Visual servoing is a guaranteed high-precision robotic control scheme but inevitably encounters several constraints. This paper aims to propose a planner yielding feasible feature trajectories, accounting for physical and image constraints that inevitably hinder visual servoing applications at the planning stage.
The proposed planner explores the image space for permissible paths by progressively extending a search tree in this space. Along with the global planning framework, an artificial potential field-based local planner handles three significant constraints (field-of-view, joint position limits and singularity) compared to existing similar planners. The planned feature paths are further time-parameterized into a desired image trajectory and, finally, tracked by an image-based visual servoing scheme to reach the goal features by continuously driving the robot.
Some constraints can be directly handled at the local planning stage in the image space to explore the search space better. Also, a sampling-based global framework can boost performance and generate a feasible trajectory for VS problems in 3 s and a 94% success rate.
Various experiments are conducted on a real robot with seven degrees of freedom to demonstrate the effectiveness of the proposed planner. Fair and exhaustive experiments of comparison against several state-of-the-art methods demonstrate the performance of the proposed method.
