A Study of Gradient Climbing Technique Using Cluster Space Control of Multi-Robot Systems

摘要

The design of the multi-robot system for distributed sensing and gradient climbing focuses on the capability to optimize the performance of tasks simultaneously. The strategy is to utilize the cluster’s redundancy and flexibility to gain and maximize the overall coverage of surveying parameters so as to surpass the performance of any single robot. The collaborative nature of the cluster provides a more efficient and effective platform for collecting data and conducting fieldwork. The purpose of this study is to explore the existing cluster space control technique to show effective gradient-based navigation, particularly that of climbing a gradient in a sensed parameter field to the local maximum. In order to achieve positive results, we need to estimate the gradient direction based on real-time measurements captured by sensors on the distributed robotic network, and then maneuver the cluster to travel in the estimated direction. Verification and characterization of this technique has been performed through both simulation and hardware-in-the-loop experimentation. In these tests, the gradient controller enabled the cluster to sense and climb the gradient in a parameterized field using kayaks in a marine environment and utilizing wheeled robots in a land based system. The successful outcome of these demonstrations proves the value of the cluster space control technique and showcases how it can be used for efficiently locating minimum and maximum features in a parameter field.