TY - GEN
T1 - A moment measure model of landmarks for local homing navigation
AU - Lee, Changmin
AU - Kim, Dae Eun
PY - 2016/1/1
Y1 - 2016/1/1
N2 - Visual navigation in robotics is one of the challenging issues, and many navigation approaches are based on localization of a mobile robot in the environment. The snapshot model is a biologically inspired model of insect behaviour to return home and it shows a simple algorithm to compare the snapshot images at the current position and the destination, instead of complex localization process. Here, we propose a new homing navigation method based on a moment measure to characterize the snapshot image efficiently. The method uses range values or pixel values of surrounding landmarks. Then it defines a moment measure to evaluate the environmental features, or landmark distributions, and the measure forms a convex shape of landscape with respect to robot positions in the environment. Based on the landscape, the mobile robot can return home successfully. Range sensors or image sensors can sufficiently provide the landscape information. Our experimental results demonstrate that the method is effective even in real environments.
AB - Visual navigation in robotics is one of the challenging issues, and many navigation approaches are based on localization of a mobile robot in the environment. The snapshot model is a biologically inspired model of insect behaviour to return home and it shows a simple algorithm to compare the snapshot images at the current position and the destination, instead of complex localization process. Here, we propose a new homing navigation method based on a moment measure to characterize the snapshot image efficiently. The method uses range values or pixel values of surrounding landmarks. Then it defines a moment measure to evaluate the environmental features, or landmark distributions, and the measure forms a convex shape of landscape with respect to robot positions in the environment. Based on the landscape, the mobile robot can return home successfully. Range sensors or image sensors can sufficiently provide the landscape information. Our experimental results demonstrate that the method is effective even in real environments.
UR - http://www.scopus.com/inward/record.url?scp=84984822326&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84984822326&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-43488-9_12
DO - 10.1007/978-3-319-43488-9_12
M3 - Conference contribution
AN - SCOPUS:84984822326
SN - 9783319434872
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 126
EP - 137
BT - From Animals to Animats - 14th International Conference on Simulation of Adaptive Behavior, SAB 2016, Proceedings
A2 - Hallam, John
A2 - Tuci, Elio
A2 - Giagkos, Alexandros
A2 - Wilson, Myra
PB - Springer Verlag
T2 - 14th International Conference on Simulation of Adaptive Behavior, SAB 2016
Y2 - 23 August 2016 through 26 August 2016
ER -