horizon
研究方向
A direction map for urban experience, spatial intelligence, and the built environment.
城市体验、空间智能与建成环境的方向地图。
Making urban life legible to machines — and governable by design.
让城市生活对机器可读——并通过设计可治理。
This is a direction map, not a deadline list. Cities are experienced by billions of partial observers every day — yet most urban intelligence still looks at the city from above, not from within it. The work ahead is to close that gap: turn lived experience into spatial intelligence, and spatial intelligence into decisions that shape the built environment.
这是一张方向地图，不是交付时间表。城市每天被数以亿计的局部观察者体验着——但大多数城市智能仍从 上方俯视城市，而非从内部理解城市。接下来的工作，是弥合这一鸿沟：把生活体验转化为空间智能， 再把空间智能转化为塑造建成环境的决策。
Direction evolves. The problems do not.
路径会变，问题不会。
Now
当下
Measuring what cities feel like
测量城市的感受
-
City in Sight 城市在眼前
Window and street as sensors — mapping urban visual experience where people actually live and move.
以窗户和街道为传感器——绘制人们真实生活与移动中的城市视觉体验。
-
Thermal Affordance 热舒适可供性
Reading thermal comfort from the visible streetscape — linking form, imagery, and felt experience.
从可见街景解读热舒适——连接形态、影像与身体感受。
-
Urban Comfort Intelligence 城市舒适智能
Toward one integrated, spatial field of urban comfort — subjective, measurable, and scalable.
走向统一、可度量的城市舒适空间场——主观、可测、可扩展。
Building
构建中
Assembling the intelligence stack
组装智能技术栈
-
Agentic Urban Science 智能体城市科学
Autonomous agents for planning and discovery — reasoning over geospatial context, not just retrieving it.
面向规划与科学发现的自主智能体——对地理空间语境进行推理，而非仅做检索。
-
Urban Foundation Models 城市基础模型
Geospatial LLMs and vision models as the cognitive substrate — a general layer for city-scale reasoning.
地理空间大语言模型与视觉模型作为认知基底——城市尺度推理的通用层。
-
Urban Digital Twins 城市数字孪生
Living twins that fuse imagery, perception, and environment — models that update as the city changes.
融合影像、感知与环境的活孪生——随城市变化而更新的模型。
-
Perception Infrastructure 感知基础设施
SP-Survey and allied platforms — human perception as reproducible, deployable urban infrastructure.
SP-Survey 及相关平台——将人类感知变成可复现、可部署的城市基础设施。
-
PhD Thesis 博士论文
Unifying and Quantifying Urban Comfort — the doctoral arc that binds perception, environment, and intelligence.
《统一与量化城市舒适》——连接感知、环境与智能的博士研究主线。
-
CityExperienceAI CityExperienceAI
Research translated into impact — spatial intelligence for government, industry, and everyday urban life.
研究转化为影响力——面向政府、产业与日常城市生活的空间智能。
Imagining
展望
Where the stack must lead
技术栈的终局方向
-
City World Models 城市世界模型
Dynamic models of how cities feel, flow, and respond — predictive intelligence, not static maps.
刻画城市如何感受、流动与响应的动态模型——预测性智能，而非静态地图。
-
Embodied Intelligence 具身智能
Agents and robots in built space — urban intelligence that perceives, navigates, and acts in the physical world.
建成环境中的智能体与机器人——在物理世界中感知、导航与行动的城市智能。
-
Legible Urban Life 可读的城市生活
Lived experience made machine-readable at city scale — the interface between people, cities, and autonomous systems.
城市尺度上可读的生活体验——人、城市与自主系统之间的接口。
-
Governable by Design 设计即可治理
From sensing to shaping — closing the loop so urban intelligence changes what we plan, build, and inhabit.
从感知到塑造——闭合回路，让城市智能改变我们规划、建造与居住的方式。