运用无人机（UAV）技术搜集工程现场险兆事件减少事故风险

Ceaseless construction accidents indicate that the problem of construction safety has not been radically solved. It is apparent that the construction industry is far from the vision of “zero accidents/injuries” espoused by many construction-related companies. Many accident investigation reports in the construction industry implied that accidents did not occurred without any precursors. There are dozens of incidents without any injuries or loss prior to fatal accidents. This type of incident is called “Near-Miss” which is usually neglected by workers and managers on construction sites. Near-Misses are regarded as free lessons for safety management and Near-Miss management can be an effective way to avoid serious accidents. The continuous development of information and communications technology (ICT), and unmanned aerial vehicle (UAV) or drone, provides the possibility to make better use of Near-Misses to promote construction safety management. On the basis of set theory, Murphy’s law, iceberg model and accident causing theory, the frame of Near-Miss theory in the field of construction safety management will be explored. The methods of expert interview, questionnaire and experiment will be adopted to identify what functions UAVs should have in the task of Near-Miss collection. If the functions of the existing UAVs are enough to collect Near-Misses, it is not necessary to amend UAVs; if not, it is necessary to amend existing UAVs, such as adding some types of sensor in UAVs. Considering the features of individual construction site, the reasonable and economic route will be identified before flight of an UAV according to topology. Based on grounded theory, videos from UAV will be used to identify Near-Misses on the site. Three steps of coding (open coding, axial coding and selection coding) will be contained. Then Near-Miss will be selected and analyzed. Corrective and preventive measures will be proposed. After proposing solutions, the whole process will be disseminated and estimated. METHONTOLOGY will be revised to be adaptive to building Near-Miss ontology in the field of construction safety management, based on IFC (Industry Foundation Class). Protégé will be selected as the ontology editor. Finally, a case based reasoning system on the basis of Near-Miss ontology will be developed including retrieving Near-Misses, reusing Near-Misses, revising solutions, and new cases retention to update the database.

频繁发生的事故表明建筑工程安全问题尚未从根本上解决，与“零事故”目标相去甚远。通过分析大量工程事故，发现每起事故并非毫无征兆地偶然发生，其发生之前会出现许多无伤害无损失的险兆事件（Near-Miss），研究表明险兆事件可以成为规避事故的有效手段。新兴信息技术和无人机技术的不断发展，为有效利用险兆事件提高工程安全管理水平提供了可能。本课题将以集合论、墨菲定律、冰山模型和事故致因理论为基础探究险兆事件在建筑工程安全管理领域的理论框架。同时，通过专家访谈、问卷调查和实验的方法，确定现有无人机系统是否满足险兆事件搜集的需要。并且，运用改良的UAV系统进行险兆事件实时搜集，基于扎根理论分析识别险兆事件，进而设计与实施险兆事件解决方案。然后采用改进的METHONTOLOGY方法，基于IFC标准体系构建险兆事件本体库。最后，基于本体库构建案例推理模型，以指导险兆事件解决方案的设计与实施。

频繁发生的事故表明建筑工程安全问题尚未从根本上解决，与“零事故”目标相去甚远。通过分析大量工程事故，发现每起事故并非毫无征兆地偶然发生，其发生之前会出现许多无伤害无损失的险兆事件（Near-Miss），研究表明险兆事件可以成为规避事故的有效手段。新兴信息技术和无人机技术的不断发展，为有效利用险兆事件提高工程安全管理水平提供了可能。本研究课题以集合论、墨菲定律、冰山模型和事故致因理论为基础探究险兆事件在建筑工程安全管理领域的理论框架。同时，通过专家访谈、问卷调查和实验的方法，确定现有无人机系统是否满足险兆事件搜集的需要。并且，运用改良的UAV系统进行险兆事件搜集，基于扎根理论分析识别险兆事件，进而设计与实施险兆事件解决方案。然后采用改进的METHONTOLOGY方法，构建Near-Miss本体库。最后基于本体库构建Near-Miss推理模型，以指导Near-Miss解决方案的设计与实施。