农林环境下RTK-UWB多传感器融合定位方法

RTK-UWB multi-sensor fusion positioning method in agroforestry environment

开发一种适用于农林环境的高精度多传感器融合定位方法,以解决全球导航卫星系统(global navigation satellite system,GNSS)在农林环境下定位中受到树木或地形遮挡等影响导致信号不稳定、对定位精度造成影响的问题。提出了一种基于无迹卡尔曼滤波(unscented Kalman filter,UKF)的多传感器融合定位方法,该方法结合了实时运动动力学模型和来自不同传感器的数据,包括差分GNSS(real-time kinematic,RTK)、超宽带(ultra-wideband,UWB)、惯性测量单元(inertial measurement unit,IMU)和轮速计。通过使用无迹卡尔曼滤波进行状态估计,使用不同传感器组合的后验估计过程实现高精度的定位,并在RTK信号失效时保持稳定性。通过实地试验验证了该方法的性能,并与其他相关方法进行了对比分析。在RTK信号正常工作时,本研究方法的最大定位误差为3.0 cm。本研究提出的多传感器融合方法在RTK失效状态下能够保持稳定的定位性能,融合定位方差为1.0 cm,最大偏差小于4.0 cm,且未出现定位跳变或发散现象,相较于原始信号误差减少约96%,具有较好的稳定性,基本消除了RTK信号造成的影响。该方法具有广泛的应用前景,可在农业、林业等领域中应对GNSS信号不稳定带来的挑战,为定位技术的进一步发展提供支持。

This investigation sets out to devise an intricate,yet robust multi-sensor fusion positioning methodology explicitly tailored for the demanding terrains of agricultural and forestry landscapes.The primary objective revolved around countering the inherent uncertainties plaguing the stability of global navigation satellite system(GNSS) signals.Central to this endeavor is the development of a multi-sensor fusion positioning technique anchored on the bedrock of the unscented Kalman filter(UKF).This innovative approach intricately weaved together real-time motion dynamics models with data streams harnessed from an extensive array of sensors.These included,but are not limited to,real-time kinematic,ultra-wideband,inertial measurement units,and wheel encoders.The brilliance of employing UKF for state estimation lies in its ability to yield positioning precision that transcends the centimeter scale,navigating seamlessly through a spectrum of sensor amalgamations,and steadfastly maintaining stability even in the wake of erratic RTK signals.To validate the potency of this novel approach,a series of meticulously designed field experiments were meticulously conducted.Through a comprehensive comparative analysis against existing methodologies,the findings under optimal RTK signal conditions showcased a mere 3.0 cm maximum positioning error-a testament to the precision of this research methodology.More strikingly,the resilience of the proposed multi-sensor fusion technique revealed its capacity to sustain positioning accuracy in the absence of a functional RTK signal.It demonstrated a fusion positioning variance of 1.0 cm,a maximal divergence under 4.0 cm,and the conspicuous absence of any positioning oscillations or divergence phenomena.Relative to the inherent inaccuracies within signals,the proposed methodology presented a staggering reduction of nearly 96%,underscoring its exceptional stability and its robustness in mitigating the adversities stemming from RTK signal degradation.This bespoke multi-sensor fusion positioning technique stands as a beacon of reliability in the specific realms of agricultural and forestry applications.Its unwavering provision of positioning data,even in the face of faltering RTK signals,offers a pragmatic solution to the persistent challenges arising from the capricious nature of GNSS signal consistencies.Consequently,this pioneering approach heralds substantial promise across a multifaceted spectrum of applications,transcending the realms of agriculture and forestry to encompass diverse domains.Its potential to effectively navigate and neutralize the impediments posed by GNSS signal instability underscores its pivotal role in fortifying the continuous evolution of positioning technology.