Periodic Reporting for period 2 - CONDUCTOR (Fleet and traffic management systems for conducting future cooperative mobility)
Reporting period: 2024-05-01 to 2025-10-31
The CONDUCTOR project’s main goal is to design, integrate and demonstrate advanced, high#level traffic and fleet management that will allow efficient and globally optimal transport of passengers and goods, while ensuring seamless multi#modality and interoperability. Using innovative dynamic balancing and priority-based management of vehicles (automated and conventional) CONDUCTOR will build upon state-of-the-art fleet and traffic management solutions in the CCAM ecosystem and develop next generation simulation models and tools enabled by machine learning and data fusion, enhancing the capabilities of transport authorities and operators, allowing them to become “conductors” of future mobility networks. We will upgrade existing technologies to place autonomous vehicles at the centre of future cities, allowing heightened safety and flexible, responsive, centralized control able to conduct traffic and fleets at a high level. These innovations will lead to less urban traffic and congestion, lowered pollution, and a higher quality of life. Project innovations will be integrated into a common, open platform, and validated in three use cases, testing the interoperability of traffic management systems and integration of different transportation means for both people and goods. Use case UC1 integrates traffic management with inter-modality, UC2 tests demand-response transport, and UC3 urban logistics. In each use case and its demonstrations, simulations will be validated through real life data.
The CONDUCTOR designed, integrated and demonstrated innovative traffic and fleet management concepts to improve the transport of passengers and goods, while considering the integration of CCAM vehicles into the mobility ecosystem. The developed solutions consist in a variety of advanced traffic management models, and several relevant scenarios ensuring seamless multimodality and interoperability. These models have being integrated and deployed within a tailor-made platform that generated advanced simulation and compared the results with current systems and solutions within three Uses Cases and five Pilots. During the 36-month project, the consortium developed a comprehensive portfolio of traffic, fleet, simulation, and governance models, supported by robust analytical and data-driven methods such as:
1/ The Cooperative Traffic Management System is a multi-agent reinforcement learning-based traffic controller combining signal control with dynamic bus-lane allocation via LED/VMS systems.
2/ A cloud-native Demand-Responsive Platform (DRP) providing AI-powered demand prediction, continuous planning, and ad-hoc planning.
3/ Fleet Management System includes modules for incident management, headway and performance monitoring, and layover management with ETA-based predictions.
4/ Ride–Parcel Pooling integrates passenger and parcel transport with three levels of integration, featuring soft delivery time windows and agent-based simulation coupled with Aimsun traffic models for realistic assessment of combined operations.
5/ The Pickup and Delivery with Crossdock for Perishable Goods family includes EV variants and dynamic rolling-horizon approaches for real-time logistics optimisation.
6/ For multimodal coordination, the Vehicle Scheduling Model for Autonomous and Connected Vehicles optimises trunk–feeder bus coordination through dynamic dispatching and holding strategies.
7/ A stochastic transit assignment model captures adaptive passenger route choice under uncertainty using real-time information, enabling scenario-dependent demand prediction and real-time rebalancing decisions.
8/ The Multimodal Journey Planner and MSP Controller integrates a reactive routing toolset with real-time mode availability, vehicle status, and shared mobility service management.
9/ Analytical and optimisation methods include data fusion and harmonization pipelines combining mobile-network data, loop detectors, logistics flows, and DRT requests; AI-based anomaly detection of unusual traffic conditions; dynamic optimization algorithms, etc.
1/ The Cooperative Traffic Management System is a multi-agent reinforcement learning-based traffic controller combining signal control with dynamic bus-lane allocation via LED/VMS systems.
2/ A cloud-native Demand-Responsive Platform (DRP) providing AI-powered demand prediction, continuous planning, and ad-hoc planning.
3/ Fleet Management System includes modules for incident management, headway and performance monitoring, and layover management with ETA-based predictions.
4/ Ride–Parcel Pooling integrates passenger and parcel transport with three levels of integration, featuring soft delivery time windows and agent-based simulation coupled with Aimsun traffic models for realistic assessment of combined operations.
5/ The Pickup and Delivery with Crossdock for Perishable Goods family includes EV variants and dynamic rolling-horizon approaches for real-time logistics optimisation.
6/ For multimodal coordination, the Vehicle Scheduling Model for Autonomous and Connected Vehicles optimises trunk–feeder bus coordination through dynamic dispatching and holding strategies.
7/ A stochastic transit assignment model captures adaptive passenger route choice under uncertainty using real-time information, enabling scenario-dependent demand prediction and real-time rebalancing decisions.
8/ The Multimodal Journey Planner and MSP Controller integrates a reactive routing toolset with real-time mode availability, vehicle status, and shared mobility service management.
9/ Analytical and optimisation methods include data fusion and harmonization pipelines combining mobile-network data, loop detectors, logistics flows, and DRT requests; AI-based anomaly detection of unusual traffic conditions; dynamic optimization algorithms, etc.
ONDUCTOR advances the state of the art in several dimensions:
* CCAM‑enabled corridor management and multimodal synchronisation: The CTMS and VSMACV models go beyond conventional fixed‑time or traffic‑actuated control by jointly optimising lane allocation, signal timing and transfer synchronisation using multi‑agent and multi‑objective reinforcement learning, with person‑based and emission KPIs. This allows explicit negotiation between public transport, freight and general traffic priorities rather than fixed, pre‑weighted control rules.
* Integration of DRT, CCAM and logistics: The combination of cloud‑native DRT platforms, AI‑based demand prediction and continuous planning, ride–parcel pooling models and dynamic PDPCDPG optimisation provides a novel blueprint for integrating passenger and freight services, enabling more efficient use of vehicles, lower empty‑kilometres and improved service economics in urban and corridor contexts.
* Interoperability and data spaces for CCAM: The CCAM Interoperability Framework, Simulation Orchestrator and mobility dataspace architecture operationalise European priorities on interoperability and data sovereignty. By aligning with GAIA‑X/IDSA concepts and exposing a rich set of standardised APIs, CONDUCTOR shows how heterogeneous CCAM, traffic and fleet‑management services can be orchestrated across tools, operators and cities.
* Behaviour‑aware governance and policy tools: The combination of an EU‑wide AV acceptance model, multi‑criteria governance controllers and pilot‑tested governance schemes allows authorities to reason explicitly about trade‑offs between safety, emissions, equity and service quality, and to design CCAM deployments that are both technically efficient and socially acceptable.
These advances underpin a portfolio of 18 Key Exploitable Results covering optimisation and predictive analytics, traffic and network simulation, and interoperability and data governance. The results are designed to be scalable and reusable across European cities and corridors, supporting future deployment of CCAM services, multimodal traffic management and mobility data spaces in line with EU climate, digital and transport objectives.
* CCAM‑enabled corridor management and multimodal synchronisation: The CTMS and VSMACV models go beyond conventional fixed‑time or traffic‑actuated control by jointly optimising lane allocation, signal timing and transfer synchronisation using multi‑agent and multi‑objective reinforcement learning, with person‑based and emission KPIs. This allows explicit negotiation between public transport, freight and general traffic priorities rather than fixed, pre‑weighted control rules.
* Integration of DRT, CCAM and logistics: The combination of cloud‑native DRT platforms, AI‑based demand prediction and continuous planning, ride–parcel pooling models and dynamic PDPCDPG optimisation provides a novel blueprint for integrating passenger and freight services, enabling more efficient use of vehicles, lower empty‑kilometres and improved service economics in urban and corridor contexts.
* Interoperability and data spaces for CCAM: The CCAM Interoperability Framework, Simulation Orchestrator and mobility dataspace architecture operationalise European priorities on interoperability and data sovereignty. By aligning with GAIA‑X/IDSA concepts and exposing a rich set of standardised APIs, CONDUCTOR shows how heterogeneous CCAM, traffic and fleet‑management services can be orchestrated across tools, operators and cities.
* Behaviour‑aware governance and policy tools: The combination of an EU‑wide AV acceptance model, multi‑criteria governance controllers and pilot‑tested governance schemes allows authorities to reason explicitly about trade‑offs between safety, emissions, equity and service quality, and to design CCAM deployments that are both technically efficient and socially acceptable.
These advances underpin a portfolio of 18 Key Exploitable Results covering optimisation and predictive analytics, traffic and network simulation, and interoperability and data governance. The results are designed to be scalable and reusable across European cities and corridors, supporting future deployment of CCAM services, multimodal traffic management and mobility data spaces in line with EU climate, digital and transport objectives.