Cooperative planning for heterogeneous teams in rescue operations

On the performance of self-organizing maps for the non-Euclidean Traveling Salesman Problem in the polygonal domain

“…The problem is to find a path for a mobile robot so that the robot will ''see'' the whole working space. The practical motivation of the problem is a search and rescue mission in which possible victims need to be found quickly [25]. The problem can be formulated as the TSP, i.e., a problem to find a path that visits the given set of sensing locations, where measurements of the robot's vicinity are taken.…”

Sensing Locations Positioning for Multi-robot Inspection Planning

“…Our approach extends Somhom's algorithm described in [6]. The idea of the algorithm is to represent a path of each particular robot by a chain of neurons, where neighbouring neurons are connected [4]. At each iteration, a nearest neuron to a randomly selected guard is determined and together with its' neighbours moved closer to the guard.…”

“…C pi is the i-th city in a permutation, for which the nearest neuron to the C pi is determined. To select the nearest neuron, the guard-to-neuron distance is defined as their Euclidean geodesic distance weighted by: weight(r) = ((length(r) − AV G)/AV G) 4 . This suppresses those neurons, which overshoot AV G and prefers those, which are bellow AV G, where AV G stands for the average length of chains in the task.…”

“…Using multiple robots can reduce total execution time, but on the other hand, it needs novel algorithms for coordinating robots' movements. Typical application of the inspection task is a search and rescue mission in case of emergencies or catastrophes [4], where the goal is to find (and rescue) insured people or other objects of interest (fire alarms, seats of fire, places, where dangerous materials are placed, etc) as fast as possible .…”