Computer-Aided Civil and Infrastructure Engineering

Abstract:  The artificial neural network (ANN) is one advance approach to freeway travel time prediction. Various studies using different inputs have come to no consensus on the effects of input selections. In addition, very little discussion has been made on the temporal–spatial aspect of the ANN travel time prediction process. In this study, we employ an ANN ensemble technique to analyze the effects of various input settings on the ANN prediction performances. Volume, occupancy, and speed are used as inputs to predict travel times. The predictions are then compared against the travel times collected from the toll collection system in Houston. The results show speed or occupancy measured at the segment of interest may be used as sole input to produce acceptable predictions, but all three variables together tend to yield the best prediction results. The inclusion of inputs from both upstream and downstream segments is statistically better than using only the inputs from current segment. It also appears that the magnitude of prevailing segment travel time can be used as a guideline to set up temporal input delays for better prediction accuracies. The evaluation of spatiotemporal input interactions reveals that past information on downstream and current segments is useful in improving prediction accuracy whereas past inputs from the upstream location do not provide as much constructive information. Finally, a variant of the state‐space model (SSNN), namely time‐delayed state‐space neural network (TDSSNN), is proposed and compared against other popular ANN models. The comparison shows that the TDSSNN outperforms other networks and remains very comparable with the SSNN. Future research is needed to analyze TDSSNN's ability in corridor prediction settings.

In this article, an accurate method for the registration of point clouds returned by a 3D rangefinder is presented. The method modifies the well‐known iterative closest point (ICP) algorithm by introducing the concept of deletion mask. This term is defined starting from virtual scans of the reconstructed surfaces and using inconsistencies between measurements. In this way, spatial regions of implicit ambiguities, due to edge effects or systematical errors of the rangefinder, are automatically found. Several experiments are performed to compare the proposed method with three ICP variants. Results prove the capability of deletion masks to aid the point cloud registration, lowering the errors of the other ICP variants, regardless the presence of artifacts caused by small changes of the sensor view‐point and changes of the environment.

A number of image processing techniques (IPTs) have been implemented for detecting civil infrastructure defects to partially replace human‐conducted onsite inspections. These IPTs are primarily used to manipulate images to extract defect features, such as cracks in concrete and steel surfaces. However, the extensively varying real‐world situations (e.g., lighting and shadow changes) can lead to challenges to the wide adoption of IPTs. To overcome these challenges, this article proposes a vision‐based method using a deep architecture of convolutional neural networks (CNNs) for detecting concrete cracks without calculating the defect features. As CNNs are capable of learning image features automatically, the proposed method works without the conjugation of IPTs for extracting features. The designed CNN is trained on 40 K images of 256 × 256 pixel resolutions and, consequently, records with about 98% accuracy. The trained CNN is combined with a sliding window technique to scan any image size larger than 256 × 256 pixel resolutions. The robustness and adaptability of the proposed approach are tested on 55 images of 5,888 × 3,584 pixel resolutions taken from a different structure which is not used for training and validation processes under various conditions (e.g., strong light spot, shadows, and very thin cracks). Comparative studies are conducted to examine the performance of the proposed CNN using traditional Canny and Sobel edge detection methods. The results show that the proposed method shows quite better performances and can indeed find concrete cracks in realistic situations.

Abstract:  Today's mobile and remote construction applications, such as tracking of materials, equipment, and workers, demand high reliability and scalability of wireless sensor networks for a large‐scale construction site. In particular, identifying the location of distributed mobile entities throughout wireless communications becomes the primary task to realize the remote tracking and monitoring of the construction assets. Even though several alternative solutions have been introduced by utilizing recent technologies, such as radio frequency identification (RFID) and the global positioning system (GPS), they could not provide a solid direction to accurate and scalable tracking frameworks in large‐scale construction domains due to limited capability and inflexible networking architectures. This article introduces a new tracking architecture using wireless sensor modules and shows an accuracy performance using a numerical simulation approach based on the time‐of‐flight method. By combining radio frequency (RF) and ultrasound (US) signals, the simulation results showed an enhanced accuracy performance over the utilization of an RF signal only. The proposed approach can provide potential guidelines for further exploration of hardware/software design and for experimental analysis to implement the framework of tracking construction assets.

Abstract:  Damage detection and calibration in beams by wavelet analysis involve some key factors such as the damage model, the choice of the wavelet function, the effects of windowing, and the effects of masking due to the presence of noise during measurement. A numerical study has been performed in this article addressing these issues for single and multispan beams with an open crack. The first natural modeshapes of single and multispan beams with an open crack have been simulated considering damage models of different levels of complexity and analyzed for different crack depth ratios and crack positions. Gaussian white noise has been synthetically introduced to the simulated modeshape and the effects of varying signal‐to‐noise ratio have been studied. A wavelet‐based damage identification technique has been found to be simple, efficient, and independent of damage models and wavelet basis functions, once certain conditions regarding the modeshape and the wavelet bases are satisfied. The wavelet‐based damage calibration is found to be dependent on a number of factors including damage models and the basis function used in the analysis. A curvature‐based calibration is more sensitive than a modeshape‐based calibration of the extent of damage.

Abstract:  Traditional documentation capabilities of laser scanning technology can be further exploited for urban modeling through the transformation of resulting point clouds into solid models compatible for computational analysis. This article introduces such a technique through the combination of an angle criterion and voxelization. As part of that, a k‐nearest neighbor (kNN) searching algorithm is implemented using a predefined number of kNN points combined with a maximum radius of the neighborhood, something not previously implemented. From this sample, points are categorized as boundary or interior points based on an angle criterion. Façade features are determined based on underlying vertical and horizontal grid voxels of the feature boundaries by a grid clustering technique. The complete building model involving all full voxels is generated by employing the Flying Voxel method to relabel voxels that are inside openings or outside the façade as empty voxels. Experimental results on three different buildings, using four distinct sampling densities showed successful detection of all openings, reconstruction of all building façades, and automatic filling of all improper holes. The maximum nodal displacement divergence was 1.6% compared to manually generated meshes from measured drawings. This fully automated approach rivals processing times of other techniques with the distinct advantage of extracting more boundary points, especially in less dense data sets (<175 points/m²), which may enable its more rapid exploitation of aerial laser scanning data and ultimately preclude needing a priori knowledge.

The three‐dimensional mapping of the built environment is of particular importance for engineering applications such as monitoring work‐in‐progress and energy performance simulation. The state‐of‐the‐art methods for fitting primitives, non‐uniform B‐Spline surface (NURBS) and solid geometry to point clouds still fail to account for all the topological variations or struggle with mapping of physical space to parameter space given unordered, incomplete, and noisy point clouds. Assuming an input of points that can be described by a single non‐self‐intersecting NURBS, this article presents a new method that leverages segmented point clouds and outputs NURBS surfaces. It starts by successively fitting uniform B‐Spline curves in two‐dimensional as planar cross‐sectional cuts on each surface. An intermediate B‐Spline surface is then computed by globally optimizing and lofting over the cross‐sections. This surface is used to parameterize the points and perform final refinement to a NURBS. For cylindrical segments such as pipes, a new supervised method is also introduced to string the fitted segments, identify connection types, standardize the connections, and then refine them using NURBS optimization. Experimental results show the applicability of the proposed methods for as‐built modeling purposes.

Abstract:  Periodic infrastructure maintenance is crucial for a safe and efficient transportation system. Numerous decision models for the maintenance planning problem have been proposed in the literature. However, to the best of our knowledge, no model exists that simultaneously accounts for traffic dynamics and is intended for long‐term planning purposes. This article addresses this gap in the literature. A mixed‐integer bi‐level program is introduced that minimizes the long‐term maintenance cost as well as the total system travel time. For the solution approach we utilize a genetic algorithm in conjunction with mesoscopic traffic simulation. The model is illustrated via a numerical example.