"Groupbased exploration and mapping with sensoric under-equipped robots"
Groupbased exploration and mapping with sensoric under-equipped robots
Sprache des Titels:
Most approaches to cooperation in the field of cooperative robotic are based on dividing the tasks between different robots to decrease the needed time. Still, one robot could do the work on it´s own, as it has all the needed functionality implemented. Inspired by the function of an ant-hill the concept of robots who are not able to work alone was tested with specially desiged robots and scaled up to the use of dozends within a simulator. This simulator had to get as close to the real life robots as possible to get valid data for statistic analysis. In this thesis three robots where built to work together and give real life information to test and refine the simulated environment. The scalability and statistical analysis was done with a large amout of simulated robots as the construction and testing would have been to much effort.The basis of the cooperation was mapping of an unknown environment with triangulation. The robots scan their surroundings by marking a spot with a laser pointer by one robot and scanning the angle of this point in relation to all the other robots. This forces at least two robots to work together find one point of the environment. The results of this thesis show that the approach is valid and distributed functionality on many robots has many advantages like smaller, cheaper robot configurations, system stability in case of robot failure and specialized robots for certain tasks.