PARTIALLY-OBSERVABLE AUTOMATIC DRIVING DECISION-MAKING METHOD AND SYSTEM BASED ON CONSTRAINT ONLINE PLANNING

摘要

A partially-observable automatic driving decision-making method and system based on constraint online planning. The decision-making method is mainly used for automatic driving and making a vehicle have multiple driving modes. The system comprises a driving environment state unit, a searching unit, a simulating unit, and a cost constraint unit. By means of the decision-making method, the present invention not only can generate a driving scheme according to a current driving environment, but also adjusts the scheme in real time according to real-time road conditions and vehicle conditions so as to improve the flexibility of the scheme. A history-based Monte Carlo search tree is constructed by means of the method, and the simulating of a solution is performed on the basis of real conditions, so that reliability is improved. In addition, certain optimal selection conditions are met in the method, so that an obtained strategy is ensured to be a random strategy, and the defect of a deterministic strategy is remedied. By means of the decision-making method, the present invention fully meets the driving requirements of current general users, and in particular, multiple modes are provided for selection, so that user experience is greatly improved.