Services for databased optimalization of battery electric vehicles - planning and surveillance

The El The bus is stuck in Birkelundsbakken - battery is completely discharged!" This is a newspaper article we do not want to see for the winter. In the project "Service for data-driven optimization of battery-electric means of transport - planning and monitoring", we will work with better planning tools so that the bus takes you easily to work without running out of power. When using artificial intelligence, we will find methods and planning tools where we calculate the bus's travel route, how much up and down it will travel, and how many passengers are transported. We will also use other sources, such as the weather today or the weather next week, because electric vehicles use more electricity when it is cold than when it is hot. In this way, bus companies can get the best advice, based on artificial intelligence, machine learning, and historical data so that they can both plan the route, plan the charging of batteries, and utilize the vehicles in a better way. The project will help make the transition from fossil to zero-emission vehicles easier. By contributing to predictable and stable public transport solutions, through the reduction of downtime, it will appear as a good alternative to car use. Investment in battery-electric solutions will also be made possible in the districts. It will also focus on how smart solutions can contribute to more sustainable utilization of resources.

The climate challenges due to global warming affect our everyday life more and more. It is now urgent to reduce emissions of greenhouse gases. In order to reduce emissions from the transport industry, there is a need to replace fossil vehicles with zero-emission vehicles. This entails some challenges for the transport operator companies.For battery-electric vehicles, everyday life is different than for traditional vehicles. Battery capacity is limited, and usable capacity changes with weather and road conditions. Planning a grid, or the allocation of resources, is challenging when the framework conditions are constantly changing. At present, this is solved by oversizing battery packs and trying to have good margins so that interruptions do not occur. Here, the planner's experience and rough estimates are often the basis. The effect of missing tools is increased use of natural resources in the production of extra-large battery packs to ensure flexibility. It is also required that several vehicles are in operation and on standby to ensure available capacity to avoid service interruptions.We want to develop a service that strengthens operational security and reduces resource use. The service must be able to predict consumption on routes based on altitude profile, weather, road, traffic pattern and load. This should facilitate the planning and disposition of vehicles in the short term (days) and long term (years).Central R&D challenges are that real-time data from different sources will have different time resolutions, is the uncertainty of the models affected? When does the uncertainty in the predictions become too great due to the outcome of data sources? Is it better to use a "black-box" for modeling expected consumption or to use additive models of individual contributions? How large ensembles are needed to get a good estimate under different conditions?