RELEVANT TO MY INTERESTS!
Though due to technical interest (well, academic interest, I don't have that MEng
yet), I must nitpick incessantly.
On the limb itself:
The hand looks fine, but the arm needs a few more DOF to be practical, especially for the wrist. A rotation point in the lower arm, and a yaw axis for the wrist would allow much more natural motion (try moving your arm without rotating your wrist or lower arm at all. You'll find you need to pretty much turn your entire body to compensate). Joint motors are fine until somebody figures out how to make artificial muscle that contract quickly, powerfully AND for a reasonable distance. Current state of the art is two-out-of-three, with low power efficiency. For fingers, however, joint motors that small aren't powerful enough, so wire tendons are used with the actuators being mounted in the lower arm.
On attachment:
Rin is 'lucky' in that she still has stumps. Currently, we have no really effective prosthetic shoulder joint. 99.9% of prosthetic limbs attach with a 'stump cup', a moulded plastioc/silicon socket which fits tightly over the stump, using friction to hold the limb in place, often with some straps for re-enforcement. Even if this cup is an absolutely perfect fit, it still rubs and chafes constantly, creating painful sores, especially with long-term use. This is why many prosthetic users simply never use their limbs. There has been some advancement with using Osseointegration to attach limbs. This involves a semiporous titanium rod being inserted into a bone, where the bone itself joins permanently with the rod, leaving an attachment point sticking out the end. Attaching a limb is then as simple as plugging it into the mounting point. Very comfortable, and can be done by the prosthetic user themselves. It does have a few disadvantages though: the attachment point must pass through the skin, and this creates an infection risk if not kept clean (it's essentially an open wound). A sufficiently large impact to the mounting point could shatter the bone from the inside, causing irreparable damage (this shock loading problem is why there has yet to be testing of osseointegrated lower body prosthesis from the worries of repeated impact without the shock absorbing properties of the lower leg), but this is generally a nonissue for arm prosthesis. To my knowledge, there are only two people win the world with osseointegrated arm prosthesis.
On control:
Current state of the art is Targeted Muscle Reinnervation, where the remains of nerves that once controlled the arm (motor neurons) are surgically reattached to muscles in the chest. Signals from these nerves cause the muscles to contract, which can be measured by surface electrodes (myoelectrics). Experiments have been performed where sensory neurons have been routed to skin on the back, resulting in a limited sensation that appears to come from the missing limb. Other than this, every prosthetic limb has no feedback to the user and is totally open-loop.
Future prosthesis are expected to be able to interface directly with the motor neuron and sensory neuron remnants without needing them to be connected to other muscles, but as of yet it has been difficult to interface with nerves and keep them alive for longer than a few days. Connecting directly to the motor cortex of the brain is also promising, especially in cases of paraplegia where the spinal cord is damaged or severed, and experiments have been performed where monkeys have operated a robot arm in order to feed themselves (
independently of their natural arms), and to operate a pair of robotic legs remotely, keeping the legs walking on a treadmill whilst the monkey stood still.
OK, that was a lot more info dump than I was intending.
tl;dr: for current prosthetics, 'awesome' and 'practical' are mutual exclusives.
