Stephen’s Law of Robotics

In the previous article the notion that over a period of time the number of capabilities a robot possesses will double was introduced. While some of these capabilities will likely come from the artificial intelligence field, just as many will come from plain old engineering. The growth of capabilities will come about due to problems to be solved, and new tasks to be performed. There are many motivations driving the growth of capabilities in robotics.

Power of incentive

The competitions inspired by DARPA to develop self-driving vehicles show how quickly these capabilities can come together when there is enough motivation. Cars that could barely drive in a straight line several years ago are now able to drive straight, around curves, through traffic, while maintaining safe distances and traveling at speed. These competitions show that initial failure leads to failure analysis, finding areas to strengthen, and areas that need new or completely different solutions. In particular, the ability to deal with unreliable input has improved substantially. The race showed that many areas, such as identifying when sensors fail, still need to improve. In one case, an acceleration sensor showed no acceleration, so the car just kept accelerating. Cases of memory leaks (allocating memory without ever freeing it) took out at least one robot during the competition. While these problems kept the some vehicles from completing the competition, one can be sure that the failures encountered will be fixed. In the formative years there will be failure after failure, often due to unexpected events, or unrecognized interactions between different parts of the system. But with each failure, there will be new solutions, and steps will be made towards greater robustness and generalization.

Cash prizes for robotics include DARPA’s automated vehicle challenges, and Google’s robotic lunar lander challenge. There are also various prizes designed to motivate students such as underwater, ground, and aerial competitions. Robot soccer is gaining serious momentum, with a challenge to beat humans by 2050. The robot soccer players even have their own federation. There is a competition to build robots to fight fires. There are likely many more incentive and academic challenges out there, and if you know of any please leave a comment.

Algorithm improvements

In computer science, algorithms such as sorting, searching, indexing, are very common. These are relatively simple algorithms compared to those required for balancing, walking, running, driving, and navigating. In the early days of computer science, very simple algorithms were replaced by faster versions, then specialized versions, and substantial analysis determined which algorithms were appropriate for which circumstances. That is strengths and weaknesses were formally defined. This same level of analysis is ongoing today for the algorithms, and heuristics required to make a robot successfully complete a task or sub-task. The challenge is always greatest the first time a particular problem is solved, but once one solution is known, gradual improvement and alternate solutions soon follow. Science and engineering continue to grow by building off of work which came before. Robotics is no different in this aspect.

Economic need

Another driving force that will push the development of robotics is economics. For the Japanese the cost of assisted care for the elderly is driving development. With a declining birth rate, and growing elderly population, there will not be enough professional care takers, which is driving the development of care taker robots. The need to cut costs and reduce exposing humans to dangers has also been pushing advances in robotics. The cost of safety is generally high, and robotics is seen as one way of bringing this cost down. Economic competition with low wage countries is also driving robot development to make higher wage countries more competitive.

Coolness factor

Another motivation for developing robots is the coolness factor. Many of us have grown up seeing robots in science fiction shows and stories. From ‘Danger Will Robinson‘ to Asimov’s three law bearing robots, to Star Trek’s Data, robots inspire the imagination. Working on robotics provides the sense that one is working on something new and exotic.

Finding your niche

Now that robot development kits are becoming more common with Lego’s Mindstorm and Microsoft’s Robotics Studio it is becoming easier to find frameworks for playing around with robot development. Just as computers benefited from small groups coming together to develop new computers and software programs, the same sense of discovery will appear for robots as well. While these kits do not seem to have captured as broad interest as the original Commodore 64, Apple II, or TRS-80, it is likely only a matter of time before the robotics hobbyists kick in to gear. Just as billionaires were born from the software industry, there are sure to be new billionaires born from the robotics industry.

Helping others

Another common motivator in robotics is the desire to help others. Research in health care oriented robots, car safety, keeping people away from bombs, and other dangerous situations have drawn many people to research and develop robots. Also, the desire to avoid monotonous tasks such as pool cleaning and floor vacuuming have already seen robots enter the market, and improvements in these cleaning robots show a gradual increase in robustness and capabilities.

Moving forward

Whether it is curiosity, novelty, altruism, national pride, money, or ego; there are many personal motivators to draw researchers and engineers toward the field of robotics. Just as the early days of computers produced many notable personalities and products, the field of robotics is a new frontier opening up to a larger group of developers out to make their mark. This will drive rapid and diverse experimentation into a variety of hardware and software architectures. The metric by which this advance will be measured is the growth in the set of capabilities a robot possesses. This growth will be displayed by a doubling in the number of capabilities over a period of time: Stephen’s Law of Robotics. This will lead to an ever greater range of tasks which can be accomplished, at a progressively lower cost.

Related Links

Moore’s Law of Robotics

Be Sociable, Share!
Explore posts in the same categories: Almost there