ADArtem S. Dubrovskii

Artem S. Dubrovskii

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Artem S. Dubrovskii has a total of 6 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Speed profile optimisation for a semi-trailer truck based on a hybrid fuel consumption model with automatic calibration based on CAN bus data

    by Artem S. Dubrovskii, Sergey S. Shadrin, Daria A. Makarova, Yury M. Furletov

    The paper proposes a method for optimising the speed profile of a long-haul semi-trailer truck that combines a physical fuel consumption model with a machine-learning correction with automatic calibration using on-board CAN bus data. The physical component is built on the tractive balance equation, which parameters, i.e. the power transmission efficiency, rolling resistance coefficient, brake-specific fuel consumption and lumped mechanical losses, are identified without bench testing by minimising the difference between the computed and the measured instantaneous fuel rate using the Powell's method; calibration is verified against the cumulative fuel counter (SAE J1939). The correction is implemented with the LightGBM algorithm that approximates the rest of the physical model; for the first time the current gear signal extracted directly from the CAN bus is used as an additional feature. The optimal speed profile is obtained by dynamic programming subject to a constraint on the travel-time increase; the speed range is lumped into 13 values within 60–80 km/h. The method was verified using a field trial with a Mercedes-Benz Actros 1845 truck on the M-11 «Neva» motorway (Moscow – Saint Petersburg, 672.5 km; about 14 million CAN records processed in under two minutes). The calibration error was 3.6%; introduction of the gear feature reduced the prediction root-mean-square error from 13.7 to 11.5 L/100 km (by 16%) and the cumulative-consumption deviation down to 1.4%. The computed fuel saving reached 9.3% at the travel time increased by less than 0.6%. Numerical simulation using the Moscow – Kazan route (809.1 km) confirmed the scalability: a 10.4% saving at a 2.8% time increase without model re-calibration.

  • Архитектура сбора и передачи телематических данных для интеллектуальных транспортных систем магистрального автотранспорта

    by Артем Семенович Дубровский, Сергей Сергеевич Шадрин, Дарья Алексеевна Макарова, Андрей Владимирович Келлер, Юрий Михайлович Фурлетов

    Развитие интеллектуальных транспортных систем для магистрального грузового автотранспорта требует надёжной инфраструктуры сбора, передачи и хранения телематических данных, а также точных цифровых моделей автомобильных дорог, без которых расчёт энергетических затрат на движение и построение адаптивных алгоритмов управления теряют практический смысл. В большинстве публикаций, посвящённых энергоэффективному управлению грузовым транспортом, телематическая подсистема описывается фрагментарно – как источник сигналов CAN-шины и GPS-трека, – без раскрытия архитектуры передачи, протоколов и методологии построения цифровой модели дороги. В настоящей работе разработана и описана трёхуровневая архитектура сбора, передачи и хранения телематических данных, включающая бортовой уровень интеграции с сетями CAN по стандарту SAE J1939, уровень оперативной обработки с навигационной привязкой по каналу Bluetooth Low Energy, а также облачную платформу с типизированным бинарным протоколом поверх TCP/IP и колоночно-ориентированной моделью хранения. Для построения цифровой модели автомобильной дороги использована методика высокоточного картирования, дополненная коррекцией по глобальной модели рельефа методом сплайн-интерполяции, что обеспечивает относительную точность модели не хуже 0,4 м при дискретности опорных точек маршрута в 2 м. Систематизированы факторы внешней среды и предложены принципы их интеграции в архитектуру данных и концепцию интеллектуальной транспортной системы, включая использование каналов V2X для перехода от статической цифровой модели маршрута к адаптивному синтезу решений в реальном времени. Практическая апробация на натурном эксперименте с магистральным автопоездом КАМАЗ-54901-92 на маршруте Р-22 «Каспий» (408 км) подтвердила работоспособность предложенных решений, позволяющих рассматривать их как самостоятельный телематический модуль для интеграции в платформы логистических операторов и системы поддержки принятия решений водителя. The development of intelligent transportation systems for long-haul freight transport requires a reliable infrastructure for telematics data acquisition, transmission and storage, as well as accurate digital road models, without which the calculation of driving energy costs and the design of adaptive control algorithms lose practical value. Most publications on energy-efficient truck management describe the telematics subsystem only fragmentarily – as a source of CAN-bus signals and a GPS track – without disclosing the transmission architecture, protocols, or the methodology used to build the digital road model. In this study, a three-tier telematics data acquisition and transmission architecture is developed and described, comprising the on-board integration layer with CAN networks via SAE J1939 and the FMS profile, an edge processing layer with navigation positioning via Bluetooth Low Energy, and a cloud platform with a typed binary protocol over TCP/IP and a column-oriented storage model. For the digital road model construction, a high-precision mapping methodology is used, complemented by correction against the global WorldDEM elevation model via spline interpolation, ensuring a relative accuracy of no worse than 0.4 m at a 2 m route discretization. Environmental factors are systematized and their integration principles into the data architecture and the ITS concept are proposed, including the use of V2X channels to transition from a static digital route model to adaptive real-time decision synthesis. Practical validation in a field experiment with a KAMAZ-54901-92 long-haul truck on the R-22 «Kaspiy» route (408 km) confirmed the operability of the proposed solutions, allowing them to be considered as a standalone telematics module for integration into fleet operator platforms and driver decision support systems.

  • Open Access

    Detection of increased fuel consumption in vehicles using deep learning in telematics data analysis

    by Artem S. Dubrovskii, Sergey S. Shadrin, Andrey V. Keller, Daria A. Makarova, Yuriy M. Furletov

    BACKGROUND: Despite the widespread use of telematics data from vehicles to monitor fuel efficiency, existing approaches are mainly focused on after-the-fact analysis and do not ensure early detection of persistent operational deviations in real route conditions. Methods that make it possible to interpretably assess changes in operational power and energy losses in real time, taking into account terrain and traffic conditions, have not been sufficiently studied, which necessitates the development of approaches to the early diagnosis of such deviations. AIM: Development of the method for identifying the causes of increased fuel consumption and formulation of diagnostic recommendations for vehicles in real-world operating conditions and its verification. METHODS: 1) Regression-based identification of reference operational indicators obtained from synchronized CAN, GPS, and digital terrain model data. 2) Training of a gradient boosting regression model for fast prediction of these indicators during subsequent trips. 3) Deviation detection via statistical thresholds with persistence requirement, producing interpretable diagnostic conclusions (chassis, aerodynamics, drivetrain) under route context. RESULTS: A stable correspondence between telematics features and operational indicators was obtained, enabling high-accuracy, low-latency monitoring of their drift. On comparable routes, it is shown that hidden malfunctions caused increased fuel consumption relative to the nominal level, which returned to the baseline after targeted maintenance. The novelty lies in integrating a physically meaningful identification stage with machine learning for fast monitoring and interpretable real-time diagnostics. CONCLUSION: The method of detection of increased fuel consumption ensuring fast diagnostics and decrease of actual consumption without noticeable time losses at the given area is developed and verified experimentally. The method operates with standard telematics without additional sensors, runs on embedded computers, and integrates with corporate maintenance systems. It enables early prevention of defect progression and reduces fuel and maintenance costs. Typical operational scenarios demonstrate fuel savings after eliminating overconsumption causes. The method is relevant for heavy-duty and agricultural machinery.