摘要

智能建造（IC）技术，尤其是人工智能驱动系统与自主机器人，是建筑工程（AEC）行业转型升级的关键，但其应用表现存在显著项目差异，即场景依赖性问题。现有评价方法多局限于单一案例研究，无法刻画该差异，难以支撑可靠的投资决策。针对这一不足，本文构建全新的期望–稳健性评价框架。该框架首先采用多准则效益评价体系，将单项工程的技术应用表现量化为标准化综合效益分值（OBS）；再以多项目 OBS 分值为统计样本，定义两项核心指标：一是期望效益，以 OBS 均值表征技术平均应用水平；二是性能稳健性，以 OBS 变异系数倒数表征技术在不同场景下的表现稳定性。依托全国智能建造试点工程 678 项大规模应用案例数据完成框架验证；对地坪、铺贴等 10 类典型建筑机器人及智能建造技术的分析表明，各类技术表现特征差异显著，凸显性能稳健性在工程落地中的重要作用。该框架突破传统单点评价局限，可为相关方在技术投资、研发及落地应用方面提供科学决策依据，助力智能建造行业可持续发展。

The adoption of Intelligent Construction (IC) technologies, particularly AI-driven systems and autonomous robots, is pivotal for the architecture, engineering, and construction (AEC) industry’s transformation, yet their performance often varies significantly across projects - a challenge known as context dependency. Existing evaluation methods, focused on single-case studies, fail to capture this variance, hindering reliable investment decisions. To address this gap, we propose a novel “Expectation-Robustness”framework. The framework first utilizes a multi-criteria benefit evaluation system to quantify a technology’s performance in a single project into a standardized Overall Benefit Score (OBS). Then, by treating the OBS scores from multiple projects as a statistical sample, we define two key metrics: (1) Expected Benefit, measured by the mean of OBS, representing the technology’s average performance; and (2) Performance Robustness, measured by the inverse of the Coefficient of Variation (CV) of OBS, representing its stability across diverse contexts. We validated the framework using a large-scale dataset of 678 technology application cases from national IC pilot projects in China. An analysis of 10 typical technologies, with a special focus on construction robots (e.g., for floor-finishing and tile-laying), revealed distinct performance profiles, highlighting the critical role of performance robustness in real-world robotic deployment. Our framework moves beyond simplistic single-point estimates, offering a more scientific basis for stakeholders to make informed decisions on technology investment, research, development, and deployment, thereby promoting the sustainable development of intelligent construction.

论文下载链接

预印本下载链接

相关项目资助情况: