2023年度 人工知能学会全国大会（第37回）
2023年6月6日〜6月9日熊本城ホール(熊本県熊本市) ＋ オンライン
2023年度 人工知能学会全国大会（第37回）
2023年6月6日〜6月9日熊本城ホール(熊本県熊本市) ＋ オンライン
[3U1-IS-3-03]Using Social Force Maps to model idiosyncratic Movement of Crowd Flow Agents
〇Carl Kjaergaard2,1, Kenji Tanaka1(1. Tokyo University, 2. GRID INC)
[[Online, Working-in-progress]]
In most modern crowd flow simulations, agent movement is usually modelled for the purposes of idiosyncracies, allowing for personal attributes such as a feeling of personal space, or they are modelled logically, allowing for greater coordination and planning between agents. The objective of this paper is derived from this separation, as it attempts to introduce an in-between solution, that allows planning with idiosyncratic variables.
The result of this paper is a model that utilizes techniques from regular crowd flow modelling, physics and regular pathfinding. It starts by developing a procedurally generated map of social forces, through which it determines the most "comfortable" path that will eventually reach its target and uses that map to determine direction, which is used for other models in the simulation system.
The result of this paper is a model that utilizes techniques from regular crowd flow modelling, physics and regular pathfinding. It starts by developing a procedurally generated map of social forces, through which it determines the most "comfortable" path that will eventually reach its target and uses that map to determine direction, which is used for other models in the simulation system.