The main goal of this paper is to present an end-to-end, data-driven framework for the control of Autonomous Electric Vehicles (AEV) for Mobility-on-Demand (MoD). We present a data-driven Model Predictive Control (MPC) algorithm that rebalances (i.e. preemptively repositions) the AEV fleet in order to meet the mobility demand in the near future. The algorithm consists of Mixed Integer Linear Programming (MILP) that leverages the short-term forecast of the mobility demand as well as the charging station availability in order to optimally rebalance the AEV fleet. The proposed algorithm is evaluated by means of simulations with the New York City (NYC) taxi data. The proposed algorithm outperforms other state-of-the-art rebalancing strategies by reducing the mean customer wait time by 82.3% and the number of rejected requests by 94.6% for a given fleet size and a number of charging stations.

Applying simulation-based optimization to city-scale traffic signal optimization can be challenging due to the large search space resulting in high computational complexity. A divide-and-conquer approach can be used to partition the problem and optimized separately, which leads to faster convergence. However, the lack of coordination among the partial solutions may yield a poor-quality global solution. In this paper, we propose a new method for simulation-based optimization of traffic signal control, called spatially iterative coordination for parallel optimization (SICPO), to improve coordination among the partial solutions and reduce synchronization between the partitioned regions. The traffic scenario is simulated to obtain the interactions, which is used to spatially decompose the scenario into regions and identify interdependencies between the regions. Based on the regions, the problem is divided into subproblems which are optimized separately. To coordinate between the subproblems, the interactions between partial solutions are synchronized in two ways. First, multiple iterations of the optimization process can be executed to coordinate the partial solutions at the end of each optimization process. Second, the partial solutions can also be coordinated among the regions by synchronizing the trips across the regions. To reduce computational complexity, parallelism can be applied on two levels: each region is optimized concurrently, and each solution for a region is evaluated in parallel. We demonstrate our method on a real-world road network of Singapore, where SICPO converges to an average travel time 21.6% faster than global optimization at 62.8× shorter wall-clock time.

Converting OpenStreetMap (OSM) data to a road network suitable for microscopic traffic simulation keeps being a challenging task: both missing information and excessive details, as well as wrong typologies present in the dataset frequently confuses automatic converters. In this paper, we present a method along with a reference implementation, Traffic Simulation Map Maker (TSMM), which aims at substantially increasing the automation level of road network prototyping by simplifying the OSM data while preserving important topology information. The main objective of this work is to enable the study of traffic simulation dynamics at scale using real-world road networks, while minimizing the need for solving the long tail of problems related to the road network generation. Our proposed approach yields what we believe is a good trade-off between precision and automation, making bold yet acceptable decisions that solve most of the errors at the source, i.e., the map. While there is definitely a loss in fidelity with respect to the real world, many properties of the road network are preserved. We argue that TSMM greatly improves the availability of arbitrarily large and usable road networks on top of available OSM maps by reducing the complexity for conversion tools and traffic simulation researchers alike. A proof-of-concept study using OSM data from Binjiang, China, demonstrates that TSMM is able to generate a road network with well-preserved topological information which avoids the many errors and deadlocks that occur when building the network using the original input sources.

Programming on heterogeneous hardware architectures using OpenCL requires thorough knowledge of the hardware. Many High-Performance Domain-Specific Languages (HPDSLs) are aimed at simplifying the programming efforts by abstracting away hardware details, allowing users to program in a sequential style. However, most HPDSLs still require the users to manually map compute workloads to the best suitable hardware to achieve optimal performance. This again calls for knowledge of the underlying hardware and trial-and-error attempts. Further, very often they only consider an offloading mode where compute-intensive tasks are offloaded to accelerators. During this offloading period, CPUs remain idle, leaving parts of the available computational power untapped. In this work, we propose a tool named OptCL for existing HPDSLs to enable a heterogeneous co-execution mode when capable where CPUs and accelerators can process data simultaneously. Through a static analysis of data dependencies among compute-intensive code regions and performance predictions, the tool selects the best execution schemes out of purely CPU/accelerator execution or co-execution. We show that by enabling co-execution on dedicated and integrated CPU-GPU systems up to 13 × and 21 × speed-ups can be achieved.

In mixed traffic, the popularity of public transport (PT) is still affected by relatively low operating speeds compared to private vehicles. To overcome this, PT priority measures have been proposed and adopted extensively. However, existing solutions such as exclusive bus lanes or traffic signal priorities are often limited in terms of available road space or large-scale feasibility. In this paper, we propose a Vehicle-to-Vehicle/Infrastructure (V2X)-based dynamic PT priority concept in mixed traffic called Virtual Right of Way (VROW). Private vehicles in front of a PT vehicle make spaces through collaborative lane changes within a dynamic clearing distance computed based on the current traffic situation. This allows a more efficient allocation of road space while still maintaining a high level of PT priority. In this paper, we evaluate the potential traffic impacts of VROW on both PT and private vehicles by conducting microscopic traffic simulations within a small urban network and a highway scenario. Comparisons with mixed traffic and other existing bus lane priority strategies, in terms of operation and safety concerns, are analyzed and highlighted. Simulation results show that VROW improves the PT operational performance with only a marginal influence on private vehicles measured by their average travel time and the number of lane changes.

In an agent-based simulation (ABS), a state update scheme carries out the transitions of agents from one state to the next. To produce correct simulation results, the update scheme must respect the cause-and-effect relationships defined by the agent-based model and ensure that the resulting overall simulation state is internally consistent. At the same time, the update scheme should be efficient enough to meet a simulationist's demand for timely results. Considering the common class of synchronous time-driven ABS, a number of update schemes have been employed in the literature and simulation frameworks. In this paper, various implementations of update schemes are analyzed and contrasted with respect to their ability to maintain the simulation correctness as well as their performance characteristics. A semantic model is formulated to define the reference behavior of synchronous time-driven ABS updates and model the dependencies among agent updates using a state access graph. Relying on the formalization, conditions under which different update schemes achieve causality are shown. Further, resolution methods are categorized according to their coordination mechanisms to achieve consistency by resolving conflicts among agent state updates. Through two case studies, the empirical performance of different update schemes and resolution methods are evaluated. For sequential execution, an update scheme based on the agent's dependencies achieves the highest performance, whereas in the parallel case, the choice of update scheme involves a tradeoff between execution time and memory usage. If deterministic simulation output is required, decentralized coordination generally outperforms centralized coordination. The results can assist implementers and researchers in their selection of appropriate methods in the design and implementation of agent-based simulators.

Mass Rapid Transit (MRT) systems, i.e. subway systems, all over the world, are experiencing an increase in ridership. This also means that in case of an unplanned MRT service disruption, the number of affected passengers is larger, requiring a fast and comprehensive response. In this paper, we thus study the disruption management of MRT systems. We develop an optimization model to identify the optimal bridging plan in response to an MRT disruption, so that the negative effects of a disruption could be minimized. Our approach supports deployment of multiple types of bridging buses, reflecting the diversity of vehicle types in a typical public transportation provider’s vehicle fleet. The optimization objective of our approach is to decrease the travel delay of passengers and increase the number of passengers who can be served. We demonstrate the effectiveness of our approach on a hypothetical case study in the central business district of Singapore. Moreover, we validate our analytical results with microscopic simulation, showing that our simplified analytical optimization approach can be used for disruption response planning. Some deviations indicate, however, that a combined simulation and optimization approach yields better results to obtain an effective bridging plan.

In this article, we propose a $k$ -anonymity approach that prioritizes the generalization of attributes based on their utility. We focus on transport data, which we consider a special case in which many or all attributes are quasi-identifiers (e.g., origin, destination, ride start time), as they allow correlation with easily observable auxiliary data. The novelty in our approach lies in introducing normalization techniques as well as distance and utility metrics that allow the consideration of not only numerical attributes but also categorical attributes by representing them in tree or graph form. The prioritization of the attributes in the generalization process is based on the attributes’ utility and can further be influenced by either automatically or manually assigned attribute weights. We evaluate and compare different options for all components of our mechanism as well as present an extensive performance evaluation of our approach using real-world data. Lastly, we show in which cases suppression of records can counter-intuitively lead to higher data utility.

This paper introduces the a framework that simplifies the process of designing and describing autonomous vehicle platooning manoeuvres which implements four design principles: Standardisation, Encapsulation, Abstraction, and Decoupling (SEAD). Although a large body of research has been formulating platooning manoeuvres, it is still challenging to design, describe, read, and understand them. This difficulty largely arises from missing formalisation. To fill this gap, we analysed existing ways of describing manoeuvres, derived the causes of difficulty, and designed a framework that simplifies the manoeuvre design process. Alongside, a Manoeuvre Design Language was developed to structurally describe manoeuvres in a machine-readable format. Unlike state-of-the-art manoeuvre descriptions that require one state machine for every participating vehicle, the SEAD framework allows describing any manoeuvre from the single perspective of the platoon leader. We hope that the SEAD framework will pave the way for further research in the area of new manoeuvre design and optimisation by largely simplifying and unifying platooning manoeuvre representation.

The future of transportation will be shared, autonomous, and electric, leading to a paradigm shift in the transportation domain. This shift comes with opportunities and challenges for the electric power system. For example, the charging of large shared autonomous electric vehicle fleets can be controlled in a centralized way and these vehicles can also provide ancillary services such as demand response or frequency regulation for the electric power system. In this paper, we give an overview of the research residing on this boundary between transportation systems and electric power systems. We discuss shortcomings and identify potential topics of research that we believe need to be addressed to foster the successful integration of shared electric vehicle fleets into the power system.

A key challenge in the implementation of novel public transport systems is to maintain usability over a broad spectrum of potential users. Transport systems that increasingly emphasise dynamic adjustment to changing passenger numbers and destinations over time cannot rely on static schedules and routes like traditional systems do. In this work we are investigating the use of agent-based crowd simulation to evaluate how different passenger guidance systems affect agent navigation in a public transport hub. We study the effects of different digital signage placement strategies in terms of crowding and walking times and also analyse how the introduction of mobile phone guidance systems affects these metrics. Our results show that crowd simulation is a cost and time-efficient tool for the evaluation of guidance systems in public transport spaces that can also support the design of bus schedules and bay assignments.

The performance of Agent-based Traffic Simulations (ABTS) has been shown to benefit tremendously from offloading to accelerators such as GPUs. In the search for the most suitable hardware platform, reconfigurable hardware is a natural choice. Some recent work considered ABTS on Field-Programmable Gate Arrays (FPGAs), yet only implemented simplified cellular automaton-based models. The recent introduction of support for high-level synthesis from C, C++, and OpenCL in FPGA tool chains allows FPGA designs to be expressed in a form familiar to software developers. However, the performance achievable with this approach in a simulation context is not well-understood. In this work, to the best of our knowledge, we present the first FPGA-accelerated ABTS based on widely-accepted microscopic traffic simulation models, and the first to be generated from high-level code. The achieved speedup of up to 24.3 over a sequential CPU-based execution indicates that recent FPGA toolchains allow simulationists to unlock the performance benefits of reconfigurable hardware without the need to express the simulation models in low-level hardware description languages.

This paper explores the interaction between autonomous and human-driven cars on a microscopic level using an agent-based traffic simulator. More specifically, it deals with the design of driving logic models of "socially-aware" autonomous vehicles that can improve the performance of surrounding vehicles on the road. Congestion waves, which are created as a result of an abrupt stopping or a car joining a highway, are a known phenomenon in current traffic systems. Experiments performed, demonstrate how the presence of intelligent social vehicles on the road can reduce such effects by acting as a flexible medium between human-driven cars. Metrics to evaluate benefits ot our AV behaviour models under various states of traffic conditions/congestion are also proposed. Finally, results showing the effectiveness of these models are presented.

State fast-forwarding has been proposed as a method to reduce the computational cost of microscopic traffic simulations while retaining per-vehicle trajectories. However, since fast-forwarding relies on vehicles isolated on the road, its benefits extend only to situations of sparse traffic. In this paper, we propose fast-forwarding of vehicle clusters by training artificial neural networks to capture the interactions between vehicles across multiple simulation time steps. We explore various configurations of neural networks in light of the trade-off between accuracy and performance. Measurements in road network simulations demonstrate that cluster fast-forwarding can substantially outperform both time-driven state updates and single-vehicle fast-forwarding, while introducing only a small deviation in travel times.

Common car-following models for microscopic traffic simulation assume a time advancement using fixed-sized time steps. However, a purely time-driven execution is inefficient when the states of some agents are independent of other agents and thus predictable far into the simulated future. We propose a method to accelerate microscopic traffic simulations based on identifying independence among agent state updates. Instead of iteratively updating an agent’s state throughout a sequence of time steps, a computationally inexpensive “fast-forward” function advances the agent’s state to the time of its earliest possible interaction with other agents. We present an algorithm to determine independence intervals in microscopic traffic simulations and derive fast-forward functions for several well-known traffic models. In contrast to existing approaches based on reducing the level of detail, our approach retains the microscopic nature of the simulation. An evaluation is performed for a synthetic scenario and on the road network of Singapore. At low traffic densities, maximum speedup factors of about 2.6 and 1.6 are achieved, while at the highest considered densities, only few opportunities for fast-forwarding exist. We show that the deviation from purely time-driven execution is reduced to a minimum when choosing an adequate numerical integration scheme to execute the time-driven updates. Verification results show that the overall deviation in vehicle travel times is marginal.

Microscopic traffic simulation is an essential tool for the evaluation of intelligent transportation systems (ITS). With the increasing complexity of ITS applications, higher-detail simulation models, and the need to analyze large-scale scenarios, simulation run-times can grow exceedingly large. One way to counter this problem is the use of parallel computing techniques, such as shared-memory multi-thread parallelism. While the foundations of parallel traffic simulation are well-known, the effects of different synchronization and agent-update mechanisms on simulation performance have not been explored systematically. In this paper, we first analyze the common properties of models used in microscopic traffic simulation to understand the impact of their data dependencies. We discuss synchronous and asynchronous agent update schemes and compare them in terms of performance and requirements. We conclude that although it requires more memory and additional conflict handling, the synchronous agent-state updating approach is favourable in terms of scalability.

The execution of agent-based simulations (ABSs) on hardware accelerator devices such as graphics processing units (GPUs) has been shown to offer great performance potentials. However, in heterogeneous hardware environments, it can become increasingly difficult to find viable partitions of the simulation and provide implementations for different hardware devices. To automate this process, we present OpenABLext, an extension to OpenABL, a model specification language for ABSs. By providing a device-aware OpenCL backend, OpenABLext enables the co-execution of ABS on heterogeneous hardware platforms consisting of central processing units, GPUs, and field programmable gate arrays (FPGAs). We present a novel online dispatching method that efficiently profiles partitions of the simulation during run-time to optimize the hardware assignment while using the profiling results to advance the simulation itself. In addition, OpenABLext features automated conflict resolution based on user-specified rules, supports graph-based simulation spaces, and utilizes an efficient neighbor search algorithm. We show the improved performance of OpenABLext and demonstrate the potential of FPGAs in the context of ABS. We illustrate how co-execution can be used to further lower execution times. OpenABLext can be seen as an enabler to tap the computing power of heterogeneous hardware platforms for ABS.