Using Evidence Accumulation Method for SolvingTask of Collective Perception

This study advances collective perception in swarm robotics by introducing the Automated Swarm Opinion Diffusion Model, addressing limitations in the original Swarm Opinion Diffusion Model. While Swarm Opinion Diffusion Model integrates social and personal information for decision-making, its reliance on manually predetermined social factor parameter restricts adaptability to diverse task complexities. The novel Automated Swarm Opinion Diffusion Model eliminates this dependency by introducing an adaptive personal factor parameter, which automatically adjust the weighting of personal and social information based on the gathered information about the environment. This automated approach improves robustness and reduces the dissemination of erroneus information in early phases of the task. Comparative simulations against baseline methods (Voter model and Majority rule) demonstrate Swarm Opinion Diffusion Model's and Automated Swarm Opinion Diffusion Model's superior performance in both accuracy and consensus speed, particularly at the tasks with higher complexities. Additionally, Automated Swarm Opinion Diffusion Model maintains comparable efficiency to Swarm Opinion Diffusion Model while automating critical parameter selection, making it more suitable for real-world applications where prior knowledge about the environment and complexity of the task is unavailable. Future research will focus mostly on extending Automated Swarm Opinion Diffusion Model to best-of-n decision-making problem, where n is greater than 2, enhancing its applicability in various real-world swarm robotics scenarios.