Development of Argonne VIL workflow for CAVs to demonstrate the energy-efficient automated driving controls in real vehicles
2022-01-0500
03/29/2022
- Event
- Content
- Argonne National Laboratory (ANL) has developed automated driving control strategies that aim at maximizing energy-efficiency. Up to 22% energy savings were demonstrated [1] in a large-scale study performed in RoadRunner, a tool for the energy-centric modeling of connected and automated vehicles (CAVs). This paper describes how these controls were implemented and tested in real vehicles on dynamometer setting, ANL vehicle in the loop (VIL) workflow, to demonstrate both the efficacy of the controls themselves, but also to validate the RoadRunner models. The main concept of the ANL VIL workflow is to build a digital twin and connect to a real vehicle on dynamometer as if it can drive on the real road. The ANL VIL workflow mainly consists of three steps, integration, preparation, and experimentation. A digital twin of the vehicle and the testing environment was built by the help of RoadRunner for the integration. The digital twin includes virtual road such as intersection, traffic signals, speed limit changes, and other vehicles. According to the developer’s test scenario design, the digital twin is automatically generated using RoadRunner and will be implemented into hardware to feed the virtual inputs to the control and actual vehicle on dynamometer. The energy-efficient automated driving controls are also modified for the real-time hardware integration and pre-tested using RoadRunner during this step. Second, the integration models were implemented into the hardware, dSpace Micro-Autobox in the preparation step. For the ease of implementation, a toolbox model required for the hardware implementation were developed in RoadRunner and it can be automatically built with other digital twin models simultaneously in the integration step. After all the digital twin models and control have been implemented, a preparation process such as warm-up has been conducted. In the experimentation step, test has been conducted on dynamometer as planned. An automated quality-check process was setup in RoadRunner to check the proper functionality of the control in a broad range of situations - a subset of these checks will be also performed in the actual vehicles during or after the tests. The test results are then analyzed and compared to results from simulations focusing on basic functionality and replicating the QC scenarios first. As a result, the optimal controllers with RoadRunner were validated to confirm the impact of optimal control for CAVs for a single vehicle. Future testing will involve more complex scenarios, including the presence of other vehicles. [1] E. Rask et al., “SMATR Mobility Connected and Automated Vehicles Capstone Report,” DOE EEMS Vehicle Technologies Office Capstone Reports, July, 2020.
- Citation
- JEONG, J., Kim, N., Di Russo, M., Grave, J. et al., "Development of Argonne VIL workflow for CAVs to demonstrate the energy-efficient automated driving controls in real vehicles ," SAE Technical Paper 2022-01-0500, 2022, .