HOW many times have you waited for an elevator to make that blessed halt at where you are and found it too crammed even for a fly? Well, if Bruce Powell, a researcher at the Otis Elevators Inc has his way, those long, irritating waits for an elevator may finally end.

Powell, a software engineer, works in the research and development wing at the headquarters of the world's largest elevator company at Farmington, Connecticut, the US.

The average elevator of today is pretty dumb, responding to calls at random. It can neither analyse traffic patterns nor figure out where to stop first when called simultaneously to several floors. Elevator A, say, is on the fourth floor, nearly empty and rising. Elevator B, quite packed, is on the eighth floor and rising. A few passengers are waiting on the 10th floor. Conventionally, elevator B, being nearer would respond to the call and some of the passengers won't be able to squeeze in, though the best decision would have been to dispatch elevator A. Powell and his colleagues are using the so called 'fuzzy logic' to assign priorities to factors as how close an elevator is to the caller vis-a-vis how crowded it is. Also, the researchers are hoping to build a 'memory' software for the elevator.

It would then recognise a rush-hour (say, the lunch period, when more crowds head towards the cafeteria at a certain floor) and automatically an elevator, whenever free, would be dispatched to that particular floor. Moreover, incorporated with the fuzzy-logic software, the elevator would soon recognise the traffic pattern on a pleasant day when office-goers would rather go out and eat. The fuzzy logic software module is based on the science of neural network - a sophisticated computer system modelled roughly on the wiring and functions of the human brain. Powell is also researching over more futuristic fixes: like installing scanners with memory in an elevator. The elevator would then 'recognise' the regular users and 'memorise' where they get down, and stop accordingly.

Otis researchers sit hunchbacked over their computers in a 28-storey test tower to analyse a test elevator's speed, acceleration and retardation, possible electromagnetic interferences from the antennae of the building and other computers. Powell's ideas have met with initial success in the 48-storey TransAmerica Building in San Francisco, California, where the average Iwaiting time' for elevators at certain busy floors have been cut down to 21 seconds from 48 seconds.

His team is also working on a project to develop elevators that glide sideways to move out of the way of another 'busier' elevator. An office building would, then, have a single 'hoistway' and several elevators can be stacked one on top of the otber. This would save the trouble of building an individual chute for every elevator. The 'hoistway' would have bifurcations sideways - and the elevator would glide either on rails or be cushioned by magnetic suspension. "Technically it's very possible," asserts Powell, "...and economical too". Conventional elevators are supported by heavy steel cables that raise and lower the cab. They must be eliminated for an elevator capable of moving laterally. Toward that aim, Otis researchers have developed a prototype electric 'linear motor' along a single track, raising and lowering the cab without the cables.