Student Projects
Contact Information
Competition Year: 2016
University: University of Leeds
Team Members: Simon Bye 2015
Faculty Advisers: Professor Ian Roberston
Email Address: Simonbye82@hotmail.co.uk
Country: UK
Project Information
Title: Strength Enhancing Exoskeleton
Description: The project involved the creation of a strength enhancing exoskeleton
Products: Compact Rio, LabVIEW RT and FPGA, V3 Muscle Sensors.
The Challenge
To design and develop a strength enhancing exoskeleton that is cheap, reliable, easily controlled and provided enough degrees of freedom for a good range of movements that could be used to aid rescue operations after natural disasters.
The Solution
Use a CompactRIO to control the actuation of an exoskeleton through the use of muscle sensors and force resistive sensors to achieve robust and accurate control over the exoskeleton’s movements.
Exoskeleton demonstration video link
https://www.youtube.com/watch?v=A91GoxHYcBM
About the author
This was my individual project whilst working towards an MSc in Mechatronics and Robotics at the University of Leeds. Since completing my masters, I am now working as an Animatronic Control Engineer building interactive robotics for the entertainment industry.
Introduction
In recent history, there has been a global increase in the number and intensity of naturally occurring and man-made disasters. Many of these disasters occur in inaccessible areas, particularly in less economically developed countries. In 2010, Haiti was hit by an earthquake which left 19 million cubic metres of debris. Accessibility for heavy plant machinery is difficult and so removal of large heavy objects is not possible, potentially leaving people buried with no hope of rescue. This is one example of the many possible scenarios where having a highly mobile lifting unit would be of great benefit. A strength enhancing exoskeleton suit would give search and rescue teams easier access to areas cut off by rough terrain or other obstacles and would enable the operator to remove objects that would usually be beyond the capabilities of normal human strength.
This project was set out to create a prototype strength enhancing exoskeleton with limited degrees of freedom, providing a solid initial platform to develop the suit further. Although Leeds University is developing assistive exoskeletons, I chose this project due to the fact that there has been no work on strength enhancing suits to date. I was confident that I could develop one within the short time scale of the project that could out perform an average person’s lifting capabilities. The aim of the initial prototype was to be able to bicep curl 150 kg, which is roughly equivalent to the weight of 2 people. The suit needed to be easily transportable, with the exoskeleton being able to be split into individual components (arms, legs and body) so it can be packed up in a convenient storage case. It should also be easily reassembled by the operator without need for extra equipment; less than 10 minutes was the goal. The exoskeleton must run off its own power; this is partly due to portability but also as a consequence of the suit being deployed in areas where infrastructure has been severely damaged so it is unlikely that any power networks would be available. I tried to develop the prototype to be as inexpensive as possible. The lower the cost of the final product, the easier it will
be for search and rescue organisations to have numerous exoskeletons, allowing for greater availability and an increased chance of saving lives. Finally making the suit robust and durable was essential due to the nature of the environments it will be used in. The final design needed to be able to deal with impacts from debris, a variety of weather conditions and prolonged operating times.
Development
The final prototype works by firstly detecting the intended direction the operator wants to move through a sensitive force resistance sensor. Once the motion required has been identified, the relevant muscle sensor becomes active. The more the muscle is strained, the higher the potential difference output is. This PD is rectified to give out a positive signal between 0 and 10 volts. The raw signal is fed into the CompactRIO’s analogue input module and filtered to produce a smooth output signal to a PID controlled pressure valve. The valve outputs 1 bar of pressure for every volt it receives giving a maximum output range of 10 bar. The more the muscle is tensed the higher the voltage transmitted. A percentage strength enhancement can be assigned to the operator by modifying a variable in the code. This allows the exoskeleton to be quickly tuned to work well with operators of varying builds with a simple recalibration of the power output. The variable power and pressure outputs allow the exoskeleton to operate efficiently by conserving energy when only light loads are being moved. When out in the field energy, efficiency is a huge bonus as the suit can operate for longer; potentially saving more lives.
Figure 1 Finished prototype front view (left) and view of CompactRIO connected into the back (right)
I used a myDAQ to develop the initial code and test the functionality of the V3 muscle sensors I had selected. I chose myDAQ because it was small, light and portable and I could also run simple tests to
evaluate the effectiveness of different code with just my laptop. I could also easily modify the code developed with the myDAQ to work with other hardware, making the transition from initial tests to deployment much simpler. I found the raw signal from the muscle sensors to be uneven and noisy and so it did not work well with the electric pressure control valves. To make the signal smoother I used a Butterworth filter resulting in a more fluid arm movement.
Figure 2 – Raw signal from sensors (left) and filtered signal (right)
For the full prototype I needed an embedded controller that could read all of the input signals from the muscle sensors then send the control signal to the hydraulics with minimal delay. Any lag between the operators arm motion and the activation of the actuators would make the exoskeleton harder to control. The controller also needed to have large enough memory and fast processing speed to control all aspects of the exoskeleton simultaneously, such that loads can be lifted evenly with both arms. This is crucial for the safety and effectiveness of the exoskeleton. After researching the various options on the market, including a number of Altera FPGA boards and the NI Single-Board RIO and myRIO controllers, I decided to use an NI CompactRIO. This controller appealed to me because of its scalability, customisable I/O modules and the large FPGA chip size. None of the other devices had the number or variety of analogue I/O channels required for the amount of sensors and actuators in my design and the hot-swappable C series modules allowed for more flexibility during development. The NI CompactRIO real time operating system is deterministic and when combined with the high speed FPGA chip is extremely reliable. A unique benefit of the CompactRIO for this project is its Class 1 Division 2 hazardous environments rating. This rating qualifies the device to be used in conditions with an ignitable concentration of flammable gas within the atmosphere up to temperatures of 135ᵒC. It also has an impact rating of 50g over 3ms. For the hazardous environments this prototype is designed to be used in. This is ideal. When powered the FPGA and RT programs run headlessly. The C series analogue input module continuously reads all the sensors and the collected data is processed in sets of 2500 samples every 2 micro seconds. The filtered data is transferred to the RT portion of the controller to utilise the increased memory during signal analysis. This signal is then outputted from the CompactRIO at a low voltage to control a set of solid state relays which send a 24V signal to the pressure valves.
Conclusion
This project has demonstrated that it is possible to develop low cost reliable exoskeletons for search and rescue. Due to the flexible and scalable nature of the CompactRIO controller adding more sensors and actuators to increase the manoeuvrability is a simple task and this would greatly increase the prototype’s lifting potential. I believe it is a very real possibility that this exoskeleton prototype could be developed into a fully functioning search and rescue unit capable of rapid
deployment in all areas and conditions. With more time and money I would like to upgrade to hydraulics for more power and more control over the motion. I would also add actuators to the legs.
Exoskeleton demonstration video link
https://www.youtube.com/watch?v=A91GoxHYcBM
