Simultaneous localization and mapping (SLAM) is a core technology that is required for teams of mobile robots to cooperate in everyday environments. This thesis proposes a multi-robot SLAM architecture that enables multiple robots to operate both indoors and outdoors, using only onboard sensors, and without GPS or pre-existing maps. It contributes robust and efficient algorithms that share sensor data between robots and distributes the SLAM back-end computation, while allowing teams to operate for periods independent of a centralized server. The architecture and algorithms are demonstrated In real-world conditions, with up to 23 heterogeneous robots exploring a 500x500 meter urban environment.
|Qualification||Doctor of Philosophy|
|Award date||29 Jul 2016|
|Publication status||Unpublished - 2016|