Koordinacija više robota za učinkovite pretraživanje prostora

This paper addresses the problem of exploration of an unknown environment by developing effective exploration strategies for a team of mobile robots equipped with continuously rotating 3D scanners. The main aim of the new strategies is to reduce the exploration time of unknown environment. Unlike most of other published works, to save time, the laser scanners rotate and scan the environment while robots are in motion. Furthermore, the new strategies are able to explore large outdoor environments as a considerable reduction of the required computations, especially those required for path planning, have been achieved. Moreover, another new exploration strategy has been developed so that robots continuously replan the order to visit the remaining unexplored areas according to the new data (i.e. updated map) collected by the robot in question or by the other team members. This new extension led to further enhancements over the above mentioned ones, but with slightly higher computational costs. Finally, to assess our new exploration strategies with different levels of environment complexity, new set of experiments were conducted in environments where obstacles are distributed according to the Hilbert curve. The results of these experiments show the effectiveness of the proposed technique to effectively distribute the robots over the environment. More importantly, we show how the optimal number of robots is related to the environment complexity.Ovaj članak istražuje problem pretraživanja nepoznatog prostora razvijanjem učinkovite strategije za tim mobilnih robota s rotirajućim 3D laserskim senzorom. Glavni cilj ove nove strategije je smanjenje vremena pretraživanja nepoznatog prostora. Za razliku od većine objavljenih radova, u ovome članku, radi smanjenja vremena, laserski senzori rotiraju i snimaju prostor dok su roboti još u pokretu. Predložene strategije, pošto se njima znatno smanjuje računska složenost, pogotovo za planiranje gibanja, omogućuju pretraživanje i vanjskih prostora prostora velikih dimenzija. Nadalje, razvijena je još jedna strategija pretraživanja koja omogućuje robotima da kontinuirano replaniraju poredak kojim će posjetiti ostatak neistraženog prostora, prema novim podacima (ažuriranoj karti) prikupljenim od njih samih ili drugih članova tima. Ovo novo proširenje nadalje unaprjeđuje performanse algoritma, ali uz nešto veću računsku složenost. Kako bi se u konačnici testirale nove strategije pretraživanja na prostorima različite složenosti, provedeni su eksperimenti s preprekama raspoređenim po Hilbertovoj krivulji. Rezultati eksperimenata pokazuju učinkovitost predložene metode u prostornom raspoređivanju robota. Od posebne je važnosti istaknuti da se u članku također istražuje odnos između broja robota i kompleksnosti prostora