Showing papers on "Fleet management published in 2018"

Planning and Decision-Making for Autonomous Vehicles

Abstract: In this review, we provide an overview of emerging trends and challenges in the field of intelligent and autonomous, or self-driving, vehicles. Recent advances in the field of perception, planning,...

Efficient Large-Scale Fleet Management via Multi-Agent Deep Reinforcement Learning

Abstract: Large-scale online ride-sharing platforms have substantially transformed our lives by reallocating transportation resources to alleviate traffic congestion and promote transportation efficiency. An efficient fleet management strategy not only can significantly improve the utilization of transportation resources but also increase the revenue and customer satisfaction. It is a challenging task to design an effective fleet management strategy that can adapt to an environment involving complex dynamics between demand and supply. Existing studies usually work on a simplified problem setting that can hardly capture the complicated stochastic demand-supply variations in high-dimensional space. In this paper we propose to tackle the large-scale fleet management problem using reinforcement learning, and propose a contextual multi-agent reinforcement learning framework including two concrete algorithms, namely contextual deep Q-learning and contextual multi-agent actor-critic, to achieve explicit coordination among a large number of agents adaptive to different contexts. We show significant improvements of the proposed framework over state-of-the-art approaches through extensive empirical studies.

A modeling framework for the dynamic management of free-floating bike-sharing systems

Abstract: Given the growing importance of bike-sharing systems nowadays, in this paper we suggest an alternative approach to mitigate the most crucial problem related to them: the imbalance of bicycles between zones owing to one-way trips. In particular, we focus on the emerging free-floating systems, where bikes can be delivered or picked-up almost everywhere in the network and not just at dedicated docking stations. We propose a new comprehensive dynamic bike redistribution methodology that starts from the prediction of the number and position of bikes over a system operating area and ends with a relocation Decision Support System. The relocation process is activated at constant gap times in order to carry out dynamic bike redistribution, mainly aimed at achieving a high degree of user satisfaction and keeping the vehicle repositioning costs as low as possible. An application to a test case study, together with a detailed sensitivity analysis, shows the effectiveness of the suggested novel methodology for the real-time management of the free-floating bike-sharing systems.

Dynamic autonomous vehicle fleet operations: Optimization-based strategies to assign AVs to immediate traveler demand requests

Abstract: Motivated by the growth of ridesourcing services and the expected advent of fully-autonomous vehicles (AVs), this paper defines, models, and compares assignment strategies for a shared-use AV mobility service (SAMS). Specifically, the paper presents the on-demand SAMS with no shared rides, defined as a fleet of AVs, controlled by a central operator, that provides direct origin-to-destination service to travelers who request rides via a mobile application and expect to be picked up within a few minutes. The underlying operational problem associated with the on-demand SAMS with no shared rides is a sequential (i.e. dynamic or time-dependent) stochastic control problem. The AV fleet operator must assign AVs to open traveler requests in real-time as traveler requests enter the system dynamically and stochastically. As there is likely no optimal policy for this sequential stochastic control problem, this paper presents and compares six AV-traveler assignment strategies (i.e. control policies). An agent-based simulation tool is employed to model the dynamic system of AVs, travelers, and the intelligent SAMS fleet operator, as well as, to compare assignment strategies across various scenarios. The results show that optimization-based AV-traveler assignment strategies, strategies that allow en-route pickup AVs to be diverted to new traveler requests, and strategies that incorporate en-route drop-off AVs in the assignment problem, reduce fleet miles and decrease traveler wait times. The more-sophisticated AV-traveler assignment strategies significantly improve operational efficiency when fleet utilization is high (e.g. during the morning or evening peak); conversely, when fleet utilization is low, simply assigning traveler requests sequentially to the nearest idle AV is comparable to more-advanced strategies. Simulation results also indicate that the spatial distribution of traveler requests significantly impacts the empty fleet miles generated by the on-demand SAMS.

Efficient Large-Scale Fleet Management via Multi-Agent Deep Reinforcement Learning.

Abstract: Large-scale online ride-sharing platforms have substantially transformed our lives by reallocating transportation resources to alleviate traffic congestion and promote transportation efficiency. An efficient fleet management strategy not only can significantly improve the utilization of transportation resources but also increase the revenue and customer satisfaction. It is a challenging task to design an effective fleet management strategy that can adapt to an environment involving complex dynamics between demand and supply. Existing studies usually work on a simplified problem setting that can hardly capture the complicated stochastic demand-supply variations in high-dimensional space. In this paper we propose to tackle the large-scale fleet management problem using reinforcement learning, and propose a contextual multi-agent reinforcement learning framework including two concrete algorithms, namely contextual deep Q-learning and contextual multi-agent actor-critic, to achieve explicit coordination among a large number of agents adaptive to different contexts. We show significant improvements of the proposed framework over state-of-the-art approaches through extensive empirical studies.

MOVI: A Model-Free Approach to Dynamic Fleet Management

Abstract: Modern vehicle fleets, e.g., for ridesharing platforms and taxi companies, can reduce passengers' waiting times by proactively dispatching vehicles to locations where pickup requests are anticipated in the future. Yet it is unclear how to best do this: optimal dispatching requires optimizing over several sources of uncertainty, including vehicles' travel times to their dispatched locations, as well as coordinating between vehicles so that they do not attempt to pick up the same passenger. While prior works have developed models for this uncertainty and used them to optimize dispatch policies, in this work we introduce a model-free approach. Specifically, we propose MOVI, a Deep Q-network (DQN)-based framework that directly learns the optimal vehicle dispatch policy. Since DQNs scale poorly with a large number of possible dispatches, we streamline our DQN training and suppose that each individual vehicle independently learns its own optimal policy, ensuring scalability at the cost of less coordination between vehicles. We then formulate a centralized receding-horizon control (RHC) policy to compare with our DQN policies. To compare these policies, we design and build MOVI as a large-scale realistic simulator based on 15 million taxi trip records that simulates policy-agnostic responses to dispatch decisions. We show that the DQN dispatch policy reduces the number of unserviced requests by 76% compared to without dispatch and 20% compared to the RHC approach, emphasizing the benefits of a model-free approach and suggesting that there is limited value to coordinating vehicle actions. This finding may help to explain the success of ridesharing platforms, for which drivers make individual decisions.

Dynamic Autonomous Vehicle Fleet Operations: Optimization-Based Strategies to Assign AVs to Immediate Traveler Demand Requests

Abstract: Motivated by the growth of ridesourcing services and the expected advent of fully-autonomous vehicles (AVs), this paper defines, models, and compares assignment strategies for a shared-use AV mobility service (SAMS). Specifically, the paper presents the on-demand SAMS with no shared rides, defined as a fleet of AVs, controlled by a central operator, that provides direct origin-to-destination service to travelers who request rides via a mobile application and expect to be picked up within a few minutes. The underlying operational problem associated with the on-demand SAMS with no shared rides is a sequential (i.e. dynamic or time-dependent) stochastic control problem. The AV fleet operator must assign AVs to open traveler requests in real-time as traveler requests enter the system dynamically and stochastically. As there is likely no optimal policy for this sequential stochastic control problem, this paper presents and compares six AV-traveler assignment strategies (i.e. control policies). An agent-based simulation tool is employed to model the dynamic system of AVs, travelers, and the intelligent SAMS fleet operator, as well as, to compare assignment strategies across various scenarios. The results show that optimization-based AV-traveler assignment strategies, strategies that allow en-route pickup AVs to be diverted to new traveler requests, and strategies that incorporate en-route drop-off AVs in the assignment problem, reduce fleet miles and decrease traveler wait times. The more-sophisticated AV-traveler assignment strategies significantly improve operational efficiency when fleet utilization is high (e.g. during the morning or evening peak); conversely, when fleet utilization is low, simply assigning traveler requests sequentially to the nearest idle AV is comparable to more-advanced strategies. Simulation results also indicate that the spatial distribution of traveler requests significantly impacts the empty fleet miles generated by the on-demand SAMS.

A decision support system to investigate dynamic last-mile distribution facilitating cargo-bikes

Abstract: This work presents a decision support system to facilitate efficient urban last-mile distribution. Orders are collected and delivered by a fleet of both conventional vehicles owned by a logistics provider and cargo-bikes operated by freelancers. Additionally, micro-hubs are operated to perform transshipments between multiple vehicles. To investigate the corresponding problem setting, an agent-based simulation is developed, which uses dynamic optimisation procedures to generate and select vehicle routes and transshipment points. Experiments motivated by dynamic real-world urban restaurant delivery services investigate the impact of cargo-bikes, urban consolidation and guaranteed delivery times. Potentials are discussed and implications for successful implementations are provided. Results highlight the importance of having a sufficient number of active cargo-bikes available and benefits of incorporating consolidation strategies to guarantee timely deliveries.

Optimizing the Scheduling of Autonomous Guided Vehicle in a Manufacturing Process

Abstract: Autonomous Guided Vehicles (AGVs) are considered as one of the key enablers of smart factories which make possible smart and flexible transportation of pallets and material on shopfloor. However, existing AGV fleet management solutions often suffer from poor integration with real-time manufacturing operations information systems, which negatively affects scheduling of AGVs. To exploit the full potential of AGVs in achieving just-intime (JIT) transportation, there is a need for intelligent AGV fleet management system which not only integrate with manufacturing information technology (IT) and operational technology (OT) but also provide prediction for the shop-floor logistic based on real-time manufacturing operations information to optimize scheduling of AGVs. This paper presents an approach for a Smart AGV Management System (SAMS), which combines the real-time data analysis and digital twin models that can be deployed within complex manufacturing environments for optimized scheduling. For a proof of concept, a case study of a line side supply of components to a manual assembly station is presented.

Fleet control algorithms for automated mobility: A simulation assessment for Zurich

Abstract: 1 The performance of four different dispatching and rebalancing algorithms for the control of 2 an automated mobility-on-demand system is evaluated in a simulation environment. The case 3 study conducted with an agent-based simulation scenario of the city of Zurich shows that the 4 choice of an intelligent rebalancing algorithm decreases the average wait time in the system. 5 For a wait time of four minutes at peak hours the most performant algorithm requires the same 6 price per vehicle kilometer as a private car today. The results show that such an automated 7 mobility on demand service can be offered while maintaining a higher fleet occupancy than with 8 conventional private cars. 9 Hörl, S., Ruch, C., Becker, F., Frazzoli, E. and Axhausen, K. W. 2

Data-Driven Resilient Fleet Management for Cloud Asset-enabled Urban Flood Control

Abstract: Emergency fleet management has become one of the determinant success factors for post-disaster responses in urban flood control. However, it is challenging as multiple types of emergency vehicles are involved, and its performance is frequently threatened by the fluctuation of rescue demands and fleet capacity. Aiming at coping with the imbalances between rescue demands and vehicle supplies, and maintaining required service level of fleet management after flood occurs, this paper proposes a data-driven resilient fleet management solution under the context of cloud asset-enabled urban flood control. First, the problem of resilient fleet management is quantitatively defined, and then a data-driven dynamic management mechanism is proposed, which is highly effective on realizing resilient fleet management. Furthermore, considering the cooperation among different types of emergency vehicles, a greedy-based algorithm is proposed for resilient vehicle dispatching based on real-time scenarios. Finally, a simulation case is also conducted to verify the effectiveness and performance of the proposed solution.

MOVI: A Model-Free Approach to Dynamic Fleet Management

Abstract: Modern vehicle fleets, e.g., for ridesharing platforms and taxi companies, can reduce passengers' waiting times by proactively dispatching vehicles to locations where pickup requests are anticipated in the future. Yet it is unclear how to best do this: optimal dispatching requires optimizing over several sources of uncertainty, including vehicles' travel times to their dispatched locations, as well as coordinating between vehicles so that they do not attempt to pick up the same passenger. While prior works have developed models for this uncertainty and used them to optimize dispatch policies, in this work we introduce a model-free approach. Specifically, we propose MOVI, a Deep Q-network (DQN)-based framework that directly learns the optimal vehicle dispatch policy. Since DQNs scale poorly with a large number of possible dispatches, we streamline our DQN training and suppose that each individual vehicle independently learns its own optimal policy, ensuring scalability at the cost of less coordination between vehicles. We then formulate a centralized receding-horizon control (RHC) policy to compare with our DQN policies. To compare these policies, we design and build MOVI as a large-scale realistic simulator based on 15 million taxi trip records that simulates policy-agnostic responses to dispatch decisions. We show that the DQN dispatch policy reduces the number of unserviced requests by 76% compared to without dispatch and 20% compared to the RHC approach, emphasizing the benefits of a model-free approach and suggesting that there is limited value to coordinating vehicle actions. This finding may help to explain the success of ridesharing platforms, for which drivers make individual decisions.

Battery Health Monitoring and Degradation Prognosis in Fleet Management Systems

Abstract: Today, fleet management systems with battery health monitoring capabilities are in the focus more than ever. This paper addresses the development of a novel battery health monitoring algorithm with a degradation prognosis feasibility particularly adapted for usage in fleet management systems. Moreover, the chosen degradation prognosis approach adapts itself continuously on varying environmental conditions or utilization modes by identifying the impact factors which lead to a certain degradation trend. Such findings, when accessible with a fleet management system, offer various possibilities for fleet analysis techniques e.g., to identify an imminent battery failure.

Analysis of smart city transportation using IoT

Abstract: As a noteworthy part of present day economy, transportation represents 6–12% of the (GDP) in may create nations. In spite of the fact that transportation has extraordinarily enhanced our lives, many exorbitant issues stay unsolved, including auto collision, blockage and vehicle discharge. Brilliant transportation has as of late turned into an interesting issue in the Internet of Things (IoT) region and is considered as the answer for the issues said above. In this report, we inspected smart transportation from an inventive point of view and separation the theme into a couple of sub subjects. In view of the market measure and the requests for versatile system, we have recognized four noteworthy open doors for mobile network operators (MNOs), incorporating telematics benefit with Usage Based Insurance (UBI). What's more, fleet management, smart parking administration based on the Narrow-band IoT (NB-IoT) arrange Innovation, emergency service Network (ESN) in light of long Term Evolution (LTE) organize and upgraded Advanced Driver Assistance Systems (ADAS) in view of LTE-V or 5G arrange.

RASberry - Robotic and Autonomous Systems for Berry Production

Abstract: The soft fruit industry is facing unprecedented challenges due to its reliance of manual labour. We are presenting a newly launched robotics initiative which will help to address the issues faced by the industry and enable automation of the main processes involved in soft fruit production. The RASberry project (Robotics and Autonomous Systems for Berry Production) aims to develop autonomous fleets of robots for horticultural industry. To achieve this goal, the project will bridge several current technological gaps including the development of a mobile platform suitable for the strawberry fields, software components for fleet management, in-field navigation and mapping, long-term operation, and safe human-robot collaboration. In this paper, we provide a general overview of the project, describe the main system components, highlight interesting challenges from a control point of view and then present three specific applications of the robotic fleets in soft fruit production. The applications demonstrate how robotic fleets can benefit the soft fruit industry by significantly decreasing production costs, addressing labour shortages and being the first step towards fully autonomous robotic systems for agriculture.

The technician routing and scheduling problem with conventional and electric vehicle

Abstract: Motivated by stricter environmental regulations, government incentives, branding opportunities, and potential cost reductions, companies are replacing their conventional vehicles (CVs) with electric vehicles (EVs). However, due to financial and operational restrictions, these fleet transitions usually occur in several stages. This introduces new operational challenges, because most existing fleet management and routing tools are not designed to handle hybrid fleets of CVs and EVs. In this paper, we study a problem arising in the daily operations of telecommunication companies, public utilities, home healthcare providers, and other businesses: the technician routing and scheduling problem with conventional and electric vehicles (TRSP-CEV). To solve this problem we propose a two-phase parallel matheuristic. In the first phase the matheuristic decomposes the problem into several vehicle routing problems with time windows, and it solves these problems (in parallel) using a greedy randomized adaptive search procedure (GRASP). At the end of each GRASP iteration, the routes making up the local optimum are stored in long-term memory. In the second phase, the method uses the stored routes to find a TRSP-CEV solution. We discuss computational experiments on industrial instances provided by a French utility. We provide managerial insight into the impact of the proportion of EVs in the fleet on metrics such as the number of routes, the total operational cost, and the CO2 emissions. Additionally, we present state-of-the-art results for the closely related electric fleet size and mix vehicle routing problem with time windows and recharging stations.

Three-state markov model of using transport means

Abstract: Guarantee of the high level of tasks execution in enterprise is a proper organization of processes and provision of necessary resources for their implementation. Particularly important, especially from the distribution company point of view, is reliability of held transport means. It depends on rational fleet management, adherence to service intervals, proper their use, as well as even workload and avoidance of unnecessary mileage, which contributes to accelerated wear. The analysis presented in this article showed that exploitation of transport means is also affected by factors not directly related to them, such as personnel decisions, which strongly determined the degree and manner of their use in the investigated company. Incompetent employees of the customer service department caused that there were generated unavoidable mileage, which could be avoided. They contributed not only to the increase of process costs, but also increased the degree of transport means consumption, unnecessarily reducing their efficiency and effectiveness. In the study, Markov models were proposed in both discreet and continuous physical time. It concerned two stages of the process, before and after implementation of changes. An analysis of the process evolution over time was also made, calculating boundary probabilities. The obtained models not only allowed for a description of the analyzed system and prediction of selected logistic indicators, but also indicated the directions for possible improvements.

Transformers Fleet Management Through the use of an Advanced Health Index

Abstract: Power transformers represent the highest value of the equipment installed in transmission substations, comprising up the 60% of the total investment. They are expected to operate for several decades without faults and possibly without relevant unscheduled maintenance practice. The new approach is developed for reducing time based maintenance and, increasing condition based maintenance and to introduce predictive maintenance as well. The aim of predictive maintenance is first to predict when transformer failure might occur, and secondly, to prevent occurrence of the failure by performing maintenance. Diagnostic information can be evaluated individually or better by a complex algorithm which merges all the single inputs and their Rate of Increase (RoI) creating a mono-dimensional figure called Health Index (HI). This concept represents a real ‘shifting of paradigm’, as it deeply affects the criteria for transformers grid management and selection of the electrical utilities and grid companies.

GPS-data-driven dynamic destination prediction for on-demand one-way carsharing system

Abstract: On-demand one-way carsharing systems are increasingly gaining popularity nowadays and the market is growing unprecedentedly. However, operating techniques such as vehicle surveillance and fleet management fall behind the industrial development. Practice on EVCARD - an on-demand one-way electric vehicle carsharing system operating in Shanghai - raises an issue on dynamically predicting the user destinations, in order to support decisions on dynamic fleet management. This study presents a global positioning system (GPS)-data-driven method to solve the problem. The historical vehicle GPS data is enabled to match the user current trajectories and infer their possible destinations. Based on the GPS trajectory similarity measurement, the study presents a four-step procedure, including (i) similarity calculation, (ii) most-similar track detection, (iii) adjustment, and (iv) sorting. The method also takes the station correlations and the user historical destinations into account. A case study is given to demonstrate the dynamic prediction procedure of this method. Experiment on 96,821 valid test tracks shows that the positive prediction rate can be above 92% if the test trip has been completed over 70%. Factors that may influence the prediction result are additionally discussed, which include the existence of round-trips, the coverage of samples, and the alternatives of destinations.

System and method for tire sensor-based autonomous vehicle fleet management

Abstract: A system for sensing tire parameters in the remote monitoring and management of a fleet of autonomous vehicles is provided. The system includes at least one autonomous vehicle that is supported by at least one tire. At least one sensor is affixed to the tire for sensing tire parameters. Means are provided for communicating data generated by the sensor to a control system on the vehicle, and a mobile network receives the sensor data from the vehicle control system. A fleet management server receives the sensor data from the mobile network, and means are provided to generate commands for the autonomous vehicle in real time based upon the data generated by the sensor. A method for sensing tire parameters in the remote monitoring and management of a fleet of autonomous vehicles is also provided.

Identifying and Classifying Highway Bottlenecks Based on Spatial and Temporal Variation of Speed

Abstract: In this paper, we develop a novel approach to identify bottlenecks on highways using probe data collected by commercial global positioning system (GPS) fleet management devices installed in...

Passenger classification-based autonomous vehicle routing

Abstract: An autonomous vehicle (AV) fleet management system includes a communications interface, one or more processors, and a memory that stores instructions that cause the processors to receive, a request for a ride using an AV from a wireless communication device of a user. The request may include a pick-up location of the user. The instructions also cause the processors to identify demographic information related to the user, determine a vulnerability score and a priority of the user based at least in part on the demographic information, and receive location information from a plurality of AVs in a fleet of AVs. The instructions further cause the processors to identify a particular AV from the fleet of AVs based at least in part on the vulnerability score, the priority, the pick-up location of the user, and the location of the particular AV and cause the particular AV to pick-up the user.

GNSS Performance Modelling for Positioning and Navigation in Urban Environments

Abstract: The Global Navigation Satellite System (GNSS) supports a growing number of several Intelligent Transport System (ITS) applications and location based services including collision avoidance, electronic toll collection, fleet management and Unmanned Aircraft System (UAS) navigation This paper performs a detailed performance analysis of GNSS with a focus on failure modes in urban environments as a critical case study A guidance-based augmentation strategy based on trajectory optimization is developedwhich accounts for the influence of urban structures on GNSS performance A simulation case study representative of UAS operations in urban environments was performed as a preliminary assessment to corroborate the developed modules

Safe Human-Robot-Interaction in Highly Flexible Warehouses using Augmented Reality and Heterogenous Fleet Management System

Abstract: Nowadays, production systems and warehouses still lack human-robot collaboration. Often due to safety issues robots are kept behind safety fences and the overall system is shutdown once a human enters. This paper is based on a new concept to overcome these issues. It consists of a safety concept which ensures a more efficient human-robot collaboration. The human worker can be localized via a safety vest which he/she wears and the nearby robots are also aware of this information. We propose a more efficient and safe overall system by using an Augmented Reality device which offers a simple interface between humans and robots and a Heterogeneous Fleet Management System which plans and provides collision-free paths for humans and robots and ensures that the resulting environment is safe for both robots and humans.

A duopoly of transportation network companies and traditional radio-taxi dispatch service agencies

Abstract: Transportation network companies commonly enter the market for taxi ride intermediation and alter the market outcome. Compared to cooperatively organized radio-taxi dispatch service agencies, transportation network companies run larger fleets and serve more customers with lower fares, when the fixed costs of the dispatch office are relatively small. The same holds for private dispatch firms, when the fixed costs of a taxicab are not too small. These results are shown in a two-stage duopoly of fare and fleet size competition with fare- and waiting-time-dependent demand.