Environmental Perception Approach For The Whole Process Of Intelligent Lifting In Precast Construction

As the mainstay of Chinese economy,the construction industry is being with increasingly serious problems such as aging workers,frequent safety accidents and low production efficiency.Intelligent construction provides a new solution or direction for the above problems,while the promotion of prefabricated buildings strongly supports the development of intelligent construction.Lifting machinery is the most commonly-seen but indispensable equipment in precast construction,its intelligence is the key of intelligent construction as well.But it still lacks suitable intelligent lifting machinery.The intelligent perception of external environment is the precondition of intelligent lifting.However,the complexity,dynamics and disorder of lifting environment,as well as the diversity,timeliness and accuracy requirement of environmental perception,make it difficult to realize the real-time intelligent perception of lifting environment.Based on the lifting operation process,a fundamental framework of environmental perception for intelligent lifting operation is proposesd specifically focusing on the recognition of targeted components,dynamic obstacles,and aligning references.In terms of the recognition of targeted components,an automated positioning algorithm for components is presented to determine their exact location and poses on site by extracting the key geometric elements of components and analyzing their spatial relationship;then,an algorithm to determine a specific component is developed by extracting the features of components and matching with the database of components.In terms of the recognition of dynamic obstacles,an automated tracking algorithm for multiple obstacles in different scenarios is developed through the establishment of integrated features of obstacles,feature screening and feature clustering;then,a quadratic motion model is constructed to predict the trajectories of obstacles combined with Cubature Kalman Filter,and the trajectories are further optimized by integrating site environment.In terms of the recognition of aligning references,a line fitting method for connection rebars is first built to examine the verticality of rebars;then,an automated generation algorithm for lifting control planes is developed based on the dilated rebar contours and the smallest outer envelope;furthermore,the precise alignment problem of control planes is transformed into the alignment problem of three groups of non-collinear rebars and sleeves,and thekey geometric elements are extracted to realize the precise alignment.Besides,an environmental perception prototype system is developed by integrating the above methods,and used in site trials or indoor experiments.The experiment results indicate that the proposed approach is effective and workable.The conclusions include:(1)The intelligent recognition method of targeted components satisfies the corresponding lifting machinery’s perception requirements,with a recognition accuracy of 20 mm and a recognition time of 660.4 ms;(2)The intelligent recognition method of dynamic obstacles satisfies the corresponding lifting machinery’s perception requirements,with a tracking accuracy of 97.7%,a prediction error of 59 mm,and a recognition time of 140 ms;(3)The intelligent recognition method of aligning references satisfies the corresponding lifting machinery’s perception requirements,with a recognition accuracy of 2 mm and a recognition time of 1600 ms.In summary,this research not only enables the intelligentization of lifting machinery,with the improvement of efficiency and safety of precast construction,but also provides a reference for the intelligentization of other construction equipments,thereby promoting the rapid development of intelligent construction.