THE quickest way to find people buried in an avalanche is to make sure no
single person is in charge of the rescue team. At least, that鈥檚 the view of
researchers at the Sandia National Laboratories in Albuquerque, New Mexico, who
say they鈥檝e worked out how to locate trapped skiers at least four times as fast
as conventional means, by adopting a swarm-like search strategy.
Many off-piste skiers now carry radio beacons to help locate them in an
accident. Traditionally, teams sweep the area in a grid pattern in an attempt to
home in on their signals. But precious time can be wasted pinpointing them, says
Blyth Wright of the Scottish Avalanche Information Service in Aviemore.
By opting to do without a leader, says Sandia computer scientist Rush
Robinett, search parties can adopt a 鈥渄istributed intelligence鈥� strategy that
makes better use of their resources. In his scheme, each member follows an
algorithm that Robinett originally developed for tiny robots sniffing out
sources of chemical or biological attacks. The problem of finding buried skiers,
he says, is similar.
Advertisement
Robinett鈥檚 algorithm is still under wraps as his patent has yet to be
granted, but each rescuer would be equipped with a Global Positioning System
receiver and a palmtop computer that monitors the position of each rescuer and
the strength of the signal each is receiving from a buried victim. The algorithm
then works out where each searcher should go next, having taken into account the
strength of the signals received by all the others. In this way, no single
person鈥攐r computer鈥攊s leading the search and they quickly converge
on the buried skier.
Wright says the technology could be promising if it can be incorporated into
user-friendly compact personal transceivers. But, he says, snow rescuers might
find it hard to beat a good search dog: 鈥淭he thousands of dollars spent
developing this could buy an awful lot of Bonio,鈥� he says.