Optimization-Based Real-Time-Capable Energy Strategy for Autonomous Electric Race Cars

Solving a Minimum Lap Time Problem (MLTP), under the constraints stemming from a race car’s driving dynamics, can be considered to be state of the art. Nevertheless, when dealing with electric race vehicles as in Formula E or the Roborace competition, solving an MLTP is not enough to form an appropriate competition strategy: Maximum performance over the entire race can only be achieved by an optimization horizon spanning all the subsequent laps of a race. This results in a Minimum Race Time Problem (MRTP). To solve this, the thermodynamic and energetic limitations of the electric powertrain components must be taken into account, as exceeding them leads to safety shutdowns. Therefore, we present an Optimal Control Problem (OCP) to calculate an Energy Strategy (ES) for electric race cars, which contain physically detailed descriptions of its powertrain components. Leveraging a Sequential Quadratic Programming (SQP) solver, the OCP can be solved numerically in real time. This enables the ES to be recalculated during a race. As a consequence, powertrain overheating can be omitted and the limited amount of stored battery energy utilized as efficiently as possible. Simultaneously, the race can be completed in minimum time.     «

Solving a Minimum Lap Time Problem (MLTP), under the constraints stemming from a race car’s driving dynamics, can be considered to be state of the art. Nevertheless, when dealing with electric race vehicles as in Formula E or the Roborace competition, solving an MLTP is not enough to form an appropriate competition strategy: Maximum performance over the entire race can only be achieved by an optimization horizon spanning all the subsequent laps of a race. This results in a Minimum Race Time Prob...     »