GNSS/INS松组合导航是目前应用最广泛的车载导航系统之一,但在长隧道、地库等遮蔽区域,卫星信号长时间失锁导致定位精度显著下降。本研究提出一种Tent-ASO-BP辅助的GNSS/INS松组合导航算法。首先,结合混沌帐篷映射(Tent)改进的原子搜索算法(ASO)优化BP神经网络模型的权值及阈值,构建Tent-ASO-BP智能预测模型;然后,利用开阔环境下GNSS/INS导航数据训练Tent-ASO-BP智能模型,在GNSS隧道失锁环境下利用自主学习后的Tent-ASO-BP模型预测隧道内的位置参数;最后,利用车载实测数据进行验证。结果表明,Tent-ASO-BP预测模型总体精度明显高于GNSS/INS松组合模型精度,Tent-ASO-BP预测模型的水平方向误差为15.4394m;GNSS/INS松组合误差为20.4292m,水平精度提升了24.42%,预测模型能够有效解决卫星信号长时间失锁时GNSS/INS松组合导航连续高精度定位难题。

GNSS/INS integrated navigation is one of the most widely used vehicle navigation systems. However, the positioning accuracy in shielded regions such as long tunnels and basements is significantly degraded due to the long- term locking of satellite signals. To address this problem, we proposed a Tent-ASO-BP aided GNSS/INS integrated navigation algorithm. Firstly, the weight and threshold of (back propagation,BP) neural network model were optimized by combining chaotic tent map and atom search algorithm (ASO) to construct Tent-ASO-BP intelligent prediction model. Then, the intelligent prediction model was trained by using GNSS/INS integrated algorithm data collected on outdoor open areas. The well-trained Tent-ASO-BP model was used to predict the position parameters in the GNSS outage regions. Finally, vehicle field tests were performed to verify the availability of the Tent-ASO- BP model. Experimental results show that the overall accuracy of the Tent-ASO-BP prediction model is significantly higher than that of the GNSS/INS model. The root mean square error of the Tent-ASO-BP prediction model in the horizontal direction is 15.439 m while the GNSS/INS model is 20.429 m, and the horizontal accuracy is increased by 24.42%. The proposed model can effectively address the problem of continuous high-precision positioning of GNSS/INS integrated navigation during GNSS outages.