The Safety and Reliability Issue of Robotic System with Artificial Intelligence - Term Paper Example

? The Safety and Reliability Issue of Robotic System with Artificial Intelligence Zihao Deng ENL213 Lisnyczyj Laura December 13, ABSTRACT This paper reviews the safety and reliability issues associated with Robotic System with Artificial Intelligence. The applications of robotic systems with artificial intelligence are growing day by day. Unlike human, robots can perform complicated works with utmost perfection. Moreover, robots can work continuously for longer periods without wear and tear. Even though the utility of using artificial intelligence in doing complicated work is unquestionable, there are many concerns about the safety and reliability of using it. Since artificial intelligence is dissimilar to human intelligence, it is incapable of taking independent decisions. Moreover, it is incapable to respond properly in unfamiliar situations. Artificial intelligence need to be advanced a lot before human can trust it for doing independent works. INTRODUCTION “Robots are physical agents that perform tasks by manipulating the physical world. They are equipped with sensors to perceive their environment and effectors to assert physical forces on it”1. The huge success of robots in doing extremely risky works with utmost precision has developed huge interests in the scientific world about using robotic systems for artificial intelligence. Even though human like behavior cannot be expected from robots, it is widely believed that robots can match human intelligence to a much closer level than once anticipated. It is a fact that robots need human supervision or intervention for safe functioning. However, current robots need less supervision from human elements and they are intelligent enough to make necessary changes in its functioning based on the changes in the environment. Robots are currently used for people to conduct remote controlled video conferencing, navigation, rescuing, assisting and lifting of soldiers in the war fronts etc. It is difficult to define artificial intelligence or AI in few words. Generally speaking "Artificial intelligence is the science of making machines do things that would require intelligence if done by men" 2. The history of artificial intelligence started along with the discovery of computers in by Charles Babbage. Even though the possibilities of using artificial intelligence or robots are infinite, still people have many concerns about the reliability of using robots for artificial intelligence purposes. After all, robots are machines and machines cannot think as human does. It should be noted that even well experienced individuals are struggling to take appropriate decisions in this ever changing world. Under such circumstances, the reliability of depending artificial intelligence for executing complex jobs which need logical and quick decision making is highly questionable. This paper analyses the Safety and Reliability issue of Robotic System with Artificial Intelligence. SAFETY AND RELIABILITY ISSUE OF ROBOTIC SYSTEM WITH ARTIFICIAL INTELLIGENCE Incredible speed in execution is the major advantage of robots controlled by artificial intelligence. Moreover, robots exhaust much slower than humans and therefore it can perform huge volumes of works compared to humans and that also within no time. In manufacturing units, robots can perform complicated works with utmost precision so that manpower requirements can be reduced considerably. Robots with artificial intelligence can handle million of calculations within seconds and therefore, it can perform preprogrammed jobs with ease. Because of the above ability a computer or robot can beat a human in a chess match. The development of mobile robots was driven by the desire to automate transportation in production processes and autonomous transport systems. The former lead to driver-less transport systems used on factory floors to move objects to different points in the production process in the late seventies. New forms of mobile robots have been constructed lately like insectoid robots with many legs modeled after examples nature gave us or autonomous robots for underwater usage1. It is difficult for human to operate in extreme weather conditions. In other words, in coldest or hottest regions, human has limitation in performing certain works. For example Polar Regions are the coldest regions on earth. Antarctic and artic explorations are currently going on and robots are used extensively in such explorations because of its ability to resist the challenges raised by extreme weather conditions. In under water explorations also robots are used as a secure equipment to collect data. Space exploration is another area in which robots controlled by artificial intelligence are reliable. In outer space and in some planets, oxygen and suitable weather conditions may not be available for human to directly conduct his explorations or experiments. In such cases, robots can be used as the safest equipment for collecting data. A signal from NASA's Mars Science Laboratory spacecraft, including the new Curiosity rover, was received by officials on the ground shortly after spacecraft separation. The spacecraft is flying free and headed for Mars after separation from the United Launch Alliance Atlas V rocket that started it on its journey to the Red Planet. Liftoff was on time at 10:02 a.m. EST from Space Launch Complex 41 on Cape Canaveral Air Force Station in Florida 3 At the time of writing this paper, television news reports show that NASA has sent a satellite called curiosity to Mars for exploring the possibilities of life in mars. It may reach Mars only after nine months and it is estimated that Curiosity can handle mars exploration for two years. It may land on a mountain in Mars. It should be noted that a robot in curiosity is controlling the mars exploration. It is programmed based on whatever the knowledge NASA has so far about Mars. It is quite possible that Curiosity or the robot in curiosity may face extremely unfamiliar situations in mars. Nobody knows, what will be the response of curiosity on such occasions. One of the major drawback of artificial intelligence is that “in trying to match human intelligence, artificial intelligence attempts to replicate nature's coded behaviors accumulated over millions of years” 4. It should be noted that nature has made lot of modifications in its behavior in order to transform uncivilized or unintelligent human into a civilized and intelligent human of present kind. Artificial intelligence can never work in that manner. In a human body, the DNA codes alone occupy cubic miles of pages. Apart from genetic codes, nature has assembled similar volumes of code in myriad subsystems” At the moment of conception, a single fertilized human egg contains information equivalent to about six billion chemical letters, which can be recorded in a thousand 500 page books. In a grown human body, with the DNA in each cell containing a sequence of over 3 billion chemical nucleotide bases, the total of those codes in the body would fill the Grand Canyon fifty times over with 500 page manuals. This was just the code in the nucleus of cells4. The architecture and functioning of human memory and intelligence is unknown to the scientific world even though there are some assumptions and hypotheses. Robots, controlled by artificial intelligence cannot think in terms of constructing such complex architecture of memories in its brain or central processing units. Moreover, natural modification of behavior based on the needs of the environment is impossible for the robots controlled by artificial intelligence. In many industries, robots are working under the supervision of humans. In majority of the Japanese car manufacturing plants, robots are helping human in executing sophisticated jobs at a rapid pace and thereby the productivity of the manufacturing firm increases considerably. However, in such a human-robot interaction scenario, safety, reliability and efficiency of the robots are debatable. According to Ding, “study of human behavior is indispensable when robots interact with humans. This requires a better understanding and modeling of the human motion”5. The reliability of robotic system with artificial intelligence is highly questionable which can be proved with the help of the following example. Suppose a person asks his robot to go out and purchase one ice-cream for him. He might have fed every instruction needed for the robot to go out, purchase the ice-cream, pay the correct price, collect the balance and come back to him using the safest routes. He might have even given instructions to the robot in how to travel across the roads when heavy traffic is going on. The robot may wait till the traffic gets cleared for crossing the road because of the instructions fed into its memory. Now consider heavy construction works were going on in a particular road at the time when the person was sending his robot for ice-cream purchase. Suppose the person failed to feed instructions to the robot about what to do when constructions works are going on. In such case, the robot may face an entirely new situation and it may try to cross the road through the construction areas. It may face many problems on its way and may not succeed in executing the mission given to it by its creator. In other words, on unfamiliar circumstances, robots controlled by artificial intelligence may fail. Human intelligence work on trial and error method. In other words, human normally experiment many things in order to arrive at right solutions or conclusions in unfamiliar situations. However, robots do not possess such abilities for the time being and therefore the reliability of robots controlled by artificial intelligence is debatable. Even though robots cannot think as effectively as humans, it can perform programmed works more cleverly than humans. In other words, the possibility of committing mistake because of lapse in concentration cannot be ruled out while human perform certain jobs. On the other hand, in the case of robots controlled by artificial intelligence, the possibility of mistake is zero if all the environmental variables do not change. As mentioned in the earlier example, a robot can repeat the purchasing of ice-cream millions of times, if nothing changed on its way to the shop and back home. At the same time, human can commit many mistakes while performing that job quite regularly. In other words, it is quite possible that human may cross the road without checking the traffic and cause accidents because of the lapses in concentration. On the other hand, robots will never face such concentration problems. This is because of the fact that robots have no emotions whereas humans have emotions. Emotions quite often spoil concentration. Artificial intelligence work on IF and THEN conditions or rules. A rule consists of two parts: condition (antecedent) part and conclusion (action, consequent) part, i.e.: IF (conditions) THEN (actions). Antecedent part of the rule describes the facts or conditions that must exist for the rule to fire. Consequent describes – the facts that will be established, or– the action that will be taken or conclusion that will be made 6. In other words, robots can function only based on certain rules and conditions. If the rules and conditions were satisfied, it will work perfectly. On the other hand, if the rules and conditions do not met, it will not function perfectly. It is necessary to perform on occasions in which the IF-THEN conditions do not materialize. For example, suppose a manufacturing unit catches fire. Workers will act quickly even if they failed to get any instructions from the top. On the other hand, robots may wait for the instructions from its controller in order to act positively. In short, Machines or robots do not have the ability to estimate the consequences of an action. It may wait for the instructions before taking any actions against or in favor of something. “Since their inception, computers have always acted on a "yes" or "no" basis. They simply have not been able to recognize "maybe." Even the most sophisticated computers, capable of performing millions of calculations per second, cannot distinguish between "slightly" or "very"”2. It should be noted that computers can recognize only the machine language even though human feed instructions to computers using different high level languages such as BASIC, PASCAL, COBOL, FORTRAN, JAVA, VISUAL BASIC etc. Computers will convert these high level languages into machine language before analyzing a particular instruction fed to it though these languages. Machine language is simply of a collection of zeros and ones. A logical one in machine language represents a YES condition whereas a logical zero represents a NO condition in machine language. There is no notations or digits for representing a “MAY BE” in machine language. In other words, there is no logic condition like .5 in between a 0 and 1 in machine language. Human can think in terms of many fractions in between 0 and one 1 such as ?, ?, 1/8 etc. However, computers or robots controlled by artificial intelligence do not have the ability to realize the differences between ? and ? or 1/8 and 1/6. “Many robot designers use a single-channel "safety rated" circuit for emergency stop. The circuit is normally closed; therefore, if a wire is accidentally cut, the robot will fail in the emergency stop condition” 7. It should be noted that all the functioning of robots depends on the reliability of the wires soldered in the robot circuits. Even if a single wire connection gets damaged, the entire functioning of the robot could be stopped. Since robots are working in extremely complicated environments, the chances of damaging the wire connections are extremely high. In other words, periodical repairing and maintenances are necessary for the safe operation of robots. In short, constant supervision of human intelligence is necessary for the safe and secure functioning of artificial intelligence. “A robot can remain in a position for a long period of time. Such a capability offers important civilian and military advantages such as surveillance, observation, search and rescue and even for entertainment and games”8. However, the ability to stay in a particular position for longer period may create problems also. For example, while using in war fronts, enemies can locate or trace the presence of robots easily because of the inability of the robots in identifying enemies and friends. Thus, enemies can seize the robots and decode the sensitive information stored in it. Thus robots can bring more harm than good in war fronts in certain occasions. Planar snake robots are not controllable when the viscous ground friction is isotropic, but that a snake robot becomes strongly accessible when the viscous ground friction is anisotropic. The analysis also shows that the snake robot does not satisfy sufficient conditions for small-time local controllability9. Robots are designed in different shapes for executing different types of jobs. Robots designed in the shape of snakes are called snake robots. Such robots have the ability to stay away from the eyes of enemies. However, such robots do not have the ability to move on all surfaces. It is capable of moving only on some particular types of surfaces which is definitely a disadvantage when used in exploration missions. For example, NASA has already sent a robot to the exploration of mars. Since NASA has not much knowledge about the surfaces in mars, the reliability of this robot in conducting the mars exploration mission is in doubt. A robot navigation in the presence of humans and other robots is rarely considered in the field of robotics. When navigating through a space filled with humans and robots with different functions, a robot should not only pay attention to obstacle avoidance and goal seeking, it should also take into account whether it interferes with other people or robots10. Since robots are controlled by artificial intelligence, it is incapable of segregating life forms with other objects. In other words, many robots are incapable of differentiating human from other objects. Under such circumstances, a person accidently comes in the vicinity of a robot could be treated as an object by the robot. For example, suppose a robot is engaged in the duty of recycling of certain scrap materials. Suppose a person accidently fell on top of the scrap materials. The robot may not stop its operation and it will recycle the person also along with other scrap materials. “For robots to be able coexist with people in future everyday human environments, they must be able to act in a safe, natural and comfortable way”11. In other words, robots need more intelligence to identify living and nonliving things. At present, it considers all living and nonliving things as objects. In short, artificial intelligence technology should be improved more to use robots safely. Robots should acquire the capability to identify the environment in which it operates and it should be able to make changes in its operations based on the needs of the environment. The supervision of human intelligence over artificial intelligence should be minimized as much as possible and for that purpose, artificial intelligences needed to be improved a lot. Robots are normally operated with the power from batteries. One of the major problems associated with the working of robots is the reliability of batteries. Robot batteries will be exhausted over a period of time and it should be recharged or replaced for the continuity of an operation. If robot batteries create problems midway through a complicated operation, the whole mission could be end up in big failure. In order to avoid that, robot batteries should use some kind of self-charging mechanisms. Solar cells can be considered as an option for solving the battery failure problems in robots. CONCLUSIONS To conclude, robots controlled by artificial intelligence are extremely useful in assisting human in conducting dangerous missions. Since robots are machines, it has no emotions or life. So it can concentrate fully on what it does at a particular movement and therefore the chances of committing errors are extremely low. Moreover, robots can perform works at a rapid pace so that it can be used in manufacturing units for increasing productivity and also to perform dangerous and sophisticated jobs. At the same time robots, controlled with artificial intelligence cannot perform any jobs when it face unfamiliar circumstances. The memory of robots cannot be compared with the memory of the human. When it comes to quick decision making in unfamiliar conditions, robots fail drastically and therefore its reliability can be ensured only for the programmed purpose alone. Robots can understand only YES or NO conditions and a MAY BE condition in between the YES or NO conditions is unfamiliar to robots. There are many instances in which human wanted to consider the MAY BE conditions also. Robots with artificial intelligence fail in such situations. Robots perform on conditional basis. It can respond based on IF and THEN clauses. There are many instances in which human wants robots to perform in conditions other than IF and THEN. In short, reliability and safety of using robots cannot be ensured beyond certain limits. Robots can never ensure complete safety and reliability to human beings as it is controlled by the human elements. However, robots can be used as reliable equipment for working in familiar conditions. Cited References 1. Niemueller T. and Widyadharma S. 2003. Artificial Intelligence – An Introduction to Robotics. http://www.niemueller.de/uni/roboticsintro/AI-Robotics.pdf. p.1-2. Accessed on 27 November 2011. 2. Lyons D. Artificial Intelligence Gets Real. Forbes, 30 November 1998. http://www.forbes.com/global/1998/1130/0118096a.html Accessed on 27 November 2011 3. NASA. 26 Nov 2011. MSL Spacecraft in Excellent Health”. NASA. 2011. http://www.nasa.gov/mission_pages/msl/index.html Accessed on 27 November 2011 4. Ebbi T. The Artificial Intelligence Disadvantage. Effective Mind Control. 2011. http://www.effective-mind-control.com/artificial-intelligence-disadvantage.html. Accessed on 27 November 2011. 5. Ding H. 27 November 2011. Control of Robotic Systems for Safe Interaction with Human Operators. Proceedings of the Twenty-Second International Joint Conference on Artificial Intelligence. http://ijcai.org/papers11/Papers/IJCAI11-468.pdf. p.2792. Accessed on 27 November 2011 6. Eric.H. Artificial Intelligence. 2004. Mississippi State University http://www.cse.msstate.edu/~hansen/classes/AIspring04/slides/expertsystem.pdf. Accessed on 27 November 2011 7. Anderson C. Robot Safety Faces Deadlines. Welding Design & Fabrication; Aug 2001, Vol. 74 Issue 8, p.22. 8. Sintov A. Avramovich T. and Shapiro A. Design and motion planning of an autonomous climbing robot with claws. Robotics & Autonomous Systems; Nov2011, Vol. 59 Issue 11, p1008-1019 9. Liljeback P. Pettersen K Y. Stavdahl, O. and Gravdahl, J. T. Controllability and Stability Analysis of Planar Snake Robot Locomotion. IEEE Transactions on Automatic Control; Jun 2011, Vol. 56 Issue 6, p1365-1380. 10. Lam C. P. Chou C.T. Chiang K.H. & Fu L.C. Human-Centered Robot Navigation—Towards a Harmoniously Human–Robot Coexisting Environment.IEEE Transactions on Robotics, 02/01/2011, Vol. 27 Issue 1, p99-112. 11. Svenstrup M. Hansen S. Tranberg A. Hans J. and Bak T. Adaptive Human-Aware Robot Navigation in Close Proximity to Humans. International Journal of Advanced Robotic Systems, Jun 2011, Vol. 8 Issue 2, p7-21 Read More