In the NeurIPS 2018 Artificial Intelligence for Prosthetics challenge, participants were tasked with building a controller for a musculoskeletal model to match a given time-varying velocity vector. Many solutions use similar relaxations and heuristics, such as reward shaping, frame skipping, discretization of the action space, symmetry, and policy blending. However, each team implemented different modifications of the known algorithms by, for example, dividing the task into subtasks, learning low-level control, or incorporating expert knowledge and using imitation learning.

Technological integration of AI and ML in the Prosthetic and Orthotic industry and the field of assistive technology has become a boon for Persons with Disabilities. The leading manufacturers of rehabilitation aids have used the concept of neural networks to simulate various anatomical and biomechanical functions of the lost part of the human body. The involvement of human interaction with multiple agents,' i.e., electronic circuitry, software, robotics, etc., has revolutionized the rehabilitation field to develop devices like Bionic legs, mind or thought control prostheses and exoskeletons. The application of AI and robotics technology significantly impacts achieving independent mobility and enhances the quality of life in Persons with Disabilities (PwDs).

Meet Atom Limbs

Atom Limbs is a company in California seeking to take the technology to the next level. The company leverages advanced sensors and machine learning - where computers train themselves to become more accurate - to interpret electrical signals from a person's brain and use them to move and manipulate a prosthetic limb.

The arm is connected with a full range of human motion in the elbow, wrist, and individual fingers- and it provides feedback to the wearer on their grip strength. The strengthened sportswear-style vest through which the arm is attached distributes the arm's weight evenly. Although it still has some weight to it, it is considerably lighter.

Free of surgery

The patient will not need any surgery or implants to function. It first connects to the wearer's residual limb with bands of sensors that measure electrical signals, then via a cup that fits over the top, with the arm connecting via an interface.

According to BBC's Paul Carter, who has personally tried out the prosthetics, "Despite avoiding upper-arm prosthetics before, when Atom Limbs said I could have a try at operating a digital version of the arm on a computer screen, via their control software, I was interested enough to say yes".

High cost

One of the significant limitations of these novel technologies coming to place is their cost and affordability for people. The assistive devices landscape is littered with products that, while impressive, can cost many times an average yearly salary. 

According to Ian Ada, a lecturer in prosthetics and orthotics at the University of Dervy, while this may sound like a lot, it is a reasonable price in the industry - though it won't be for everyone.

Also, there are ethical and practical issues around such products. Some criticize that companies are focusing too much on innovation rather than application. However, developments like these are shedding light on a hopeful future for at least some.