Search Results

Millions of configurations of power-split hybrid powertrain can be generated due to variation in number of planetary-gear sets (PG), difference in number, type and installation location of shift actuators (clutches or brakes), and difference in connection positions of power components. Considering the large number of configurations, complex structures and control modes, it is vital to construct an appropriate multi-index system evaluation method, which directly affects the requirement fulfillment, the time and cost of 2-PG system configuration design. Considering one-sidedness (dynamics and economic performance), simplicity (linear combination of indicators) and subjectivity (relying on expert experience) of previous system evaluation method of 2-PG system design, a more systematic evaluation method is proposed in this paper. The proposed evaluation system consists of five aspects, involving dynamic, economy, comfort, reliability and cost, and more than 20 indexes.

In this work, the influences of various real-timely available energy management strategies on vehicle fuel consumption (VFC) and energy flow of a range-extender electric vehicle were studied The strategies include single-point, multi-point, speed-following, and equivalent consumption minimization strategy. In addition, the dynamic programming method which cannot be used in real time, but can provide the optimal solution for a known drive situation was used for comparison. VFCs and energy flow characteristics with different strategies under Worldwide Harmonized Light Vehicles Test Cycle (WLTC) were obtained through computer modeling, and the results were verified experimentally on a range-extender test bench. The experimental results are consistent with the modeled ones in general with a maximum deviation of 4.11%, which verifies the accuracy of the simulation models.

The on-ramp merging driving scenario is challenging for achieving the highest-level autonomous driving. Current research using reinforcement learning methods to address the on-ramp merging problem of automated vehicles (AVs) is mainly designed for a single AV, treating other vehicles as part of the environment. This paper proposes a control framework for cooperative on-ramp merging of multiple AVs based on multi-agent deep reinforcement learning (MADRL). This framework facilitates AVs on the ramp and adjacent mainline to learn a coordinate control policy for their longitudinal and lateral motions based on the environment observations. Unlike the hierarchical architecture, this paper integrates decision and control into a unified optimal control problem to solve an on-ramp merging strategy through MADRL.