多特征融合的道路车辆检测方法

建立了当贝叶斯风险一定时，发送时间与干扰之间关系的模型.在此基础上，提出了2种满足干扰约束的算法，单用户感知模式下最大化系统吞吐量时的帧长设计算法和多用户合作感知模式下基于纽曼-皮尔逊准则的帧长设计算法.仿真结果表明，同已有算法相比，由于帧长的优化，即使贝叶斯风险较大，新算法也能取得相对较优的吞吐量.

Fusing Multiple Features to Detect On-road Vehicles

This paper first derived the function between the transmission time and the interference for given Bayesian risk value. On this basis, the following two interference constraint algorithms were proposed, i.e., the non-cooperative sensing mode frame length design algorithm to maximize system throughput and the cooperative sensing mode frame length design algorithm based on Newman-Pearson criterion. Simulations results have shown that, even if the Bayesian risk value is large, the proposed algorithm can achieved better throughput due to the optimization of the frame length, compared with the existing schemes.