Comparative study of orchestrated, centralised and decentralised approaches for orchestrator based task allocation and collision avoidance using network controlled robots

Abstract

We report a design for an orchestrator controlled robotic network that organises and manages a set of activities through a task allocation model for the robotic nodes. The model also helps to study the robotic-orchestration with the help of centralised and decentralised approaches. Additionally, our model incorporates an innovative collision avoidance algorithm that avoids collision between parallel running robots without any communication between them. The results of the orchestrated, centralised and decentralised approaches are compared using the Paired signed and Wilcoxon signed rank tests. Simulated results, when compared with the experimental data, show that the orchestrated approach is better than the decentralised approach in terms of communication bandwidth and performance’s consistency. In addition, the orchestrated approach is better than the centralised approach, in terms of both communication overloads and task execution time.