Search Results

With the development of brake-by-wire technology, electro-hydraulic composite braking technology came into being. This technology distributes the total braking force demand into motor regenerative braking force and hydraulic braking force, and can achieve a high energy recovery rate. The existing composite braking control belongs to single-channel control, i.e., the four wheel braking pressures are always the same, so the hydraulic braking force distribution relationship of the front and rear wheels does not change. For single-axle-driven electric vehicles, the additional regenerative braking force on the driven wheels will destroy the original braking force distribution relationship, resulting in reduced braking efficiency of the driven wheels, which are much easier to lock under poor road adhesion conditions.

For intelligent vehicles, a fast and accurate estimation of road slope is of great significance for many aspects, including the steering comfort, fuel economy, vehicle stability control, driving decision-making, etc. But the commonly used estimation methods nowadays usually demand additional sensors or complex dynamic models, causing increase in system complexity as well as decrease in accuracy. To solve these problems, this paper puts forward a real-time road slope estimation algorithm leveraging the relationship between pitch angle and road slope, which only requires low sensors cost and computational complexity. Firstly, a GNSS/INS fusion system is established to obtain the pitch angle with respect to the navigation frame, which couples the vehicle’s pitch angle in vehicle frame and road slope angle.

Increasing the degree of individuality of the autopilot and adapting it to the habits of drivers with different driving styles will help to increase occupant acceptance of the autopilot function. Inspired by the Twin Delayed Deep Deterministic policy gradient algorithm(TD3) algorithm to increase action spontaneity, this paper proposes a Soft Actor-Critic(SAC) based personalized following control strategy to increase the degree of strategy personalization through driver data. In order to obtain real driver data, this paper collected driving data based on driver-in-the-loop experiments conducted on a simulated driving platform, and selected data from three drivers with distinctive driving characteristics for model training. A continuous action space model was developed by vehicle following kinematics. A temporal Gate Recurrent Unit (GRU) based reference model is trained to receive temporal state signals and output acceleration actions according to the current state.