Robotics

---

PROLAB Technologies Research and Development effort is focused on Robotics and Artificial Intelligence; with the intention of both providing a platform for us to develop products and product enhancements in these fields, and as a tool for study and education, for all ages.

It is clear that both Robotics and AI are on the edge of a breakthrough - on the engineering side robots are becoming far more flexible, smaller and more capable of movement and adaption; on the AI front moderately capable AI bots are spreading fast and the major tech companies are making significant investment in areas like driverless cars, where fast, adaptive thinking machines will be key to success and safety.

Challenges & Questions in Robotics and AI

Right now there are a number of challenges and questions within the field of robotics; below we offer a brief run through of a few current issues:

Energy

Probably the biggest challenge to the concept of fully autonomous anthropoid (or biomorphic) robots is simply one of power - how to create motor systems and processors so efficient, and power units with high enough capacity, that autonomous machines can run for days (or weeks) without needing to shutdown and recharge. In Sci-fi this is achieved by hand-wavium, unobtainium, or miniturized nuclear fuel cells - sadly none of these look likely to make it into the designs of any current robot concepts. Wave, wind and solar power become options for non-humanoid robot constructs and for robots operating in non-domestic environments, and continual incremental improvements in the power management of devices, sensors and actuators assist; but we are a long way away from having anthropoid human-like robots walking and talking with the total independence of, say, C3PO, Ava or the T1. We will need some radical changes in power capture, storage and consumption for this to become reality.

Motorisation

Mobility design trends in robotics are moving away from wheeled and tracked systems to systems that behave more like humans or animals, the requirement for faster and more sophisticated motors, actuators and sensors grows. Complex motorised designs like biomechanical legs, DARPA Robotics Challengers, and biomorphic spider- or bird-shaped designs, all require smaller, faster, actuators and complex programming; or alternative methods of propulsion based on actual biological systems. Systems like MIT's Cheetah show that it is possible to create dynamic forms that can achieve speeds and stability that are close to those achieved in nature, but there is still some way to go to creating safe robots that can inter-operate with humans and deal with the surprises and stresses that a real world environment can throw at them.

Advancing AI

The premise of the movie Ex Machina is an interesting one - what is the real Turing test that indicates the existence of true Artificial Intelligence? Is it successfully convincing a human that he or she is talking to another human, and not a machine? Is it winning in complex games? A good enough smart computer program, with enough inputs and data can probably just about pass the Turing test - even Amazon's Alexa gets pretty close at times - but the real test of a self-aware AI system is really whether the system is so self-aware that it can deliberately set out to convince a human that it is something it is not, or deliberately set out to influence the human tester into believing or wanting something that the human didn't start off desiring - in other words, a true self aware AI has to have the ability to lie. If Ava convinces you that she needs to be loved, and makes you fall in love with her, even when she doesn't really care about you at all, then she passes the test.