Welcome to my home-page
Alfred Schurmann


In contrast to the very specialized, only little intelligent and partly autonomous robots used now, the intelligent full-autonomous and motivated robot (called below in short: full-autonomous robot) can do many different and complex activities in its surrounding. It has not only models of objects and situations appearing in its surrounding, but also models of activities (called activity schemas) which it can do in this surrounding. This has important practical implications - in many cases, we do not need 3 to 10 specialized robots which do several different activities, because one full-autonomous robot could do these activities. The following examples show this.

Full-autonomous assistance robot for the elderly
The surrounding of this full-autonomous assistance robot is a home where are one or more elderly or disabled persons who need help. The full-autonomous assistance robot would have (a) models of objects (e.g the rooms are objects), with their localizations, and possible situations in the home, and (b) models of activities which it can do in the home e.g. getting a glass from the kitchen, cook dinner on the electric stove, operate washing machine and vacuum cleaner e.t.c. Some so-called "ambient-intelligence-components" could also be applied in this home.
The application of the full-autonomous assistance robot would have the following advantages, compared with the "ambient-intelligence-home" (or smart house):
i)       specialized robots, smart appliances and other equipment (e.g. cameras, sensors and communication net) would be reduced at least by 80%,
ii)      a computer for controlling the remaining smart components is not necessary (the full-autonomous robot would control them),
iii)      the full autonomous robot would be able to adapt easily to changes in a house (e.g. to operate new washing machine),
iv)      in contrast to the full-autonomous robot, "intelligent components" do only few moving activities, which restrict essentially the use of such components, e.g. an "intelligent cooking pot" which says when and what I should put in the pot would me help very little, when I must put in the food myself (cooking recipes fulfill also this "intelligent" function of this pot,
v)      the full-autonomous robot would do similar activities as a care helper and would be perceived as a substitute for such service person - the elderly would not need to learn new behavior patterns,
vi)      a home in which are embedded many smart devices and specialized robots would be strange and frustrating for elder persons, because they are not used to such new surrounding and ought have much time and great motivation to adapt them-self to the very technical new home,
vii)     the full-autonomous robot could communicate with people using simple closes of a natural language.

Full-autonomous robot for domestic works
When said assistance robot for the elderly would be build, then it would be relative easy to modif'y its software, so that it could do domestic works (as housemaid helper) in a normal home (also taking care of children elder than 5 years).

Full-autonomous robot used as assistance worker
When said care or domestic robot would be built, we could modify its software so that it could be used (a) as assistance worker of a skilled worker in manufacturers, and (b) for full-autonomous operation of machines, specialized robots and devices. Application (b) would have great future because it leads to manufacturers where all work would be done only by machines, specialized robots and full-autonomous robots, where the last would operate these machines and specialized robots, and control the whole production process.
This solution is cheaper, more reliable and adaptable than a manufacture with machines, specialized robots and many cameras and sensors networked with computers which control them and the whole production process, because:
i)      if the full autonomous robots are used, then 70% to 80% of cameras, sensors, communication net and control computers would be reduced,
ii)     full-autonomous robots would not only control and steer the production process, but also better recognize errors and irregularities, and repair them if they are not serious,
iii)    whole sections of a manufacture could produce without people, where some full-autonomous robots would communicate  all important situations to the control person and this person would sent instructions to some full-autonomous robots.
How difficult is to build the full-autonomous robots?
In http://www.home-a-schurmann.de/makerob.html (or http://mywebpage.netscape.com/alschurm1/page10.html) is described the idea how to build the full-autonomous robot. It is much easier to build said assistance robots than to build a reliable and useful system for driving a car, because e.g.:
i)     the number of situations the driver should perceive and react is much more greater than the number of situations in a house,
ii)     at most 98% of traffic situations could be recognized by the driving system, but at least 2% of important situations (some of them may be very dangerous) would not be recognized by the driving system, and this would lead to serious road accidents,
iii)     even in some recognized situations (e.g. an appearance of three important traffic situations at the same time) the driving system would not be able to react properly in 1-2 sec, and the human driver would not have time to take-over the driving, recognize the current traffic situations and react properly (if the human driver should recognize driving situations and be able to react properly simultaneous with the driving system, then the human driver does not need such assistance system).
Despite this great difficulties, some computer scientists are developing such driving assistance system, and some car companies finance this.
Difficulties like (ii) and (iii) do not occur when the full-autonomous assistance robot for elderly would be built and used.

Some robot experts say that the robot doing household chores (e.g. cooking dinner and washing dishes) could not be built in the next 20 years. This opinion follows from (a) to great focusing of rather simple specialized robots, (b) the belief that modeling of the world is not realistic and not necessary, and (c) the opinion that a robot doing many different activities must have goals and must plan its activities. These experts have in mind a robot which has abilities and intelligence of a human. Such robot indeed could not be built in the next 20 years. In my opinion we need not a human like robot, because it would not be reliable.
The full-autonomous robot is not human like - it does not model the world but only surroundings which are in a house or a manufacture; it can however do many activities of a human in said surroundings, although it has not goals and do not plan activities. Said full-autonomous robot for house works could be built in 2 - 4 years.

Where should not or cannot be applied the full-autonomous robot?
It is not reasonable or possible to apply full-autonomous robots every-where. If a specialized robot do well its job (e.g. a robot which keep eye on rooms in a bank, robots which clean sewers and specialized mini-robots), then it should not be replaced by the full-autonomous robot.
For reasons given in "How difficult is to build....", I do not see how a reliable full-autonomous robot for driving a car could be built.
Also the use of full-autonomous robot in armed forces would not be reasonable or possible, because:
i)    full-autonomous robots could not substitute fighting soldiers,
ii)    some simple military operations can do specialized robots.

Dr. Alfred Schurmann                                                                                                  Sept. 11, 2007

E-mail: alfschurman@yahoo.de
goto startsite