine. The gyroscopes will be tested to to assure they sense the correct direction and speed the front legs are moving so the hind braces can follow accordingly. 31

33 Chapter 10: Conclusion and Future Work 10.1 Uncertainties Due to the nature of this project working with live animals, it is necessary to be certain that the device meets all of the requirements needed to assist the canine in real time safely and effectively Springs Uncertainties moving forward include are the spring mechanism. This is a design change that came up late into the semester, a strong proof of concept has yet to be designed. The team has incorporated a small spring, but not in the location that was recently decided on. The spring constant of the spring needs to be calculated. Once the spring constant has been calculated the team will either find a spring on the market or come up with a plan to design one. After this is complete the design in its entirety will be able to be tested and adjusted as needed Coding Speed The speed of the code is a major problem with doing a real time feedback system. With as many equations as the gyroscope values have to go through to get to the on-time for the motors, the speed of the code could slow down because of the calculating speed. The belief is that with only twelve values being run through the equations, it should not slow it down a significant amount but we may have a speed issue if it is run for a long period of time Dogs Reaction Every dog will have a different and unique reaction to being placed in the brace system. The group will not know how the dog will react until they are already in the brace. Some dogs might freak out and try to wiggle their way out of the braces. Other dogs might not want to move once they are in the braces and it might be hard for the team to get the dog to take a first step Plans for Second Semester Second semester is the beginning of the testing phase for the canine exoskeleton. The redesign of the brace with the new motor and pulley system must be finished before moving forward with testing on dogs with partially paralyzed hind limbs. 32

34 Finding Dogs with Partially Paralyzed Limbs The team is in the midst of acquiring dogs that fit the design s specifications. Dogs that have partially paralyzed hind limbs must pass a neurological and physical examination at the VTH in order to continue through the testing phase Customized Canine Exoskeletons Once the dogs pass the examinations, customized braces will be fitted to each dog. Each dog will receive it s own linkage, pulley and motorized systems Final System for the Second Semester At the end of this year the Generation 2 Canine Exoskeleton System will be at a point to where angular measurements of dog s hind limbs will be taken and braces manufactured so the electrical and mechanical systems can be imposed upon and allow the disabled dog to walk. The Generation 2 Canine Exoskeleton System will include the updated crank linkage system for actuating the hips, brushless motor with piston system for actuating the stifle and hock joints, wireless gyroscopes for sensing canine movement, and a dog pack for housing all electrical components. In addition to this the sensing and mechanical systems will be fully integrated and able to function in real time at a physiologically acceptable walking speed of 1.5 mph while on a disabled canine Plans for Future Years The canine exoskeleton should provide the insight into how an exoskeleton system should be designed and the constraints that are important to keep in mind for the safety of the subject using the device. Our design has the short term goal of creating a usable device that could improve the lives of canines with hind limb paralysis for both rehabilitation and augmentation purposes. This year, the design team may find better alternatives to the sensing system currently being used and in future years, the design may implement those ideas. Additionally, if the design shows promise for the a real market product, custom manufactured parts and processes could be implemented. The longer term goal is to translate the design to fit the needs of a paralyzed or partially paralyzed human. This project will give insight to the necessary precautions and certifications to work with live test subjects in an ethical and safe way for future designs Final Thoughts The team has done a lot of brainstorming and working with ideas to try and address the current limitations of the device. The team believes their current design path will be successful. Come 33

35 the beginning of second semester the team will be ready to fully test the 2nd iteration of the design. The testing trials, successful and unsuccessful, will be extensively documented to help us prevent remaking mistakes, as well as helping those in the future continue work on this project without having to redo work this team has already completed. 34

36 References 1. Bos et al. Journal of NeuroEngineering and Rehabilitation (2016) 13:62 DOI / s z Ekso Bionics. Web. 21 Sept Bowler, Niel. Rise of the human exoskeletons. BBC.com. 4 March Web. 21 Sept Indego. Web. 21 Sept Generation 1 project report ines_report.pdf 6. Generation 1 project website

37 Appendix A - Abbreviations CSU: Colorado State University DM: Degenerative Myelopathy IACUC: Institutional Animal Care and Use Committee RICRO: Research Integrity and Compliance Review Office OSHA: Occupational Safety and Health Administration VTH: Veterinary Teaching Hospital Appendix B - Budget Figure 16: Budget 36

38 Table 2: List of costs incurred and deposits made into the team s account. Starting Balance ($) What? Total ($) Paid For By Last Year s Team Balance BIOM Contribution Contribution ECE Contribution Contribution Brushless Motor Personal Steel Cable Personal ( x 2) Batteries (Donation) Donation (-79.95) Ruffwear Dog Pack (Donation) Donation Arduinos Personal Transmitter/Receivers (x2) Personal +200 Donation from JB s father Donation Linkage Materials Personal Keysight Sponsorship Sponsorship Tales of a Lab Donation Donation In regard to the braces, Martin Kauffman is donating the first set of braces made. Hinges will be recycled from last year s team to keep costs down. All subsequent braces made will cost $400 per leg. Each dog that is examined at the VTH will incur a $15 exam fee. Dogs that are already patients at the VTH will have had their examinations and won t need to be examined again. The fundraising campaign has not ended yet so the team will have extra money added to the budget. In regard to donations, two batteries were donated from Colleen s parents while JB s father donated $200. Ruffwear Performance Dog Gear donated a pack to house all of the electronics. A sponsorship from Keysight Technologies was secured through a competition. 37

39 Appendix C - Project Plan Evolution Project Plane 12/9/16 Table 3: Timeline for the holiday break Due Date Dec 16 Dec 21 Dec 26 Task Final decision on Kono Calculate torque and power specifications needed to move canine Final decision on disabled dog (if didn t go with Kono) Test torque and power specifications needed to move canine Start testing gyroscopes Spring design finalized Jan 2 Integrate brushless motor and spring design into device including manufacturing the piston, mounting the hardware, and testing on cart Dog pack set up with pouches to house all electrical and mechanical components Measurements of new canine completed New design for braces completed (cleared with Martin) Ensure sensors will be sufficient for tracking proper dog gait Jan 9 Go down to Orthopets to manufacture new braces for the disabled canine Complete testing for torque and speed on cart and design a metric to establish if the torques and speed are sufficient for testing on an animal Use Maya to test cart to see if it moves in real time Jan 10 Jan 17 Start adding mechanical parts to the braces for disabled canine Completed new device and ready to test on disabled canine 38

40 Table 4: Timeline for the second semester till E-Days Date Jan 20 Jan 23 Feb 1 Feb 14 Feb 17 Mar 1 March 7 Task Place disabled canine in brace but do not test walking, just test comfort and familiarity Start collecting data on disabled dog according to IACUC protocol Have collected all data Have partial data analysed, decide if data is quality enough, retest /redesign if needed Started any retests/design tweaks if need Analyzed all data Finished revising any design changes Start retesting if needed Start E-Days report March 14 March 21 Apr 14 Secondary testing completed Secondary testing data analyzed Ready to present for E-Days 39

41 Project Timeline 10/28/16 Table 5: Timeline for the second semester till E-Days. Due Date Oct 28 Task Find forces on a dog hindlimb Feedback on brushless motor testing - Order more if they work well, if not research other options Nov 4 Linkage system changes with crank Gyroscope set up to receive and output proper data Nov 18 Decided on stepper motor replacements Integrate brushless motors into design Set up another meeting with Martin to discuss making braces for a diseased dog Dec 2 Dec 9 Have a setup with the dog pack for carrying all the electronics and the stepper motors Design changes completed Found a diseased dog with hindlimb paralysis New device for diseased dog that contains integrated systems Jan 17 Apr 7 Apr 14 Have fully functioning device ready to be tested on a paralyzed dog Have tested the device on a paralyzed dog and revised any flaws in the design E-Days 40

42 Project Timeline 10/6/16 Figure 17: Original Gantt Chart for

43 Appendix D - Project Code Files //include libraries #include <MPU6050_9Axis_MotionApps41.h> #include <MPU6050.h> #include <math.h> #include "I2Cdev.h" #include <Wire.h> boolean calibrationmode=true; int count = 0; int accfactor; int velfactor; double gfactor=9.81; double accx1, accy1, accz1,accx2, accy2, accz2; double gyrox1, gyroy1, gyroz1, gyrox2, gyroy2, gyroz2; double accx1offset, accy1offset, accz1offset,accx2offset, accy2offset, accz2offset; double gyrox1offset, gyroy1offset, gyroz1offset, gyrox2offset, gyroy2offset, gyroz2offset; int16_t tempraw; int16_t ax1, ax2, ay1, ay2, az1, az2, gx1, gx2, gy1, gy2, gz1, gz2; // make the gyros MPU6050 gyro1(0x68); MPU6050 gyro2(0x69); void setup() { Wire.begin(); Serial.begin(115200); gyro1.initialize(); gyro2.initialize(); delay(100); // this grabs the initial raw values from the gyros gyro1.getacceleration(&ax1, &ay1, &az1); gyro1.getrotation(&gx1, &gy1, &gz1); accx1 = ax1; accz1 = az1; accy1 = ay1; gyrox1 = gx1; gyroy1= gy1; gyroz1= gz1; gyro2.getacceleration(&ax2, &ay2, &az2); gyro2.getrotation(&gx2, &gy2, &gz2); 42

44 accx2 = (double)ax2; accz2 = (double)az2; accy2 = (double)ay2; gyrox2 = gx2; gyroy2= gy2; gyroz2= gz2; } void loop() { if(loopconst==1){ for(int i=1; i<=1000; i++){ gyro1.getacceleration(&ax1, &ay1, &az1); gyro1.getrotation(&gx1, &gy1, &gz1); accx1 = ax1; accz1 = az1; accy1 = ay1; gyrox1 = gx1; gyroy1= gy1; gyroz1= gz1; gyro2.getacceleration(&ax2, &ay2, &az2); gyro2.getrotation(&gx2, &gy2, &gz2); accx2 = ax2; accz2 = az2; accy2 = ay2; gyrox2 = gx2; gyroy2= gy2; gyroz2= gz2; accx1offset=accx1offset+accx1; accy1offset=accy1offset+accy1; accz1offset=accz1offset+accz1; gyrox1offset=gyrox1offset+gyrox1; gyroy1offset=gyroy1offset+gyroy1; gyroz1offset=gyroz1offset+gyroz1; accx2offset=accx2offset+accx2; accy2offset=accy2offset+accy2; accz2offset=accz2offset+accz2; gyrox2offset=gyrox2offset+gyrox2; gyroy2offset=gyroy2offset+gyroy2; 43

45 gyroz2offset=gyroz2offset+gyroz2; } calibrationmode=false; accx1offset=accx1offset/1000; accy1offset=accy1offset/1000; accz1offset=accz1offset/1000; gyrox1offset=gyrox1offset/1000; gyroy1offset=gyroy1offset/1000; gyroz1offset=gyroz1offset/1000; accx2offset=accx2offset/1000; accy2offset=accy2offset/1000; accz2offset=accz2offset/1000; gyrox2offset=gyrox2offset/1000; gyroy2offset=gyroy2offset/1000; gyroz2offset=gyroz2offset/1000; }else{ gyro1.getacceleration(&ax1, &ay1, &az1); gyro1.getrotation(&gx1, &gy1, &gz1); accx1 = ax1 - accx1offset; accy1 = ay1 - accy1offset; accz1 = az1 - accz1offset; gyrox1 = gx1 - gyrox1offset; gyroy1 = gy1 - gyroy1offset; gyroz1 = gz1 - gyroz1offset; gyro2.getacceleration(&ax2, &ay2, &az2); gyro2.getrotation(&gx2, &gy2, &gz2); accx2 = ax2 - accx2offset; accy2 = ay2 - accy2offset; accz2 = az2 - accz2offset gyrox2 = gx2 - gyrox2offset; gyroy2 = gy2 - gyroy2offset; gyroz2 = gz2 - gyroz2offset; //first set print // Serial.print(accX1); Serial.print("\t"); // Serial.print(accY1); Serial.print("\t"); // Serial.print(accZ1); Serial.print("\t"); // Serial.print("\n"); // 44

46 // Serial.print(gyroX1); Serial.print("\t"); // Serial.print(gyroY1); Serial.print("\t"); // Serial.print(gyroZ1); Serial.print("\t"); // Serial.print("\n\n"); // second set print // Serial.print(accX2); Serial.print("\t"); // Serial.print(accY2); Serial.print("\t"); // Serial.print(accZ2); Serial.print("\t"); // Serial.print("\n"); // // Serial.print(gyroX2); Serial.print("\t"); // Serial.print(gyroY2); Serial.print("\t"); // Serial.print(gyroZ2); Serial.print("\t"); // Serial.print("\n\n"); // delay(2000); } } 45

47 Appendix E - IACUC and RICRO Certifications Each member of the group completed IACUC and RICRO certifications through CSU to work with and test on live animals. The certifications are below. 46

48 47

49 Appendix F - Donation Letter Canine Orthotronic Mobility System Colorado State University Senior Design Project Colleen Jones, Computer Engineering Student Jordan Bernhardt, Biomedical & Electrical Engineering Student Kat Killingsworth, Mechanical & Biomedical Engineering Student Lindy Gillette, Mechanical & Biomedical Engineering Student Ryan Leuenberger, Mechanical & Biomedical Engineering Student Dr. Anura Jayasumana, Electrical & Computer Engineering (ECE) Project Advisor To whom it may concern, Every day, some dogs lose their ability to walk due to hind limb weakness or paralysis from a variety of diseases. These symptoms can appear with no warning and leads to the paralysis of dogs hind legs and sometimes even the front legs. Many of these conditions are nonprogressive and dogs can recover their ability to move over time with physical rehabilitation. However, many owners do not have the capability to provide treatment and thus end up using wheelchairs or other relief devices that do not help the dog regain strength. They remain unable to walk properly, and so there is a need for a physical therapy device that pet owners can use. We are the Canine Orthotronic Mobility System Senior Design team from Colorado State University, comprised of biomedical, computer, electrical, and mechanical engineering students. The goal of our project is to create a rehabilitation exoskeleton for dogs suffering from these mobility impairments. This device would be placed on the paralyzed dog and aid in motor functions until the dog regains its strength. The research accomplished for this project will, one day, also be applied to modern human physical medicine to ultimately develop a device that will mobilize paralyzed individuals. This second year of the project we are building off of what the previous year accomplished. We will work towards fine tuning the sensor system which was comprised of accelerometers, gyroscopes, pressure sensors, and Arduino microcontrollers, which will then process the data being collected and will replicate the angles of movement onto a motorized exoskeleton comprised of linear actuators and stepper motors. In addition to improving the standing device our team also plans to create four additional devices and to put the integrated device onto a canine with hind leg paralysis. We are writing to you to ask about possible funding opportunities to aid our efforts in creating this system by the Senior Design Showcase Event in April of We rely on the support of industry members, consumers, mentors, and those who are interested in the cause in order to be successful. Movement of the exoskeleton will require powerful, expensive actuators to compensate for large loads since we will be working on modeling a larger dog breed, a Labrador retriever. In addition to load capacity, these actuators must be extremely precise and fast since we are modeling live animal movement. The sensors must also be extremely accurate when it comes to leg angles, speed, and pressures. We are also working diligently with our academic advisor and the CSU Vet Teaching Hospital to more thoroughly understand what type of conditions and problems our device must be capable of handling. These expensive devices, construction of our moving model, electrical and mechanical device testing, and 48