北斗+5G时空基准的TDOA定位方法研究

TDOA Positioning with BDS+5G Space-Time Datum

第5代移动通信技术(the 5th generation mobile communication technology,5G)引入的毫米波、大规模天线阵列、超密集组网及终端直通等新技术,不但能大幅度提升通信性能,而且具备解决卫星“最后一千米”定位难题的潜力,有望满足城市峡谷、工业物联网等复杂场景下的高精度定位需求。利用5Gn78频段的实际测试数据对基于北斗高精度时空基准的5G下行定位方法进行验证,使用北斗载波相位差分定位与全站仪在室内实验场建立厘米级空间基准,使用北斗授时及带内光纤同步达到纳秒级时间同步精度。对比了Chan算法、泰勒(Taylor)级数展开算法与残差加权(residual weighting,RWGH)算法3种基于到达时间差(time ofdifferential arrival,TDOA)的定位算法在室内热点场景下的定位精度,结果表明,静态定位时残差加权算法精度最优,动态定位时Taylor级数展开算法优于RWGH算法,Taylor级数展开算法与RWGH算法的静态与动态定位精度均可达分米级。其中,非视距误差对TDOA观测值产生约5m的偏差,使用RWGH算法可改善Chan算法在非视距条件下的定位精度。

Objectives:With the establishment of the positioning enhancement project in the 3rd generation partnership project to meet further positioning requirements,the positioning performance of the 5th genera-tion mobile communication technology(5G)has been greatly improved.Currently,researches on 5G com-munication positioning are mainly conducted through channel simulators.Methods:We validated the Chan algorithm,the Taylor series expansion algorithm,and the residual weighting(RWGH)algorithm using real 5G n78 positioning reference signal(PRS)at 3400-3600 MHz from eight pico base stations.Centimeter-level space datum and nanosecond-level time synchronization were built relying on BeiDou satellite naviga-tion system in two typical indoor test scenarios.One 5G receiver was set at several reference points with two antennas to collect static positioning data.For kinematic positioning,we used a handcart with 5G re-ceiver on it to walk along with the reference square traces.Results:(1)The error rates of 5G PRS signal de-modulation were lower than 30%,and the integrity rates were over 80%at 1 Hz sample frequency.(2)The accuracy of time of differential arrival(TDOA)was below O.5 m on average.But during some periods,the TDOA errors caused by non-line-of-sight(NLOS)could be 5 m or even more.(3)In static positioning tests,the accuracy of Chan algorithm was better than 1 m with line-of-sight condition,but the positioning accuracy of Chan algorithm decreased significantly when NLOS occurs.The Taylor algorithm also failed to provide accurate positioning results in those epochs.(4)In kinematic positioning tests,Taylor algorithm achieved 0.77 m accuracy,while the accuracy of RWGH algorithm was 0.91 m.The error of Chan algorithm was larger than 1 m.Conclusions:We have a better understanding about those maximum likelihood estimations and their applications and limitations.The RWGH algorithm shows improvement both in static and kinematic positioning tests,while the Chan algorithm is not suitable for complex indoor scenario with certain multipath effect.Still,the multipath effect and the NLOS error are the key factors to be solved in indoor positioning.Further research on 5G indoor and outdoor location combination with satellite observation and inertial navigation system is also needed.