Decision and finite-time cooperative control of multi-robot systems for odour source localisation

In the last two decades, the problem of odour source localisation has been widely studied by using a single robot due to its practical significance for human security, such as searching for the source of toxic gas leakage and locating the origin of a fire at its initial stage. Recently, using a multi-robot system to locate the odour source has received increasing interest from researchers because of several major benefits over a single robot such as a wider detection range and multiple detection information. In this dissertation, two aspects on the odour source localisation problem are studied; one is to learn from data detected and collected by the multi-robot system to make a decision on the position of the odour source while the other is to coordinate and control the multi-robot system to locate the source of odour in terms of the decision results. In particular, a distributed coordination control architecture including two levels: a decision level and a control level, is designed. In the decision level, a new distributed decision algorithm, which can make a decision on the position of the odour source, is formulated. In the control level, a particle swarm optimization based finite-time motion control algorithm, a consensus-based finite-time motion control algorithm, and a potential-based finite-time motion control algorithm, are then developed to control the robot group to locate the odour source, respectively. Finally, the effectiveness of the proposed solutions consisting of the architecture, the decision algorithm, and the control algorithms for odour source localisation is illustrated through simulations.