Browsing by Keyword "Predictive models"
Now showing 1 - 2 of 2
Results Per Page
Sort Options
Item Transfer Learning and Online Learning for Traffic Forecasting under Different Data Availability Conditions: Alternatives and Pitfalls(Institute of Electrical and Electronics Engineers Inc., 2020-09-20) Manibardo, Eric L.; Laña, Ibai; Del Ser, Javier; IAThis work aims at unveiling the potential of Transfer Learning (TL) for developing a traffic flow forecasting model in scenarios of absent data. Knowledge transfer from high-quality predictive models becomes feasible under the TL paradigm, enabling the generation of new proper models with few data. In order to explore this capability, we identify three different levels of data absent scenarios, where TL techniques are applied among Deep Learning (DL) methods for traffic forecasting. Then, traditional batch learning is compared against TL based models using real traffic flow data, collected by deployed loops managed by the City Council of Madrid (Spain). In addition, we apply Online Learning (OL) techniques, where model receives an update after each prediction, in order to adapt to traffic flow trend changes and incrementally learn from new incoming traffic data. The obtained experimental results shed light on the advantages of transfer and online learning for traffic flow forecasting, and draw practical insights on their interplay with the amount of available training data at the location of interest.Item A Vehicle Simulation Model and Automated Driving Features Validation for Low-Speed High Automation Applications(2021-12-01) Matute-Peaspan, Jose Angel; Zubizarreta-Pico, Asier; Diaz-Briceno, Sergio E.; Tecnalia Research & Innovation; CCAMThe low-speed high automation (LSHA) is foreseen as a development path for new types of mobility, improving road safety and addressing transit problems in urban infrastructures. As these automation approaches are still in the development phase, methods to improve their design and validation are required. The use of vehicle simulation models allows reducing significantly the time deployment on real test tracks, which would not consider all the scenarios or complexity related to automated driving features. However, to ensure safety and accuracy while evaluating the proper operation of LSHA features, adequate validation methodologies are mandatory. In this study a two-step validation methodology is proposed: Firstly, an open-loop test set attempts to tune the required vehicle simulation models using experimental data considering also the dynamics of the actuation devices required for vehicle automation. Secondly, a closed-loop test strives to validate the selected automated driving functionality based on test plans, also improving the vehicle dynamics response. To illustrate the methodology, a study case is proposed using an automated Renault Twizy. In the first step, the brake pedal and steering wheel actuators' behavior is modeled, as well as its longitudinal dynamics and turning capacity. Then, in a second step, an LSHA functionality for Traffic Jam Assist based on a Model Predictive Control approach is evaluated and validated. Results demonstrate that the proposed methodology is capable not only to tune vehicle simulation models for automated driving development purposes but also to validate LSHA functionalities.