So if it wasn’t bad enough that humans continue to threaten the peace and stability of the world, it now appears as if the robots also want to demonstrate their prowess in the time honoured tradition of killing - or so you would believe based upon the internet hype around the Russian Robot, FEDOR
(Final Experimental Demonstration Object Research).
Well, fear not humans! Despite Russia’s Deputy Prime Minister's enthusiasm for this nascent Terminators ability to fire a 9mm pistol while honing its fine motor skills and decision making algorithms
, FEDOR has a long way to go before it can be considered a threat to humans maintaining the monopoly on violence. Yes, unlike its imaginative human creators who have excelled in the art of war, this invention comes replete with a range of technical obstacles of which the internet hype should consider and, instead, understand that it was built for an entirely different purpose altogether.
First, some observations. See the tether?
FEDOR is unable to achieve enough on board power to provide it with locomotion or manipulation, and this constraint may also extend to its on-board processing. To achieve a suitable level of machine intelligence to operate, FEDOR requires some significant inputs for processing and power, which means that it is going to be bound for some time to that tether if this system is to perform the broad range of functions and behaviours that are being advertised.
Secondly, locomotion and dynamic balance are lacking. This is a critical development milestone, where achieving mobility is necessary for manoeuvre to perform tasks in the natural environment. Videos of its movement on flat surfaces
demonstrates an inefficient and unpredictable gait where it requires substantial support to progress along the laboratory test track. It also appears that FEDOR is synthesising human movement captured in its software without accurately analysing the terrain which it encounters and applying the required gait. This is much more noticeable when you compare it to an advanced humanoid such as Boston Dynamics Atlas V2
operating in an unpredictable field environment. Hence, FEDOR is unable to traverse a real human environment let alone satisfactorily perform in a lab environment.
Thirdly, the video imagery is significantly edited. It does not show FEDOR picking up the pistol, entering or exiting vehicles, crouching down to crawl, or maneuvering without human assistance. In short, the shots are staged, probably due to the fact that they are struggling with one of the more complex areas of robotics related to mobile manipulation tasks. In fact, all of the videos I could find of FEDOR are short edited fragments where the robot is already in the required configuration, including in this instance with pistols already in hand, surely calibrated to the initial target set prior to performing its firing sequence. In addition, in many of these tasks, the humanoid is likely being remotely operated.
These are just some of the simple technical limitations to be considered when assessing these capabilities. However, in outlining these limitations, I am not discounting the larger impact that the Robotics industry will have on human society in the next decades. Surely it will be profound, though unlike many other observers I do not believe that it will be adversarial, or result in the rise of the machines to make man their slaves.
Like any tool developed throughout our evolution, we will wield it and improve our ability to complete tasks or functions more efficiently and effectively. This was the purpose for FEDOR’s development - to interact with human tools and environments, hence its proposed deployment to the international space station and conceptual deployment to the moon to establish a base. Before these deployments can be considered, however, there are significant obstacles to overcome if it is to perform the most basic of human tasks, let alone pose an autonomous threat to humans.