Food4Rhino

NeosExplorer is an agent simulation system developed on the Rhino-Grasshopper platform, with source code written in C#. Its primary function is to simulate exploratory movement or predictive path-guided behaviors of agents in 2D space, applicable to pedestrian simulations or behavioral studies of other organisms.

Version 2.0.0 has added input ports for evacuation zones and output ports for evacuation time, enabling building space optimization based on evacuation duration.

Development Background

NeosExplorer was developed as a viable alternative to the Grasshopper (GH) pedestrian simulation tool PedSim, which is no longer available through public platforms or channels. While PedSim's internal implementation details remain unclear to users, its simulation results suggest it was designed to guide agents toward goals via shortest paths.

Core Algorithm & Modes

NeosExplorer shares the same core algorithm as PedSim—the Social Force Model from pedestrian dynamics. Version since 1.2.2 offers two simulation modes:

Exploration Mode (using Explorer Simulator):

This mode operates without path guidance. Upon encountering obstacles, agents identify the nearest feature point within their field of view, adjust their direction, visit points of interest (PoI) for a designated duration, and then proceed toward the destination. PoIs are optional, and multiple PoIs are supported.

Note: This mode currently has limited adaptability for complex scenes; regular geometric shapes are recommended for obstacle outlines.

Path Guidance Mode (using Pedestrian Simulator):

This mode pre-computes the shortest path from the start point to the destination via PoIs using a 2D visibility graph and the A* algorithm, guiding agent behavior accordingly. It includes a reliable movement assurance mechanism suitable for scenes with complex obstacle contours.

Critical Note: If closed polylines acting as obstacles are manually drawn point-by-point, they must be drawn counter-clockwise (unless a value of 0 is input to the Navigation Offset, NO parameter).

Flexibility & Accessible Data

NeosExplorer exposes most adjustable parameters to users for flexibility, including desired speed, maximum speed, mass, field-of-view angle, and visibility range. Adjusting these parameters allows adaptation to diverse scenarios. Users can also access comprehensive agent data during simulations—such as current position, velocity, resultant force, movement trajectory, and field of view—to support varied research needs.

Core Engine & Operation

The core engines are the Explorer Simulator and Pedestrian Simulator, requiring iterative computation via Grasshopper’s Trigger (Timer) component.

Key Concept: "Time" in this simulation refers to iteration time steps. For example:

If the Trigger’s time interval = 100 ms (i.e., iterates every 100 ms),

and an agent’s stay duration at a PoI = 5,

the actual stay duration is only 500 milliseconds.

This may render stays imperceptible during real-time simulation. The same principle applies to inertia duration and departure interval.

Component Info & Test Environment

NeosExplorer is part of the author’s Neos series of Grasshopper components. Upon installation, it appears under the Neos Category and NeosExplorer Subcategory in the Grasshopper toolbar.

Tested on: Windows 11, Rhino 8 SR20 (v8.20.25157.13001, 2025-06-06) – Grasshopper 1.0.0008 (2025-06-06).

Acknowledgement

We sincerely hope NeosExplorer supports your research and work.

(Developer: 黄华飞/Wong Wah Fai / Huang Hua-fei)

NeosExplorer是一个基于Rhino-Grasshopper平台开发的智能体仿真系统，源码为C#，其功能为在二维空间模拟智能体的探索性行进行为或预知性路径引导行为，可用于行人模拟或其他生物的行为模拟。

版本2.0.0简化了双向人流的模拟实现方法，并增加了疏散区域输入端与疏散时间输出端，使基于疏散时长的建筑空间优化成为可能。

NeosExplorer的开发初衷是寻求替代GH行人模拟工具PedSim的可行方案，因为GH插件PedSim已经不能在公开的平台或渠道上获得。PedSim的内部实现细节用户并不清楚，根据模拟结果，猜测其设计理念应该是以最短路径引导智能体走向目标。

NeosExplorer的算法核心与PedSim一样，同为行人动力学中的社会力模型，1.2.2以后的版本提供了两种模拟模式：使用Explorer Simulator的探索模式与使用Pedestrian Simulator的路径引导模式。

探索模式不依赖路径引导，智能体遇到障碍物后会寻找视野内最近的特征点并改变运动方向，找到感兴趣的地方并访问一段时间后再走向终点。兴趣点非必需，也支持多个兴趣点。这种模式目前并不适用于复杂场景，建议使用规则的几何形作为障碍物的平面轮廓。

路径引导模式基于二维可见性图与A*算法预先计算从起点经过各兴趣点直至终点的最短路径，以最短路径引导智能体的行为，同时具有可靠的移动保障机制，适用于障碍物平面轮廓复杂的场景。

需要注意，如果作为障碍物的闭合多段线是手动逐点绘制的，则需要逆时针方向绘制（如果"Navigation Offset, NO"输入端输入0则可忽略该要点）。

NeosExplorer为了保持灵活性对用户开放了绝大部分可调参数，如期望速度、最大速度、质量、视角、视距等，通过调节各类输入参数可适应不同的应用场景。同时，用户可获得模拟场景中智能体的绝大部分信息，如当前位置、速度、合力、轨迹、视野等，以适配不同的研究需求。

NeosExplorer的核心引擎为Explorer Simulator与Pedestrian Simulator，需要借助GH的Trigger（Timer）组件进行迭代运算。需要注意，行人模拟中涉及的“时间”概念指迭代时间步数（time step），例如，当Trigger的时间步长（time interval）为100毫秒，即100毫秒迭代一次，行人在兴趣点的停留时长（stay duration）设置为5，则实际停留时长仅为500毫秒，在实时模拟过程中将观察不到明显的停留迹象。对惯性持续时长（inertia duration）、出发间隔（departure interval）的理解同理。

NeosExplorer是作者开发的Neos系列GH组件之一，安装后其Category为Neos，Subcategory为NeosExplorer。

NeosExplorer在Window11，Rhino8 SR20 (8.20.25157.13001, 2025-06-06)-Grasshopper 1.0.0008（2025-06-06）平台进行了测试。

衷心希望NeosExplorer能支持你的研究与工作。

（开发者：黄华飞/WongWahFai/Huang Hua-fei）