Mobile robot formation following method based on homography

Abstract

The invention discloses a mobile robot formation following method based on homography. The method comprises the steps that under the premise that a spacing distance of an ideal formation pattern and an ideal expectation image are given, a homography matrix is used to construct a virtual robot capable of reflecting real-time position poses of a following robot in the ideal formation pattern, and anoriginal formation problem is transformed into a trajectory tracking problem of the virtual robot. During the formation following, the speed of a piloting robot adopts an estimation mode, the piloting speed can be estimated more accurately by adopting a relation model between the homography and the speed and the real-time speed of the following robot, thereby avoiding adopting a local communication mode and saving the formation experiment cost. The method is simple and feasible, and can satisfy the requirements of the mobile robot formation following.

Description

technical field [0001] The invention relates to a trajectory tracking method, in particular to a homography-based mobile robot formation following method. Background technique [0002] Mobile robot formation control is a typical multi-robot collaborative problem, which has broad application prospects in military, life, industry, etc. Research on this type of problem mainly focuses on intelligent transportation fields such as dispatching and tracking. It has been a multi-robot problem in recent decades The hotspots and difficulties of robot collaboration related research, the research on this issue plays an important role when multi-robots need to work together to complete the job. [0003] Formation following is essentially a trajectory tracking problem. Therefore, most of the current research on formation control is based on trajectory tracking research. The traditional research methods mainly include: decomposing the original nonlinear system into multiple low-order subsys...

AI Technical Summary

Problems solved by technology

[0004] In recent years, the rapid development of robot visual servo control based on monocular cameras has made some related researches gradually focus on monocular visual feedback. Depth information, constructing a local three-dimensional map to locate the robot pose information in real time, and then complete the formation task, the cost is relatively high and the controller design is difficult; the relevant pose is estimated by adding mark points, provided that the relevant information needs to be marked and recorded in advance , providing more prior knowledge; methods based on multi-view geometry are more extensive, such as pole and homography decomposition, but there are also problems such as pole singularity or non-unique decomposition that affect system stability

Images