Stefania Costantini

Bridge health monitoring and machine learning are increasingly intertwined for civil engineers and artificial intelligence experts. Bridges’ poor health can result in severe outcomes if not addressed in time. Therefore, continuous monitoring is required to detect any anomaly or damage. Sensors, such as accelerometers, inclinometers, thermistors, etc., can help actively monitor these bridges. The signals from these sensors help record physiological activities. Such activities are helpful for anomaly detection, damage localization, and bridge health predictions with the help of machine learning algorithms. The proposed method extracts features from the dynamic response of a bridge to ambient excitation. It focuses on processing the signal received from different accelerometers installed on a steel railway bridge to determine the location of the damage and the level of the damage predictions. Initially, features are extracted from time-series data; then, they are fed to a deep neural network after some pre-processing. Normal and augmented data are used with different parameter tuning for results. Original data is also subdivided, and the effect of data slicing on the predictions is investigated. The results show that one-fourth of the slicing of the original data gives the best results for training and testing accuracy with a deep neural network. The results show that the reduced matrix representation, particularly the 40 × 40 feature slicing, improved the classification performance for the predefined bridge scenario classes under the considered experimental settings. For bridge scenario classification, the best reported accuracy was 93.54%, while for damage intensity classification the best reported accuracy was 98.21%. In the DNN-based optimizer comparison, the Adam optimizer achieved higher and more stable performance than Stochastic Gradient Descent (SGD), with test accuracies of 92.3% and 93.7% compared with 75.2% and 86.4%, respectively. It is also observed that the Adam optimizer outperformed Stochastic Gradient Descent (SGD) in terms of both damage localization and damage intensity estimation.

In the near future, transportation systems will include both autonomous vehicles and human-operated vehicles sharing the same traffic conditions. Human drivers will have difficulty predicting the actions of autonomous vehicles and the latter will face challenges due to complex decision-making algorithms and dynamic environments. The lack of standardized interaction protocols between autonomous vehicles and human drivers further complicates safe decision-making. This paper proposes an AI-based advisory framework to enhance human driving skills in mixed autonomy traffic and improve autonomous vehicles in a Human–AI teaming fashion. Our framework is composed of both a centralized component and a decentralized component. The centralized component primarily identifies driving style and trajectory parameters that impact traffic efficiency across large-scale traffic networks shared by human drivers and autonomous vehicles. At the local level, however, our proposed framework features a decentralized, agent-based strategy to enable effective coordination between human and autonomous vehicles—especially at complex intersections. An initial prototype is modeled and implemented in a desktop virtual reality environment for testing and training.

How to avoid unethical practices like bias, manipulation, causing harm, and instead building ethical machines has been a topic of investigation and discussion within the Artificial Intelligence (AI) field for more than two decades. If there were clear rules to follow, AI would have long ago demonstrated how to avoid unethical practices. We suggest that such moral norms could perhaps emerge from experience by ethical reasoning about particular situations in different domains, and evolve over time. In this work, we review the field of machine ethics in the last two decades or so, discussing the challenges and outlining possible future directions and potential developments. Building on these insights, we propose a modular framework for deriving practical ethical rules for AI systems from experience, enabling a more transparent and adaptive approach to moral decision making in AI. The potential of the framework is illustrated by means of a couple of case studies in the medical domain.