An Efficient Centralized Planner for Multiple AGVs at the Crossroad of Polynomial Curves

In this paper, we introduce a centralized planner with low computational cost to schedule the motions of multiple Automated Guided Vehicles (AGVs) at the intersection of pre-defined polynomial curves. In particular, we find that the collision conditions between two AGVs along polynomial paths can be formulated as a set of polynomial inequalities. Furthermore, by solving these inequalities, the continuous boundaries of potential collision areas for vehicles can be determined offline and stored in a table. During the online phase, by also taking into account the priority setup among AGVs, the planner will use efficient table lookup to determine a list of intermediate goals for each AGV to move toward based on their real-time position feedback. In this way, the multiple robots are able to navigate on polynomial guide paths safely and efficiently at the crossroad, and the performance of our approach is demonstrated in a number of simulated scenarios. 

This work is supported by Hong Kong General Research Fund (GRF) 11207818, 11202119, and the Centre for Transformative Garment Production. 

We would like to thank Mr. T. Liu and Mr. C. Z. Kuang for meaningful discussions.