A path-based multi-agent navigation model

Abstract

The quality of a crowd simulation model is determined by its agents' local and global trajectory efficiency. While an agent-based model can accurately handle the local trajectories, global decisions usually are handled by a global path planner. However, most of the global path planning techniques do not consider other agents and their possible paths and the future global flow in the environment. In this paper, we propose a composite system that takes future agent configurations into account via a modified A* algorithm to create a global path plan and combines the global path plan with a local navigation model. We show that the agents using the proposed model intelligently plan their paths based on the dynamic configuration of the environment. In order to balance the performance vs. trajectory quality trade-off, we propose a hierarchical grid structure and discuss its effects on both trajectory quality and computational performance.

A time‐based global path planning strategy for crowd navigation

“…We performed the tests for three different k values corresponding to 0.1, 0.3, and 0.5, similar to the original study . In the following sections, these three different versions are referred to as P 1 , P 3 , and P 5 , respectively.…”

“…We use RVO for the local navigation purposes in the minisimulations. As widely known, RVO may cause the agents to significantly deviate from their route to avoid collisions when they come across large groups moving in a different direction . This flaw is negligible in the real simulation as the proposed method (assuming it works as expected) encourages the agents to move in a flow with the other agents for most of the time.…”

“…A variety of composite models have been proposed by extant studies with the intention of combining the agent‐specific short‐term capabilities of the local navigation methods and high‐quality long‐term paths of the global navigation methods. An example of the composite models is the PBM presented by Ozcan et al in which the agent density in the environment is represented by a value termed as the potential information . They store the potential information, which includes the flow direction, in the cells of a grid structure that covers the simulation environment.…”

“…We compare the proposed method with a recent and competitive approach where the global navigation information of the crowd members are represented by some values termed as potential values . The simulation results for the mass scenarios (conventional circle scenario and cross scenario) show that the proposed model slightly outperforms the potential‐based method (PBM).…”

Editorial

“…Three papers by Barut and Haciomeroglu [9], Che et al [10] and Ozcan and Haciomeroglu [11] are about crowd simulation. The four other papers by Lee et al [12], Kim et al [13], Hartmann et al [14] and Choi et al [15] focus on the problem of human motion synthesis and retargeting.…”