As the world’s urban population rapidly grows, cities all over the world are experiencing severe traffic congestion. Traffic congestion causes widespread socioeconomic and environmental issues for cities. As an example, the New York City economy is expected to incur losses of over $20 billion over the next five years because of traffic congestion that leads to unforeseen business expenses, wasted fuel consumption from increased idling, and increased vehicle maintenance costs. In addition to the rapid growth of the number of vehicles on the road, unexpected roadway conditions caused by roadside construction/maintenance, car accidents, asynchronous traffic signals, changes in driver behavior, among other factors also cause the traffic congestion in urban areas. The propagation of traffic disturbances along with a lack of rapid information sharing between drivers can easily lead to widespread delays in an entire city’s traffic network.
Natural Language Processing (NLP) for Intelligent Transportation Systems
Applying Natural Language Processing (NLP) in social media for the purpose of leveraging underutilized transportation-related posts for Intelligent Transportation Systems (ITS) applications is an active area in research. In this project, we propose to support and complement traffic reporting systems and navigation assistants by exploiting the abundance of data in social media and use it as an additional source of traffic information. To this end, we develop an automated framework that automatically processes social media data from the web, classifies it, and extracts traffic-related reports to convert them into navigation alerts. One situation where this framework has the ability to effectively work is the case when drivers are stuck in heavily-congested roadways. There is a tendency that users could post about the situation in social media, which is meaningful for other drivers. However, this is not the only unique possibility.
The framework also uses inputs from specialized agencies and from regular people who are not necessarily driving such as cyclists, pedestrians, or passengers in vehicles, ride-sharing taxis, or buses, etc. The framework automatically browses social media platforms, distinguishes traffic-related messages, extracts/understands the incidents information, and converts them into alerts in the navigation apps. A joint text processing framework based on fine-tuning BERT classification models for filtering the traffic related information, and either a Question-answering (QA) model or a name-entity recognition (NER) model for extracting real-time traffic details data from social media are developed. In this two-phase NLP pipeline, we first develop a filter that classifies numerous collected text inputs into different groups. Then, we extract necessary information from these filtered groups in order to automatically understand and characterize the reported traffic event from social media by determining its location, its occurrence time, and its nature, e.g., blocked road, accident, etc. Once the detailed traffic information is obtained, we implement an automated system that converts collected posts from social media into real-time traffic information that are used to update navigation maps and warn drivers about any traffic events.
The framework also uses inputs from specialized agencies and from regular people who are not necessarily driving such as cyclists, pedestrians, or passengers in vehicles, ride-sharing taxis, or buses, etc. The framework automatically browses social media platforms, distinguishes traffic-related messages, extracts/understands the incidents information, and converts them into alerts in the navigation apps. A joint text processing framework based on fine-tuning BERT classification models for filtering the traffic related information, and either a Question-answering (QA) model or a name-entity recognition (NER) model for extracting real-time traffic details data from social media are developed. In this two-phase NLP pipeline, we first develop a filter that classifies numerous collected text inputs into different groups. Then, we extract necessary information from these filtered groups in order to automatically understand and characterize the reported traffic event from social media by determining its location, its occurrence time, and its nature, e.g., blocked road, accident, etc. Once the detailed traffic information is obtained, we implement an automated system that converts collected posts from social media into real-time traffic information that are used to update navigation maps and warn drivers about any traffic events.
Machine Learning for IoT Service Discovery
Efficiently utilizing the vast network of devices in the Internet-of-Things (IoT) to leverage computational resources can be beneficial for edge computing services. The edge computing affirms to bring the available distributed, but yet closer resources, to the devices requesting external computational and/or storage capabilities. In many cases, the resource sharing and computing capabilities are indispensable in the IoT system, considering many of terminal devices such as sensors, mobile phones, actuators, and personal computers that may lack these resources to accomplish specific tasks. Devices, in the neighborhood, can share their available computational resources to process tasks for the profit of their peers and help relieve the load of cloud and edge servers. Service discovery in IoT platforms can be used to seek for edge computing providers. Despite that, searching in the vast network to find suitable devices remains one of the major challenges in ubiquitous IoT networks, especially when devices are heterogeneous and require a variety of services with different levels of storage and computational capabilities at irregular time instants. To achieve a productive search of mobile edge computers in the large-scale IoT, it is required to find available, trustworthy, and reliable devices that can potentially handle the targeted computational tasks. The emerging concept of Social IoT (SIoT) can help achieve these goals using the social relations built among the devices, aka} social objects. These social relations transform the IoT system into a social network of devices or “friends” having common characteristics and criteria, which can raise the level of security and trustworthiness in such diverse networks.
The ITL team is developing generic frameworks aiming at selecting appropriate edge computers that can handle the computational task of its peer in a trustworthy and rapid manner. The proposed solution reduces the complexity of the service discovery task by shrinking the search space and applying machine learning techniques that do not require recurrent training, which can be very beneficial for large-scale IoT systems.