Sample collection, construction and prediction methods for road surface type prediction model, and equipment

The invention discloses a sample collection, construction and prediction methods for a road surface type prediction model, and equipment, and relates to the technical field of automatic driving. The sample collection method comprises the following steps: selecting at least one road surface type in a virtual driving scene as a scene in an intelligent driving simulation system in which the virtual driving scene and a vehicle dynamics model are pre-stored, and carrying out a simulation driving test; collecting vehicle dynamic parameters corresponding to each road surface type; and associating the vehicle kinetic parameters with the corresponding pavement types to obtain a sample set. The pavement type can be accurately predicted. 本发明公开了路面类型预测模型的样本采集、构建、预测方法及设备，涉及自动驾驶技术领域。样本采集方法包括：在预存储有虚拟驾驶场景和车辆动力学模型的智能驾驶仿真系统中选择所述虚拟驾驶场景中的至少一种路面类型为场景进行模拟驾驶试验；采集每一种路面类型对应的车辆动力学参数；将所述车辆动力学参数和与其对应的所述路面类型关联，得到样本集。本发明能够准确地预测路面类型。