In January, the agency had issued a request for information (DE-FOA-0001473) seeking input from researchers and developers in a broad range of disciplines including automotive vehicle control, powertrain control and transportation analytics regarding the development of advanced energy efficiency optimization technologies for future connected and automated vehicles (CAVs). This input is now reflected in the NEXTCAR FOA.

The technologies to be developed for NEXTCAR will be required to demonstrate a 20% reduction in energy consumption when implemented on a 2016 baseline vehicle.

Examples of potential technologies include, but are not limited to, advanced technologies and concepts relating to full vehicle dynamic control, powertrain control, improved vehicle and powertrain operation through the automation of vehicle dynamics control functions, and improved control and optimization facilitated by connectivity.

These improvements include the reduction of the fuel and/or energy consumed by future individual vehicles undergoing either human operation or semi- or fully-automated operation, either in isolation or in cooperation with other vehicles.

Solutions that only take into account vehicle-level longitudinal (or vehicle dynamic) control or driver behavior optimization without regard for optimized powertrain operation are unlikely to achieve the energy efficiency goals sought by this Program. In essence, the co-optimization of vehicle-level (vehicle dynamic) and powertrain-level operations is sought in order to minimize the energy consumption of future vehicles.

… From a control point of view, vehicles currently operate in isolation as a collection of single ‘selfish’ entities, even in dense traffic. Developments in connectivity and automation will allow vehicles in the future to operate in a range of cooperative modes with other surrounding vehicles. While such cooperative behavior has been the subject of much recent research, the full potential of improved powertrain control (as opposed to improved vehicle longitudinal or dynamic control) on the resultant composite energy efficiency of a cohort of vehicles undertaking cooperative vehicle behavior5 has not yet been fully explored. The focus of the ARPA-E NEXTCAR Program is on increasing the energy efficiency of each individual vehicle in the automotive fleet, through the improvement of vehicle dynamic and powertrain (VD&PT) control, by utilizing emerging technologies and strategies in sensing, communications, information, decision-making, control and automation.

—DE-FOA-0001564

The optimization of the operation or energy efficiency of Level 4 (L4) autonomous vehicles is beyond the desired scope of the NEXTCAR Program, which emphasizes applications from L0 to L3 levels of automation.

Top: Vehicle dynamic and powertrain control and actuation logic flow diagram for a L0 vehicle. Bottom: Vehicle dynamic and powertrain control and actuation logic flow diagram for future L3-L4 vehicles. Under the NEXTCAR Program, connectivity will allow the use of additional information inputs to the vehicle and powertrain controllers. Click to enlarge.

Conventional powertrain control at present is almost exclusively reactive and backward-looking, with limited provision for the incorporation of sensor-based feedback except for crude or indirect measures of combustion efficiency and/or exhaust emissions, the agency notes in the FOA. As a result, powertrain operation is frequently rendered non-optimal with regard to fuel and energy consumption minimization, and considerable opportunity arises for energy efficiency optimization, with the required computation either performed on-board in real-time or on an off-line basis.

Vehicle connectivity, however, enables the use of additional, exogenous inputs for improved real-time vehicle and powertrain control. Such inputs could potentially be used to create a specific time-based trajectory of optimized powertrain control references to minimize the fuel or energy consumption of each individual vehicle across some finite future time horizon.

The creation or addition of additional high-value information that can be made available through V2X for use in powertrain control systems may also enable significantly higher individual vehicle efficiency through combustion optimization (in the case of ICVs or HEVs), energy storage optimization (in the case of HEVs and BEVs), and route optimization and optimized vehicle dynamic performance for all vehicles. For ICVs or HEVs, the addition of “perfect” information on fuel chemistry, engine and after-treatment conditions, weather and environmental conditions, traffic conditions ahead, and perhaps driver behavior (for example), could lead to meaningful enhancements in the energy efficiency of each and every vehicle under a range of operating conditions and use cases.

One promising enabling technology underlying future vehicle and powertrain control is the development of model-based control algorithms and systems—this will allow powertrain control to be fully predictive and forward-looking, and enhance the effect of real and virtual feedback, as well as utilizing a range of additional information available through connectivity. With this increased information, model-based control using real-time optimization has the potential for useful efficiency gains for individual vehicles, and hence by extension, the entire vehicle fleet.

—DE-FOA-0001564