Autonomous Mobility-on-demand Research Articles

Designing an Autonomous Mobility-on-Demand Service for Transit Last-Mile Access

Public transportation has the potential to provide a safe, convenient, affordable, and eco-friendly mobility service. However, because of its fixed routes and limited network coverage, it is sometimes difficult or impossible for passengers to walk from a transit stop to their destination. This inaccessibility problem is also known as the “transit last-mile connectivity problem.” Such a lack of connectivity can force travelers to drive, consequently increasing vehicle ownership and vehicle miles traveled (VMT) on the roads. The autonomous mobility-on-demand (AMoD) service, with characteristics such as quick fleet repositioning and demand responsiveness, has the potential to provide last-mile coverage where fixed-route transit can only provide limited coverage. This study aims to address the last-mile problem by developing an online demand-responsive AMoD service integrated with fixed-route transit. A linear time-delay dynamical system is proposed to model the AMoD system, and a model predictive control methodology is adopted to regulate the system around an equilibrium point with minimum vehicle rebalancing. To assess the impact of this new mobility service on travel demand, a simulation study is developed and integrated with a mode choice model capturing a combined transit-AMoD model. The experiment results reveal the potential to enhance transit efficiency, reduce VMT, and accommodate the increased transit demand while maintaining the quality of service.

Charging autonomous electric vehicle fleet for mobility-on-demand services: Plug in or swap out?

This paper compares two prevalent charging strategies for electric vehicles, plug-in charging and battery swapping, to investigate which charging strategy is superior for electric autonomous mobility-on-demand (AMoD) systems. To this end, we use a queueing-theoretic model to characterize the vehicle waiting time at charging stations and battery swapping stations, respectively. The model is integrated into an economic analysis of the electric AMoD system operated by a transportation network company (TNC), where the incentives of passengers, the charging/operating shift of TNC vehicles, the operational decisions of the platform, and the planning decisions of the government are captured. Overall, a bi-level optimization framework is proposed for charging infrastructure planning of the electric AMoD system. Based on the proposed framework, we compare the socio-economic performance of plug-in charging and battery swapping, and investigate how this comparison depends on the evolving charging technologies (such as charging speed, battery capacity, and infrastructure cost). At the planning level, we find that when choosing plug-in charging, increased charging speed leads to a transformation of infrastructure from sparsely distributed large stations to densely distributed small stations, while enlarged battery capacity transforms the infrastructure from densely distributed small stations to sparsely distributed large stations. On the other hand, when choosing battery swapping, both increased charging speed and enlarged battery capacity will lead to a smaller number of battery swapping stations. At the operational level, we find that improved charging speed leads to increased TNC profit when choosing plug-in charging, whereas improved charging speed may lead to smaller TNC profit under battery swapping. The above insights are validated through realistic numerical studies.

Autonomous mobility on demand: from case studies to standardized evaluation

We present an overview of ten case studies of Autonomous Mobility on Demand (AMoD) transportation systems, which are based on realistic data from different urban contexts. Comparing AMoD systems with Conventionally Driven Vehicles (CDV), the limits of reduction of vehicles, the cutting-back of parking spaces, and the increase of empty miles are investigated. As a result of introducing a shared fleet of autonomous vehicles (AV), the analysis demonstrated that 88%–93% of CDV are not required to meet realistic requirements. Parking spaces can be reduced by 83%–97%, while empty miles could be increased by 6%–15%. Nonetheless, fleet dispatching techniques that use the advanced optimization algorithms can reduce the ratio of empty miles by as much as 40%. Consequently, we propose a standard procedure for conducting intelligent transportation system studies (ITS) that can assist in the planning of traffic on urban environments at operational, tactical, and strategic levels. Furthermore, the case studies enabled us to design an Intelligent Transportation System Readiness Level (ITS-RL) scale to assess the realism of case studies, facilitate risk assessment, and provide guidance on how to incorporate AMoD system within a local context.

Coordinated flow model for strategic planning of autonomous mobility-on-demand systems

ABSTRACT High-quality strategic planning of autonomous mobility-on-demand (AMOD) systems is critical for the success of the subsequent phases of AMOD system implementation. To assist in strategic AMOD planning, we propose a dynamic and flexible flow-based model of an AMOD system. The proposed model is computationally fast while capturing the state transitions of two coordinated flows (i.e. co-flows): the AMOD service fleet vehicles and AMOD customers. Capturing important quantity dynamics and conservations through a system of ordinary differential equations, the model can economically respond to a large number and a wide range of scenario-testing requests. The paper illustrates the model efficacy through a basic example and a more realistic case study. The case study envisions replacing Manhattan's existing taxi service with a hypothetical AMOD system. The results show that even a simple co-flow model can robustly predict the systemwide AMOD dynamics and support the strategic planning of AMOD systems.

A predictive chance constraint rebalancing approach to mobility-on-demand services

This paper considers the problem of supply-demand imbalances in Mobility-on-Demand (MoD) services. These imbalances occur due to uneven stochastic travel demand and can be mitigated by proactively rebalancing empty vehicles to areas where the demand is high. To achieve this, we propose a method that takes into account uncertainties of predicted travel demand while minimizing pick-up time and rebalance mileage for autonomous MoD ride-hailing. More precisely, first travel demand is predicted using Gaussian Process Regression (GPR) which provides uncertainty bounds on the prediction. We then formulate a stochastic model predictive control (MPC) for the autonomous ride-hailing service and integrate the demand predictions with uncertainty bounds. In order to guarantee constraint satisfaction in the optimization under estimated stochastic demand prediction, we employ a probabilistic constraining method with user-defined confidence interval, using Chance Constrained MPC (CCMPC). The benefits of the proposed method are twofold. First, travel demand uncertainty prediction from data can naturally be embedded into the MoD optimization framework, allowing us to keep the imbalance at each station below a certain threshold with a user-defined probability. Second, CCMPC can be relaxed into a Mixed-Integer-Linear-Program (MILP) and the MILP can be solved as a corresponding Linear-Program, which always admits an integral solution. Our transportation simulations show that by tuning the confidence bound on the chance constraint, close to optimal oracle performance can be achieved, with a median customer wait time reduction of 4% compared to using only the mean prediction of the GPR.

Data-Driven Distributionally Robust Electric Vehicle Balancing for Autonomous Mobility-on-Demand Systems Under Demand and Supply Uncertainties

Electric vehicles (EVs) are being rapidly adopted due to their economic and societal benefits. Autonomous mobility-on-demand (AMoD) systems also embrace this trend. However, the long charging time and high recharging frequency of EVs pose challenges to efficiently managing EV AMoD systems. The complicated dynamic charging and mobility process of EV AMoD systems makes the demand and supply uncertainties significant when designing vehicle balancing algorithms. In this work, we design a data-driven distributionally robust optimization (DRO) approach to balance EVs for both the mobility service and the charging process. The optimization goal is to minimize the worst-case expected cost under both passenger mobility demand uncertainties and EV supply uncertainties. We then propose a novel distributional uncertainty sets construction algorithm that guarantees the produced parameters are contained in desired confidence regions with a given probability. To solve the proposed DRO AMoD EV balancing problem, we derive an equivalent computationally tractable convex optimization problem. Based on real-world EV data of a taxi system, we show that with our solution the average total balancing cost is reduced by 14.49%, and the average mobility fairness and charging fairness are improved by 15.78% and 34.51%, respectively, compared to solutions that do not consider uncertainties.

A Scientometric Review of Mobility-on-Demand Car-Sharing Systems

This article gives a scientometric analysis of research advances in mobility-on-demand (MoD)systems. By reviewing 191 articles with diverse perspectives from the past 10 years, a scientometric study is conducted using CiteSpace, giving an integrated knowledge mind mapping of MoD systems. The results indicate that the major contributions in MoD systems field are mainly from the United States, Singapore, China, and Germany. Meanwhile, pivotal institutions, journals, and article keywords are discussed. MoD system knowledge domains are concentrated in the categories of demand and service, system optimization, impacts and factors, and response and choice. The development of MoD systems is evolving from taxis and minibuses to autonomous vehicles. In particular, research interest in autonomous MoD systems is rapidly growing. This analysis gives a comprehensive and integral knowledge mind mapping, which contributes to the MoD system information network and provides valuable insights into future research directions and developments for investigators and software developers in the field of vehicle-sharing and traffic scheduling.

Acceptability modeling of autonomous mobility on-demand services with on-board ride sharing using interpretable Machine Learning

Research on the factors that may affect the future uptake of autonomous Mobiliyt on Demand (MoD) services are today more than ever relevant. In this paper, we attempt to investigate all aspects of acceptability of a proposed Autonomous-MoD service (AMoD). More specifically, we develop mode choice models and identify the factors that affect it in both sunny and rainy weather conditions, using state-of-the-art Machine Learning models and interpretation techniques, such as the permutation feature importance and partial dependence. Furthermore, we estimate the willingness of the service’s potential users to pay for reduced travel time and propose an on-board negotiation scheme of the travel time and cost for sharing one’s ride. For the above purposes, we conducted a questionnaire survey with 1600 participants in the city of Athens, Greece just after alleviation of the lockdown and measures related with the COVID-19 first wave. The models developed are capable of predicting the mode choice and acceptability of the negotiation scheme with an accuracy of over 80%. Except for the cost, travel and walking time of each alternative mode, the users’ mobility profile, attitude towards autonomous vehicles and demographic characteristics are identified as the most important factors affecting the respondents’ choices. Moreover, the willingness to pay for reduced travel time varies from 0.18 to 0.62€, depending on the mode and about 0.53€ for the on-board negotiation.

Learning to Branch-and-Bound to Route an Autonomous Mobility on Demand System

Autonomous Mobility on Demand (AMoD) is a system consisting of a fleet of centrally-controlled, autonomous vehicles that take customers from their requested origins to their requested destinations. In order to minimize the distance traveled by the fleet, a routing scheme must be developed to service all customer requests. This paper investigates the usage of the branch-and-bound algorithm (BB) to find such a scheme, as well as the usage of a neural network (NN) to speed up BB. Given a fixed road network, sets of randomly generated requests were passed into BB to obtain the ordering of each set that minimized the number of computations, and a NN was trained on this data. New randomly generated request sets were then passed into the trained NN, and runtimes given the NN-predicted orderings were compared to average runtimes over all permutations. For a NN with 1024 nodes in the first and 512 in the second hidden layer and a learning rate of 0.1, using the NN resulted in an average 40% decrease from average runtimes; furthermore, NN-orderings never resulted in increased runtime. Other combinations of NN parameters resulted in around 25% decrease in runtime. Also, BB performed the same number of computations for all permutations with the same first request, simplifying the problem to only finding the next request. These results show that training the NN results in a more efficient, faster routing algorithm that is therefore easier to scale up, enabling at-scale adoption of AMoD.

Quality-of-Service Aware Battery Swapping Navigation and Pricing for Autonomous Mobility-on-Demand System

The autonomous mobility-on-demand (AMoD) system is regarded as a promising shared mobility method for the sustainable transportation system. In addition, battery swapping could become an efficient energy-refueling means for electric vehicles (EVs). The quality-of-service aware battery swapping price determined by the battery swapping stations (BSSs) considers the interaction between AMoD fleet and BSSs, which is formulated as a bilevel optimization problem in this article. The interdependence between two systems is bridged by the swapping price that is both time-varying and location-varying. In the upper level, the swapping pricing is optimized to reflect the battery inventory so that the EV fleet could be navigated for the BSS strategically, the queue time would be reduced consequently. BSS operator optimizes swapping pricing scheme and battery charging management with the objective of minimizing the cost induced by the queue process and battery charging. In the lower level, a unified network flow model incorporating queue procedure in BSSs is proposed in order to characterize the operational decision of AMoD fleet. The AMoD operator develops fleet scheduling strategies with the objective of maximizing the AMoD system profit. An iteration-based algorithm is applied to attain the results and the effectiveness of the algorithm is validated by the real-world data from New York.

Joint Charging Infrastructure and Autonomous Fleet Planning Considering Distributed Renewable Resources

Considering its unique operational characteristics, the autonomous mobility-on-demand (AMoD) system is playing a more and more critical role in the future transportation system. In this article, we consider a joint planning problem of autonomous fleet (i.e., vehicle type and fleet size) and charging infrastructure (i.e., the size of charging piles with various charging rates, distributed renewable resources) while accounting for the spatial–temporal demand-side flexibility of the AMoD system. We adopt an augmented network flow model to capture the autonomous fleet's charging, order serving, and rebalancing decisions. The proposed planning problem is modeled as a two-stage stochastic program and solved by multicut Benders decomposition techniques. Simulation experiments based on real-world data are performed to demonstrate the effectiveness of the proposed planning scheme.

Understanding user perception and feelings for autonomous mobility on demand in the COVID-19 pandemic era

New mobility-on-demand services together with the emerging technology of autonomous vehicles (AV) aim to revolutionize urban transportation systems, by introducing autonomous driving and sophisticated sharing and routing schemes for efficiently serving individual’s needs and requirements. On the other hand, the COVID-19 pandemic has disrupted travel patterns due to the emerging trends of social distancing and teleworking. In this paper, we aim at investigating users’ perception on autonomous vehicles, mobility on demand schemes as well as on the future transportation landscape using data collected through a questionnaire survey in the Metropolitan Area of Athens, Greece conducted after the first COVID-19 pandemic wave. First, a statistical analysis of the responses is performed and, then, a clustering approach is followed to identify user profiles based on daily mobility patterns and attitudes towards autonomous vehicles. Subsequently, the identified profiles are exploited in the development of a Bayesian Network to reveal interrelations between user profiling, attitudes and perceptions for future mobility services. Regarding the acceptance of Autonomous Mobility on Demand (AMoD) services, as well as travelers’ level of happiness concerning future scenarios of urban transportation, results have shown that the majority of travelers in Athens will be more than happy in the case where the entire transportation system is served with AMoD services.

Reinforcement Learning for Autonomous Mobility-on-Demand Systems

Autonomous mobility-on-demand (AMoD) systems represent a rapidly developing mode of transportation wherein travel requests are dynamically handled by a coordinated fleet of robotic, self-driving vehicles. Given a graph representation of the transportation network - one where, for example, nodes represent areas of the city, and edges the connectivity between them - we argue that the AMoD control problem is naturally cast as a node-wise decision-making problem. In this paper, we propose a deep reinforcement learning framework1 to control the rebalancing of AMoD systems through graph neural networks. Crucially, we demonstrate that graph neural networks enable reinforcement learning agents to recover behavior policies that are significantly more transferable, generalizable, and scalable than policies learned through other approaches. Empirically, we show how the learned policies exhibit promising zero-shot transfer capabilities when faced with critical portability tasks such as inter-city generalization, service area expansion, and adaptation to potentially complex urban topologies.

An agent-based model for assessing the financial viability of autonomous mobility on-demand systems used as first and last-mile of public transport trips: A case-study in Rotterdam, the Netherlands

The continuing urbanization and corresponding increase in transport demand are putting pressure on the accessibility, safety, sustainability, livability, and efficiency of urbanized regions. Public transport is regarded as a sustainable mode of transport for these regions and therefore transport policies aim to increase its attractiveness. However, public transport is facing last-mile connectivity problems. The application of Autonomous Mobility on-Demand (AMoD) as a feeder service for public transport hubs can potentially improve the first and last-mile trip leg which increases the attractivity of public transport. However, will such a system be financially viable when applied in an urban area? and what kind of operation will lead to the highest system performance? In this research, this question is addressed by proposing a method that connects macro transport modeling and agent-based modeling (ABM). An existing gravity-based travel demand estimation model built in a macro simulation tool is used to predict passenger demand across all the OD pairs of a city. For those OD pairs that can use the AMoD as first /last mile this is modeled using an agent-based rationale to be able to simulate the behavior of passengers and vehicles within that specific area of the city. The simulation model is applied to the case study area of the south of Rotterdam, in The Netherlands, where metro Station Zuidplein and the rail Station Lombardijen function as two AMoD hubs. Using the case study, the impact of relocation, ridesharing, and charging strategy is assessed in regards to financial viability. Among other insights, results show that the AMoD service leads to a profit on a typical business day for the operating companies despite the high-quality level of the service (very low average waiting time for a vehicle). If this particular system would not consist of automated vehicles and one would have to pay a salary to drivers, it would not be possible to make a profit on a typical business day. Moreover, results show that activating dynamic ridesharing and using wireless fast chargers at the stations results in the most financially viable operation. Activating automatic relocations results in the most costly operation.

Deep Reinforcement Learning-Based Charging Pricing for Autonomous Mobility-on-Demand System

The autonomous mobility-on-demand (AMoD) system plays an important role in the urban transportation system. The charging behavior of AMoD fleet becomes a critical link between charging system and transportation system. In this paper, we investigate a strategic charging pricing scheme for charging station operators (CSOs) based on a non-cooperative Stackelberg game framework. The Stackelberg equilibrium investigates the pricing competition among multiple CSOs, and explores the nexus between the CSOs and AMoD operator. In the proposed framework, the responsive behavior of AMoD operator (order-serving, repositioning, and charging) is formulated as a multi-commodity network flow model to solve an energy-aware traffic flow problem. Meanwhile, a soft actor-critic based multi-agent deep reinforcement learning algorithm is developed to solve the proposed equilibrium framework while considering privacy-conservation constraints among CSOs. A numerical case study with city-scale real-world data is used to validate the effectiveness of the proposed framework.

A business class for autonomous mobility-on-demand: Modeling service quality contracts in dynamic ridesharing systems

With the popularization of transportation network companies (TNCs) (e.g., Uber, Lyft) and the rise of autonomous vehicles (AVs), even major car manufacturers are increasingly considering themselves as autonomous mobility-on-demand (AMoD) providers rather than individual vehicle sellers. However, matching the convenience of owning a vehicle requires providing consistent service quality, taking into account individual expectations. Typically, different classes of users have different service quality (SQ) expectations in terms of responsiveness, reliability, and privacy. Nonetheless, AMoD systems presented in the literature do not enable active control of service quality in the short term, especially in light of unusual demand patterns, sometimes allowing extensive delays and user rejections. This study proposes a method to control the daily operations of an AMoD system that uses the SQ expectations of heterogeneous user classes to dynamically distribute service quality among riders. Additionally, we consider an elastic vehicle supply, that is, privately-owned freelance AVs (FAVs) can be hired on short notice to help providers meeting user service-level expectations. We formalize the problem as the dial-a-ride problem with service quality contracts (DARP-SQC) and propose a multi-objective matheuristic to address real-world requests from Manhattan, New York City. Applying the proposed service-level constraints, we improve user satisfaction (in terms of reached service-level expectations) by 53% on average compared to conventional ridesharing systems, even without hiring additional vehicles. By deploying service-quality-oriented on-demand hiring, our hierarchical optimization approach allows providers to adequately cater to each segment of the customer base without necessarily owning large fleets.

Rebalancing Autonomous Vehicles using Deep Reinforcement Learning

The shared autonomous mobility-on-demand (AMoD) system is a promising business model in the coming future which provides a more efficient and affordable urban travel mode. However, to maintain the efficient operation of AMoD and address the demand and supply mismatching, a good rebalancing strategy is required. This paper proposes a reinforcement learning-based rebalancing strategy to minimize passengers’ waiting in a shared AMoD system. The state is defined as the nearby supply and demand information of a vehicle. The action is defined as moving to a nearby area with eight different directions or staying idle. A 4.6 4.4 km2 region in Cambridge, Massachusetts, is used as the case study. We trained and tested the rebalancing strategy in two different demand patterns: random and first-mile. Results show the proposed method can reduce passenger’s waiting time by 7% for random demand patterns and 10% for first-mile demand patterns.

A Simulation Assessment of Autonomous Mobility on Demand in the City of Edinburgh

We present simulation scenarios of private and public transportation modes of transportation being replaced by autonomous mobility-on-demand (AMoD) system in Edinburgh, UK, based on realistic temporal and spatial demand patterns. Edinburgh is interesting to simulate because it is selected to be one of the top 50 congested cities despite the traditional layout of the city with a low population density. For each mode of transportation, we test five different operational policies (from the literature) for different fleet sizes. We examine the results both in terms of system expenditures and profits, as well as service quality. Additionally, we evaluate Pareto solutions that combine policies’ performance to give a trade-off between wait times and empty miles. In comparison with private cars, the Autonomous Vehicles (AV) resulted in 86% reduction in the number of cars on the road. In addition, the use of the AVs per day is at least two to seven times higher than the UK private car usage. A reduction in the number of vehicles will have a positive influence on the congestion. When compared with the public mode, AMoD vehicles require 62% less time and 60% less distance than buses for the same trips.

Modeling and Optimization of Multiaction Dynamic Dispatching Problem for Shared Autonomous Electric Vehicles

The fusion of electricity, automation, and sharing is forming a new Autonomous Mobility-on-Demand (AMoD) system in current urban transportation, in which the Shared Autonomous Electric Vehicles (SAEVs) are a fleet to execute delivery, parking, recharging, and repositioning tasks automatically. To model the decision-making process of AMoD system and optimize multiaction dynamic dispatching of SAEVs over a long horizon, the dispatching problem of SAEVs is modeled according to Markov Decision Process (MDP) at first. Then two optimization models from short-sighted view and farsighted view based on combinatorial optimization theory are built, respectively. The former focuses on the instant and single-step reward, while the latter aims at the accumulative and multistep return. After that, the Kuhn–Munkres algorithm is set as the baseline method to solve the first model to achieve optimal multiaction allocation instructions for SAEVs, and the combination of deep Q-learning algorithm and Kuhn–Munkres algorithm is designed to solve the second model to realize the global optimization. Finally, a toy example, a macrosimulation of 1 month, and a microsimulation of 6 hours based on actual historical operation data are conducted. Results show that (1) the Kuhn–Munkres algorithm ensures the computational effectiveness in the large-scale real-time application of the AMoD system; (2) the second optimization model considering long-term return can decrease average user waiting time and achieve a 2.78% increase in total revenue compared with the first model; (3) and integrating combinatorial optimization theory with reinforcement learning theory is a perfect package for solving the multiaction dynamic dispatching problem of SAEVs.

Integrated Operation Model for Autonomous Mobility-on-Demand Fleet and Battery Swapping Station

The autonomous mobility-on-demand (AMoD) system provides an alternative solution for sustainable and economical transportation system. Meanwhile, battery swapping could become a promising approach to sustain the efficient operation of EV fleet. This article proposes a combined operation scheme for battery swapping station (BSS) and AMoD system. To maximizing the profit of AMoD system, an expanded network flow model is employed to determine swapping scheduling and vehicle rebalancing for EV fleet in AMoD system. The interaction between the shared mobility-on-demand system and BSS is modeled by the swapping demand and swapping price. Specifically, swapping demand is determined in the AMoD system and introduced as known parameters in the BSS problem. Fleet management operation in AMoD system will be influenced by swapping price provided by BSS. The BSS management problem is formulated as a mixed-integer linear programming model, which optimizes refueling decisions for depleted batteries. Simulation experiments based on real-world data from New York City are provided. The results demonstrate the effectiveness of this proposed integrated operation model.