• ino ino Ads by Toktok9ja
  • Balvinciglobal Balvinciglobal
  • Murya-Magazine-Banner Murya-Magazine-Banner

TOKTOK9JA MEDIA

Gossip-Entertainment-News-Music-Videos

Sci­en­tists devel­op robots that can adapt like ani­mals and recov­er from dam­age in less than TWO MINUTES

While it’s a com­mon fear that robots will one day rise up and turn against us, there’s cur­rent­ly a major obsta­cle stand­ing in the way — they’re too frag­ile.

New research is hop­ing to make robots more resilient by equip­ping them with spe­cial soft­ware that can help them learn how to bounce back from an injury in two min­utes or less.
The hope is that these learn­ing algo­rithms will help pro­duce more effec­tive autonomous robots that require less human inter­ven­tion and can last longer in crit­i­cal sit­u­a­tions like the work­place or search and res­cue sce­nar­ios.
A team of sci­en­tists equipped robots with spe­cial soft­ware that can help them learn how to bounce back from injuries in two min­utes or less, by tak­ing a page from real ani­mals.
For the study, sci­en­tists from the Pierre and Marie Curie Uni­ver­si­ty and the Uni­ver­si­ty of Wyoming, took a page from real ani­mals. 
When an ani­mal is injured, they’re able to com­pen­sate by limp­ing, shift­ing their weight or some oth­er strat­e­gy. 
Many three-legged dogs can catch Fris­bee, for exam­ple, or if some­one sprains their ankle, they can still fig­ure out how to walk even with an injury. 
‘When injured, ani­mals do not start learn­ing from scratch,’ Jean-Bap­tiste Mouret, a co-author of the study, said in a state­ment. 
‘Instead, they have intu­itions about dif­fer­ent ways to behave.
‘These intu­itions allow them to intel­li­gent­ly select a few, dif­fer­ent behav­iors to try out, and after these tests, they choose one that works in spite of the injury.   
‘We made robots that can do the same,’ he explained. 
Before the robot is deployed, it uses a nov­el algo­rithm to cre­ate a detailed map of the space. This allows it to devel­op cer­tain intu­itions about what behav­iors it can per­form and their val­ue
Before the robot is deployed, it uses a nov­el algo­rithm to cre­ate a detailed map of the space. 
Accord­ing to the researchers, this map rep­re­sents the robot­’s ‘intu­itions’ about what behav­iors it can per­form and their cor­re­spond­ing val­ue. 
Essen­tial­ly, the robot can build a library of dif­fer­ent motions and estab­lish which body parts it can rely on if it becomes injured, even if it has a bro­ken or miss­ing leg. 
Requir­ing a robot to map all these sce­nar­ios would take too long and could poten­tial­ly dam­age the device, so the sci­en­tists mapped them out in a com­put­er sim­u­la­tion. 
In doing so, they were able to test and map over 13,000 dif­fer­ent ways of walk­ing, includ­ing with ‘dam­aged, bro­ken and miss­ing legs, and for a robot­ic arm with joints bro­ken in 14 dif­fer­ent ways’, accord­ing to the study.
The sci­en­tists call this process an ‘intel­li­gent tri­al and error algo­rithm,’ ulti­mate­ly enabling the robots to adapt to sit­u­a­tions in two min­utes or less. 
With this study, the researchers are hop­ing that it will give robots greater longevi­ty in dan­ger­ous or time sen­si­tive sit­u­a­tions, i.e. res­cu­ing peo­ple from for­est fires or earth­quakes
After it becomes dam­aged, the robot acts like a ‘sci­en­tist’, try­ing out dif­fer­ent behav­iors and rul­ing out those that don’t work.  
‘We have robots store knowl­edge from pre­vi­ous expe­ri­ence in the form of a map of the behav­ior-per­for­mance space,’ the researchers not­ed. 
‘Guid­ed by this map, a dam­aged robot tries dif­fer­ent types of behav­iors that are pre­dict­ed to per­form well and, as tests are con­duct­ed, updates its esti­mates of the per­for­mance of those types of behav­iors’
‘The process ends when the robot pre­dicts that the most effec­tive behav­ior has already been dis­cov­ered. The result is a robot that quick­ly dis­cov­ers a way to com­pen­sate for dam­age with­out a detailed mech­a­nis­tic under­stand­ing of its cause, as occurs with ani­mals,’ they con­clud­ed. 
In the past, sci­en­tists have been able to devel­op self-diag­nos­ing robots, but they used expen­sive sen­sors that often took a while to devel­op a con­tin­gency plan (such as walk­ing with a limp).
COULD THIS TINY ROBOT SOME DAY BE USED TO DELIVER DRUGS INSIDE OUR BODIES?

Sci­en­tists at Ger­many’s Max Planck Insti­tute for Intel­li­gent Sys­tems have devel­oped a tiny robot that might be able to be used for med­i­c­i­nal pur­pos­es.
The robot is about a sev­enth of an inch long, mak­ing it small­er than a pen­ny.
Called a ‘mil­liro­bot,’ it can walk, crawl and jump on land, as well as swim in water. 
Sci­en­tists mod­eled it after crea­tures like jel­ly­fish, cater­pil­lars and lar­va.
With­out any legs, the robot is made of a soft, pli­able rub­ber that’s embed­ded with mag­nets
The robot is con­trolled using a mag­net­ic field
The team hopes that the soft robot can be insert­ed in human bod­ies for tar­get­ed drug deliv­ery 
It may also be used to improve min­i­mal­ly inva­sive sur­gi­cal pro­ce­dures.
The sci­en­tists could add a pock­et to the tiny robot that can house drugs.
With this study, the researchers are hop­ing that it will give robots greater longevi­ty in dan­ger­ous or time sen­si­tive sit­u­a­tions. 
For exam­ple, self-repair­ing robots would be able to wade into sit­u­a­tions deemed too dan­ger­ous for humans, such as for­est fires, retriev­ing trapped vic­tims or shut­ting down a nuclear plant.  
‘It could enable the cre­ation of robots that can help res­cuers with­out requir­ing their con­tin­u­ous atten­tion,’ Danesh Tara­pore, a co-author of the study, said in a state­ment.   
‘It also makes eas­i­er the cre­ation of per­son­al robot­ic assis­tants that can con­tin­ue to be help­ful even when a part is bro­ken.’    
Do you have a hot sto­ry or scan­dal you would like us to pub­lish, please reach us through toktok9ja@gmail.com or What­sApp: +2348033498347

One thought on “Sci­en­tists devel­op robots that can adapt like ani­mals and recov­er from dam­age in less than TWO MINUTES

Leave a Reply

Your email address will not be published. Required fields are marked *