Optimizing Traffic Management Systems: A Comparative Study of Apache Spark Deployments

March 2024

In today's fast-changing technological world, it is important to understand data quickly, especially when it comes to recognising and responding to incidents. Our recent study addresses this topic, focusing on how a powerful tool called Apache Spark performs in two different setups: one with containers and one directly on physical hardware, referred to as 'bare metal'.

Apache Spark is a kind of super-intelligent data processor. We tested how well it performs when set up in containers (like virtual boxes) with tools like Docker and Kubernetes, and how well it performs when installed directly on the computer hardware without virtual layers ('bare metal'). We wanted to understand this in order to decide how we should set up the technology, especially in situations where rapid incident detection and handling is extremely important.

We found that bare metal configurations provide direct access to the inner workings of the computer and are therefore faster. On the other hand, containerised setups managed by Kubernetes have the advantage of being scalable, resource efficient and easy to use. This difference is crucial for traffic management systems. The ability to change and improve our models quickly is key to keeping our traffic management system work well. Quick model changes mean quicker responses to unexpected traffic incidents, so our management solutions are accurate and timely.

As businesses and researchers work with large amounts of data, the way we set up Apache Spark becomes increasingly important. Using containers helps us to create really efficient dashboards that use smart algorithms to visualise data. This synergy between smart computing and container technologies is particularly useful and opens the door for future dashboard developments and emphasises the need to choose the right setup for different computing needs.

To summarise, our recent study conducted by Frontier Innovations researchers comparing how Apache Spark works in container configurations with bare-metal environments has shown us how to best set up the CONDUCTOR management system. By choosing the right setup that allows us to quickly change and improve our models, we can ensure that our incident detection technology works really well and provides fast and accurate responses. This not only helps us to decide how best to set up our technology, but also helps to improve traffic management technology. It shows how important it is to choose the right technical setup for fast and efficient data analysis and response systems.

Facilitating cross-border airport shuttle trips

February 2024

Slovenia is difficult to reach by plane, as Slovenian airports only offer a few flights and tickets are generally expensive. Many travellers prefer to travel via a nearby international airport, particularly in Italy, Austria or Croatia, and use shuttle services connecting Slovenia with the desired airport to reach a number of international destinations. To facilitate these cross-border trips, GoOpti offers an demand-responsive shuttle service to and from the airport that is synchronised with flight arrivals and departures. Compared to more conventional forms of demand-responsive transport, cross-border airport shuttle services are characterised by one-off requests, highly variable demand in terms of space and time and a high degree of personalisation.

As part of the CONDUCTOR project, we are exploring ways to improve customer satisfaction, vehicle occupancy and schedule quality under uncertainty. As demand is realised over time, future travel conditions become increasingly unpredictable and incidents and accidents can significantly affect travel times and routes. Therefore, services should be optimised a priori based on a prediction of demand and traffic conditions and re-evaluated and refined in real time, not only based on realised ad hoc requests and changing conditions, but also in anticipation of future demand and the emerging traffic situation.

Use case 2 focuses on demand-responsive transport that facilitates cross-border airport shuttle trips from Slovenia to nearby airports and vice versa. Currently, route plans with pick-up and drop-off orders and vehicle assignments are created offline, using rough predictions of travel times and a set of realised requests as inputs. The goal is to incorporate predictive analytics and dynamic routing to enable future demand planning that takes into account the evolution of network states and optimises vehicle occupancy while considering people's needs. Predictive modelling allows us to anticipate demand and fluctuations. Accurate prediction allows us to better manage the transport fleet of the demand- responsive transport service. In developing a demand responsive platform, we aim to utilise the benefits of demand prediction to improve operational efficiency by reducing wastage of resources and calculating fleet idle time.

The Dial-a-Ride Problem with Meeting Points

January 2024

Recent developments in communication technology have significantly increased the reach of Demand-Responsive Transit (DRT) to a wider audience. In the ride-sharing variant of DRT, vehicles often deviate from the most direct route to pick up and drop off passengers, resulting in increased travel times for both operators and passengers. To alleviate this issue, a novel approach proposes the use of meeting points where passengers walk a short distance from their origin to a pick-up location and from their drop-off location to their destination.

At the Transportation Research Board's (TRB) 103rd Annual Meeting in Washington, D.C., researchers introduced an extension of the Dial-A-Ride Problem (DARP) called DARP with Meeting Points (DARPmp). This innovative approach determines the optimal meeting point dynamically during the solution process, eliminating the need for predefined locations for specific requests.

This joint work of the University of Twente and the National Technical University of Athens presents a fleet optimisation model for Shared Demand Responsive Transit that addresses the challenges of pick-up and delivery logistics. The contribution comprises a mixed-integer linear program formulation that considers meeting points, leading to a significant reduction in vehicle operating costs and promoting cost-efficient operations.

Enhancing Mobility-on-Demand Services: Impact of Dynamic Travel Times

December 2023

Mobility-on-Demand (MoD) services have begun to play a vital role in urban mobility in recent years. Given the advances in autonomous vehicles (AVs), AV-based MoD services are also on the horizon. They are expected to offer cheaper fares and potentially take up less parking space. Additionally, carpooling could save total miles travelled.

So far, Automated Mobility on Demand (AMoD) services are still in the pilot phase. Research into the potential benefits of AMoD services uses simulations. These studies usually use deterministic network travel times from historical or simulated data and therefore lack the assessment of the traffic effect that the AV-based MoD service has on the overall system. They also lack a mitigation strategy to ensure that the confirmed ride requests are met even if the estimated vehicle schedule is outdated due to realistic traffic.

To answer the above questions, a bridge is built between Aimsun Next and FleetPy. The former provides a microscopic traffic simulation, while the latter provides AMoD fleet control strategies. The results of a case study in the city of Munich show that the combination of inaccurate travel time estimates and mitigation strategies for infeasible routes significantly deteriorates the performance of AMoD services. Moreover, customers are faced with unreliable pick-up times and travel time estimates. Allowing re-assignments of initial vehicle schedules according to the updated system states helps to restore system efficiency and reliability, but only to a small extent. Read more about our research at https://arxiv.org/abs/2308.05535

Forecasting the traffic flow

November 2023

Efficient traffic management is at the heart of urban living. When commuting runs smoothly, roads are less congested and travel times are reduced. This is possible thanks to ground-breaking advances in traffic flow prediction systems. In many modern cities, managing traffic flow efficiently remains a challenge. However, with accurate prediction of traffic patterns, even without real-time data, this challenge can be overcome. This is presented in a recent study aimed at enhancing existing approaches to traffic management.

The study of an innovative approach that will enhance traffic flow prediction systems presents an efficient solution that combines historical traffic patterns with key factors such as weather conditions and public holidays. By integrating extensive feature engineering, this novel methodology aims to accurately model traffic dynamics. The effectiveness of the study was tested using a new real-world data set from Ljubljana, Slovenia, which provides a new perspective on traffic behaviour. Rigorous analyses of the constructed models revealed promising accuracy and sensitivity to various parameters. An interesting aspect is that our approach offers practical solutions for real-life situations, even without relying on real-time data. It is a step forward in predicting traffic flows, aiming for more efficient transport systems and a better travel experience.

The future of traffic management is evolving, and this study paves the way for smarter, more efficient urban mobility. Read more about our research at https://dx.doi.org/10.1186/s12544-023-00600-6

The CONDUCTOR General Assembly in Bilbao, Spain

September 2023

The CONDUCTOR consortium gathered for the 2nd General Assembly on 28 and 29 September 2023. The meeting took place at the University of Deusto in Bilbao, Spain. We focused on reviewing the progress of the project and planning the next steps for piloting our use cases, while considering the progress in refining methods and models tailored to Cooperative, Connected and Automated Mobility (CCAM). We also discussed our strategies to ensure seamless integration into the platform environment and formulated the methodology for conducting the use cases, addressing potential risks.

The second day was dedicated to the cooperation with the ORCHESTRA project. In a workshop we were introduced to the concept of Multimodal Traffic Management (MTM) and the strategies developed by their project. Following the theoretical lectures, we participated to an interactive board game developed by ORCHESTRA, where the participants put into practice the concepts presented earlier. This allowed us to reconsider a possible integration of the MTM concept into the CONDUCTOR project.

ITSC 2023 conference special session

September 2023

The 26th IEEE International Conference on Intelligent Transportation Systems (ITSC 2023) was held in Bilbao, Spain, from 24 to 28 September 2023. At the conference, the CONDUCTOR and TANGENT projects, represented by Jožef Stefan Institute, University of Deusto and National technical university of Athens, joined forces to organise a special session on "Advanced traffic and network management" on 25 September 2023.

The objective of the special session was aligned with emergence of novel information and communication technologies and computer-based tools to create new opportunities to rethink existing transportation problems and propose innovative solutions to tackle the challenges. From a traffic and network management perspective, this special session focused on presenting operational concepts for an efficient, resilient and adaptable multimodal network and traffic management systems that use innovative techniques for data collection and fusion and take advantage of state-of-the-art technologies (e.g., artificial intelligence, high-performance computing). More specifically, the following papers were presented:

  • Integrated Signal and Bus Lane Control: An Intersection-Level Approach by Emmanouil Kampitakis, Konstantinos Katzilieris and Eleni Vlahogianni.
  • A Taxonomy of Signal Vehicle Coupled Control from a Mathematical Programming Perspective by Arka Ghosh, Juan Angarita-Zapata, Antonio David Masegosa and Ander Arriandiaga.
  • AI Assisted Dynamic Bus Lane Control in Connected Urban Environments: The Case of Intermittent Dynamic Bus Lanes by Konstantinos Katzilieris, Emmanouil Kampitakis and Eleni Vlahogianni.
  • The Pickup and Delivery Problem with Crossdock for Perishable Goods by Konstantinos Gkiotsalitis and Nikolopoulou Amalia.

Handbook on Transport Pricing and Financing

September 2023

CONDUCTOR envisions a world where connected and automated vehicles (CAVs) are seamlessly integrated with advanced, high-level traffic and fleet management solutions, enabling efficient and globally optimal transportation for passengers and goods. This shift towards CAVs is having a direct impact on the pricing of transportation services which needs to be considered and assessed.

Assoc. Prof. Alejandro Tirachini from the Department of Transport Engineering and Management at the University of Twente, recently released a "Handbook on Transport Pricing and Financing", including a chapter on "Connected and automated vehicles: effects on pricing".

The comprehensive book focuses on transport pricing and financing, two crucial factors with far-reaching socio-economic implications. The handbook sheds light on how pricing goes beyond financial aspects and aims at economic efficiency. You will discover the potential of new technologies such as digital payment channels, dynamic pricing techniques and information provision shaping the future of transportation.

Chapter 13 reviews the theory of transport pricing theory and its relation to automated vehicles. It discusses the expected benefits and drawbacks of automated vehicle technologies and analyses their impact on pricing-related aspects. The findings suggest that automation can lower optimal fares for private cars and public transport, making them more accessible.

Detailed use-case specifications and their key performance indicators

July 2023

The project's use case models describe the functional requirements of a system and specify how the system under consideration should interact with the user and other actors. Such specifications benefit further development and stimulate discussion between stakeholders, as the use cases are easy to understand and follow even without detailed knowledge. In close cooperation with the stakeholders, we took an iterative approach to create the use cases. We identified different scenarios for each pilot project, describing their trigger conditions, their main process flow and their termination conditions.

Recent research and reports highlight a shifting paradigm in the measurement of progress, emphasising the need to evaluate both positive and negative impacts on the economy, environment and society, taking into account differences between user groups. With this in mind, we present a comprehensive multidimensional framework to monitor and assess the impact of the functionalities of CONDUCTOR in the pilot cases. The framework considers four groups or dimensions of key performance indicators (KPIs): technical, economic, environmental and social. It also takes into account local challenges by considering use case and site-specific indicators. Although the groups are not directly transferable to other sites, they ensure an assessment along the dimensions of sustainable development.

Read more about these specifications and the KPIs on our website.

Stakeholders’ requirements, user needs and social innovations

June 2023

The CONDUCTOR report aiming to design, integrate, and demonstrate advanced traffic and fleet management for the efficient and optimal transport of passengers and goods, has been published. This report includes a careful analysis of the use cases covered by CONDUCTOR, a set of general regulatory and social requirements, and stakeholders and users needs collected from a state-of-the art review and a dedicated survey.

The CONDUCTOR consortium has developed an integrated methodology represented by a double funnel diagram. The methodology consists of:

  • A top-down approach which gathers formalised and general input from regulators and public bodies.
  • A bottom-up approach that facilitates the collection of service-specific inputs derived from user needs and the interests of other stakeholders.

The results of both the bottom-up and top-down analysis were jointly reviewed by the CONDUCTOR consortium, taking into account the input gathered during the workshop held on the 22 March 2023. The in-depth examination of the use cases produced more insightful results that highlighted the primacy of certain conditions for the specific purposes of CONDUCTOR.

The readers will find insights and suggestions for a profitable use of the results for the design process of the use cases and pilots.

The full document is available on our website.

CONDUCTOR project at EUCAD 2023 conference

May 2023

The CONDUCTOR project was invited to join the 4th European Conference on Connected and Automated Driving (EUCAD 2023) organised by the European Commission in Brussels, Belgium on 3 and 4 May 2023. The organisation team recognised the importance of CONDUCTOR project for bridging the gap between the development and deployment of Cooperative, Connected and Automated Mobility (CCAM) technologies.

Our CONDUCTOR team, consisting in representatives from the partners Netcompany-Intrasoft SA, BAX Innovation Consulting SL and Deep Blue SRL, participated to the high-level policy plenary sessions and the thematic technical breakout sessions.

In addition, CONDUCTOR was offered the opportunity to promote the project’s activities, objectives and expected results during the Exhibition session organised on the 3 May 2023 at the Autoworld Museum. The CONDUCTOR booth was perfectly located, and we received positive feedback from all the participants who stopped by our stand.

CONDUCTOR in the CINEA Brochure

May 2023

The European Climate, Environment and Infrastructure Executive Agency (CINEA) published project brochure on Cooperative, Connected and Automated Mobility (CCAM). CCAM technologies offer substantial benefits, including improved accessibility for individuals who are unable to drive, increased road safety and traffic efficiency, along with reduced costs and emissions. These innovations provide economic opportunities, creating businesses, jobs and specialised skills. Cluster 5 within the Horizon Europe Programme highlights its funding emphasis on CCAM to enhance safety, accessibility, and sustainability in transportation. The brochure provides a comprehensive overview of the CCAM projects funded under the Horizon Europe Programme. The projects are listed according to different thematic areas, i.e. data ecosystems, infrastructure support, demonstration, socio-economic and environmental aspects, vehicle technologies, validation methodologies and coordination activities. CONDUCTOR is among the projects that provide infrastructure support to enable the integration of CCAM vehicles into the transportation system.

The CONDUCTOR project is presented on page 21 of the CINEA brochure.

ITSC 2023 special session

March 2023

We are organising the special session on "Advanced traffic and network management" at the IEEE International Conference on Intelligent Transportation Systems ITSC 2023 in Bilbao, Spain.

The transition to an advanced multimodal transport system requires better coordination of traffic. Recent developments in CCAM and evolving urban mobility patterns are challenging existing practises to reduce congestion, save time and advance connected and automated mobility, new services and multimodal transport integration. This special session focuses on efficient, resilient and adaptive multimodal network and traffic management concepts that utilise advanced data collection, fusion techniques and state-of-the-art technologies such as artificial intelligence and high performance computing. We look forward to receiving submissions on network and traffic management research initiatives in Europe and worldwide, covering topics such as connected and autonomous driving systems, deep learning for traffic management, optimisation methods, travel behaviour modelling, smart road infrastructures and more.

Papers can be submitted through the ITSC2023 submission system (the special session’s code is a16f7).

Submission deadline: 15 May 2023

The CONDUCTOR General Assembly in Rome, Italy

March 2023

A two-day CONDUCTOR General Assembly was organised by Deep Blue SRL in Rome, Italy, on 21 and 22 March 2023. On the first day, the initial progress of the project was reported and discussed. A significant part of our discussions was centered around the planned data sources, their technical descriptions and the degree of their availability. The results of the survey, filled by 58 relevant stakeholders from the logistics sector, public authorities and road operators, were presented and especially the data related to their regulatory needs, requirements, benefits, expected bottlenecks and challenges. The findings from the survey were transformed into insightful graphical representations. These were discussed during the meeting and are further detailed in the project Report on stakeholders’ requirements, user needs and social innovations.

On the second day of the meeting, we organised two workshops. The "Stakeholders’ inputs and use cases" workshop aimed to discuss the collected needs and requirements and map these and the associated key performance indicators to the use cases of the CONDUCTOR project. The "System Architecture and Data Sources" workshop provided insights into the architecture components, their interactions, and the data specifics for each use case.

The CONDUCTOR project was launched

November 2022

The CONDUCTOR project officially started on 1 November 2022 and will run for three years. The project aims to design, integrate, and demonstrate advanced high-level traffic and fleet management to allow efficient and globally optimal transport of passengers and goods while ensuring seamless multimodality and interoperability. To achieve this goal, the project has been built around a strong multi-disciplinary consortium of fifteen experienced partners in the field of traffic and fleet management.

The CONDUCTOR consortium gathered on 22 and 23 November 2022 during the official kick-off meeting organised by the project Coordinator Netcompany-Intrasoft SA in Athens, Greece. The kick-off meeting was the occasion for the partners to meet each other and discuss the practical implementation of CONDUCTOR project’s activities and the achievements of its objectives via three use cases.

The consortium particularly appreciated the participation and the presentation made by our Project Officer from the European Climate Infrastructure and Environment Executive Agency (CINEA) who showed genuine interest and support in our CONDUCTOR project.

If you do not want to miss the latest news and up-to-date information relevant to our CONDUCTOR project, visit frequently our website and follow us on social media.