PRORANGE

The solution is devised as a bi-level controller, with the upper layer being the global mission manager and the lower layer acting as a short-term predictive speed controller. Mission manager is a cloud-based optimization algorithm that operates on a longer timescale and captures the complete route in order to ensure energy feasibility throughout the trip. Based on the vehicle state like position, speed and state-of-charge level, the global speed and charging plan is constantly updated and exchanged with the lower layer controller for further efficiency improvement. In contrast to the coarse road discretization on the mission manager, this on-board predictive controller uses a fine-grained road profile and shorter prediction horizon, thus capturing a more accurate representation of necessary vehicle and battery dynamics.

Using our proprietary FORCESPRO solver suite, the two mission optimization algorithms are highly efficient and tractable in real-time, both in on-board and off-board setup.

Similar to our Mission Optimization solution, Dynamic Routing is devised as a bi-level controller, with the upper layer being the fleet-level optimizer and the lower layer regulating the speed profile of each individual vehicle. As such, the upper layer acts as a fleet routing algorithm, and provides a sequence of roads and driving actions to be taken by every vehicle in order to maximize mission execution and avoid unnecessary congestions or waiting times at intersections and arrival locations. Furthermore, the routing solution respects various road regulations like takeover and bypassing as well as the priority of arrival of different vehicles. In contrast, the lower layer represents a modified version of the Mission Optimizer which optimizes the speed/acceleration profile of each vehicle in order to maximize its energy efficiency.

Using our proprietary FORCESPRO solver suite, the two mission optimization algorithms are highly efficient and tractable in real-time.

Models are either provided by the customer or developed specifically for the application. Because of the ability of our solver FORCESPRO to deal effectively with complex and nonlinear optimization, high-fidelity models can be used to reach higher performance. The technology is patent pending.

This energy budget is still not sufficient for constantly driving with maximum energy consumption, so the teams have to find smart methods to save on energy and concurrently keeping the highest possible pace. In contrary to precomputed strategies, computing optimal strategies online on the car facilitates flexible adaptation to eventualities such as full course yellow periods or safety car deployments. For this reason, the software needs to be capable of dealing with all foreseen and unforeseen eventualities during the race.