智能网联汽车定位算法计算优化

Optimization of Intelligent Connected Automobile Positioning Algorithm Calculation

智能网联汽车要实现自动驾驶,需要连续、可靠的定位信息。多传感器融合定位技术是智能网联汽车的必备解决方案。在多传感器融合系统中,惯性测量单元是必不可少的传感器。但是惯性导航的机械编排计算量大,实时定位效率较低。实际上,惯性导航算法在速度更新时,无须进行旋转效应和划桨效应补偿。从实测结果可知,省略两项高阶补偿项,不仅不会影响其定位效果,也能降低定位算法的计算量,从而满足智能网联汽车自动驾驶的低时延要求。

Intelligent connected automobiles require continuous and reliable positioning information to achieve autonomous driving.Multi-sensor fusion positioning technology is an essential solution for intelligent connected automobiles.In multi-sensor fusion systems,inertial measurement units are essential sensors.However,the mechanical arrangement of inertial navigation requires a large amount of computation,and the real-time positioning efficiency is relatively low.Actually,the inertial navigation algorithm does not require compensation for rotation and paddle effects during speed updates.From the actual test results,it can be seen that omitting two high-order compensation terms not only does not affect its positioning effect,but also reduces the computational complexity of the positioning algorithm,thereby meeting the low latency requirements of intelligent connected automobile autonomous driving.