Multi-Agent Based Driving - Range Prediction and Energy Optimization for Electric Vehicles
Although critical to a CO2 neutral environment, the acceptance of electric vehicles is still lacking in comparison to their gasoline powered counterparts. This is mainly due to their higher price in combination with their driving range limitations. One approach to solve these issues is to develop more efficient batteries, capable of storing more energy, while reducing size and costs. The other approach, addressed in this work, is to utilize the available energy at best. To this aim, an intelligent prediction and energy management system (EMS) should be developed, to be installed on board.Copyright: RWTH Aachen
ACS is involved in Project RACE, funded by the BMWi, with the goal of developing a highly extendable, robust and efficient ICT architecture for future electric vehicles. This project teams up eight partners from industry and university. RWTH Aachen University is represented by the institutes ACS and ISEA. Within this project, ACS is developing the EMS for the use in future electric vehicles, to extend their driving range. This system must be safe, reliable and highly efficient. To meet all these requirements the system is designed in a multiagent framework technology.
Within the multiagent distributed approach, each entity, the so called agent, is capable of performing control tasks on its own. In the EMS the agents are assigned to all the energy consuming and generating units in the vehicle‘s electrical onboard network. These agents cooperate with one another to lead the system to the overall optimal operating point regarding energy consumption, comfort and the ability to reach the desired destination.
For the multiagent EMS to operate it is necessary to be involved in the whole process of a full drivecycle. This covers information collection and also intervening during the drive. The process is starting with route planning and ends when the destination is reached. This requires event and energy consumption prediction as well as continuous energy management. Therefore, the EMS generates an estimation of the expected driving range along the shortest path to the destination based on realworld map data, including elevation and speed limitations as well as weather, driver and passenger data. With these data the driver is able to check how far the vehicle can drive and the EMS is able to optimize the actual energy demand by using a market based trading approach with energy as the resource to be bought and sold. This is done by the agents according to the preestimation and actual situation to make the system efficient, safe and comfortable. The EMS is designed to react immediately to aberrations by reallocating the available energy. This can lead to small variations in comfort and speed during the drive but it tries to ensure to reach the destination within certain limits. The EMS also keeps the driver informed about the actual system state and the possibility of reaching the desired destination.
This video is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.