Our proposed design uses aluminum alloy 6061. This alloy of aluminum is
commonly used and readily available, providing for low cost. It is also
light-weight and easy to work with (machining, bending, welding). In
comparison, stainless steel (eg. Type 303) is around 2 to 3 times denser, while being not as
proportionately strong. Also, it is more difficult to weld than aluminum. Polymers that do provide
the needed strength are much more expensive.
Our proposed design uses McKibben air muscles. The main method of eliminating competing designs was cost. Through Prof. William Melek's research in robotic gait motion, motors that would meet our power requirements would run in the high hundreds of dollars (per joint). Electro-active polymers are also prohibitively expensive. EAPs providing our force and actuation distance requirements would run in the order of thousands of dollars per joint. Hydraulic or pneumatic cylinders were eliminated based on cost, ease of application, and power-to-weight ratio. Although pistons typically come at an affordable price in standard sizes, but customized pistons are expensive. In contrast, air muscles can be custom designed and built for tens of dollars each. Also, the air muscle's dominant 400:1 power-to-weight ratio (of the muscle only, not including the air supply equipment) makes using air muscles a very convincing choice.
Our proposed solution uses an Arduino microcontroller. It is a cheap, easy, and popular embedded system prototyping platform. It is very cost effective for our purposes, since it is a ready solution to a lot of hard problems encountered designing embedded system hardware that would take months to solve independently, at a price of less than a fifth of our overall budget. It is easy to program for and has a large support community to help us along in case problems occur. Its popularity speaks about its quality and value as a platform for hobby and research projects.
In particular, we have decided to use the Arduino Mega for its plentiful input and output (I/O) ports and added memory over the baseline models.
Our design has both soft stops (software features which ensure the exoskeleton will not perform dangerous tasks) and hard stops (physical limits as to how much force can be applied and the positions the mechanism can take).