A proliferation of data collected by devices, sensors, robots, and people will change the way we live and prosper. Data-driven approaches will change the way we design, maintain and renovate infrastructure, and the manner by which we protect the environment.
The advent of the information era creates new needs, new performance requirements, and new threats. It creates also tremendous opportunities for performance optimization, monitoring and efficiencies.
Computational models run at computational speeds that were not considered possible a few years ago, new sensors with new measuring capabilities, next-generation, AI-trained control systems are already change the Civil, Construction and Environmental Industries.
ARGO-E GROUP can help you make sense of complex data by developing powerful visualization tools and analytics platforms. Our solutions enable you to gain valuable insights, track project performance, and make data-driven decisions to optimize your operations and improve project outcomes. Our group works with the entire data lifecycle, from collection, data analysis and then to decision making.
Our dedicated machine learning team leverages advanced algorithms and techniques to unlock the potential of data in the civil, environmental, and construction industries. From predictive analytics to anomaly detection, we develop machine learning models that enhance decision-making, improve risk assessment, and enable intelligent automation for your projects or customized apps.
Through our various initiatives, we have applied Artificial Intelligence (AI) approaches to science data, sensing data, geospatial data, as well as business data with the goal to leverage their full potential, and pave new avenues for the industry and the profession.
Example projects are highlighted below:
ARGO-E GROUP’s Machine Learning Division has developed a Python app to help engineers and scientists classify images easily without programming skills. Developed with UC Berkeley’s Professor Dimitrios Zekkos, the app is part of a rock slope characterization project.
It classifies images into categories such as good rock, poor rock, active landslides, terraces, and others. The app’s user-friendly interface allows for easy labeling and clustering of images, which can then be used to train Machine Learning algorithms.
Social media can be a truly unique source of data to analyze the characteristics of natural disasters and assess their impact on people and infrastructure in near-real-time. Data mining social media activity (e.g. on Twitter), combined with the application of Machine Learning and Natural Language Processing (NLP) can help mine the “signal” hidden in the social media that can provide unprecedented insights to natural disasters.
On-going research work on various natural disasters such as earthquakes, landslides, tsunamis, hurricanes and floods has demonstrated unprecedented insights on assessing the impact of the event on infrastructure that can be used to pinpoint expected damage in the immediate aftermath of a natural disaster.
The R&D group is actively working on a web platform whose purpose is to visually show the analysis of earthquake events immediately after they occur.
Our data analysis sheds some light on how American geotechs felt COVID-19 impacts their business. ARGO-E has been collecting data through GeoWorld’s Geotechnical Business Confidence Index (GBCI) and the results of our Geo-Business insights are particularly interesting.
A review of the role of Unmanned Aerial Vehicles in Civil Infrastructure
Our Founder, Dr. Dimitrios Zekkos has been a leader in data collection and analysis using Unmanned Aerial Vehicles with applications in civil infrastructure. He has authored numerous papers on the topic that can be found here. One of the key papers, published in 2020 is on “Applications of Unmanned Aerial Vehicles in Civil Infrastructure which was published in the Journal of Infrastructure Systems with co-authors Prof. Jerome Lynch and PhD student William Greenwood.
The paper can be found here:
https://ascelibrary.org/doi/abs/10.1061/%28ASCE%29IS.1943-555X.0000464
Unmanned aerial vehicles (UAV), or drones, have become popular tools for practitioners and researchers alike. Recent years have seen a significant increase in UAV uses for many applications in the fields of science and engineering. A broad array of research development in UAVs has been reported in the literature. This paper provides a summary review of efforts related to UAV development with a focus on civil infrastructure applications. First, guidance is provided for researchers looking to newly incorporate UAVs into their research efforts. The advantages and disadvantages between different UAV types are outlined and performance characteristics discussed. Examples of different sensor payloads that demonstrate expanded functionality are provided. The review also provides an overview of research efforts in the emerging domain of wireless sensor networks and data processing algorithms specific to UAV-collected data. Highlights of recent achievements of UAVs in post-disaster reconnaissance, infrastructure component monitoring, geotechnical engineering, and construction management are presented. Lessons learned from UAV implementation and considerations for good practice are also discussed. The paper concludes with a discussion of the emerging and future research domains that address the most pressing knowledge gaps in current practice.
Data Fusion for Infrastructure Monitoring
The proliferation of sensors on robots, drones and satellites provide unprecedented sensing opportunities to assess the interaction of the built environment with climate, people and habitat. The sensing data can be used to guide decisions that enhance resiliency and sustainability, and optimize energy and resource efficiency.
Global Monitoring of Landslides Using Social Media
A data processing pipeline that involved filtering, geoparsing, geocoding and clustering as well as context learning Machine learning algorithms was developed and used to generate a database of landslides using social media posts. The database provides an inventory of landslides that includes estimate of timing, location, size as well as consequences. An output of that database is shown in the figure for the year 2018. The algorithm is operating continuously since 2018.