首页|基于高斯过程的鸽群归巢导航轨迹预测

基于高斯过程的鸽群归巢导航轨迹预测

扫码查看
针对鸽子在熟悉区域的导航行为学特性,提出了一种基于高斯过程的鸽群"即时归巢区域"导航轨迹预测模型.基于鸽子熟悉区域释放归巢试验的试验现象,对现有的鸽子归巢数据进行统计分析,考虑个体不相关性、时间关联性以及观测误差等建立基于高斯过程的导航轨迹预测数学模型,预测鸽子在熟悉区域的飞行轨迹,并从行为学和数学模型上论证了鸽子飞行轨迹的收敛性,从不完全的观测信息中定量给出鸽子的潜在习惯路线结构及变化范围.仿真实验基于鸽群真实归巢飞行数据验证了预测模型的一致性和合理性,为信息不完全条件下的无人机自主导航轨迹预测提供方法支撑.
Navigation Trajectory Prediction of Homing Pigeon Flocks Based on Gaussian Process
Based on the navigation characteristics of homing pigeon flocks in familiar areas,a navigation trajectory prediction model based on Gaussian process for pigeon flocks in"immediate homing area"is proposed.Based on the exper-imental phenomenon of pigeon homing experiment in familiar area,the existing pigeon homing data is statistically analyzed.Considering the individual irrelevance,temporal correlation and observation errors,a mathematical model of navigation traj-ectory prediction based on Gaussian process is established to predict the flight trajectories of pigeon flocks in familiar area.The convergence of pigeon flight trajectories is also demonstrated from behavioral and mathematical models,and the poten-tial habitual route structure and the range of change of pigeons are given quantitatively from incomplete observation informa-tion.The simulation experiment verifies the consistency and rationality of the prediction model based on the real homing flight data of pigeon flocks,which provides methodological support for the prediction of autonomous navigation trajectory of unmanned aerial vehicles with the incomplete information.

homing pigeon flockbio-inspired navigationtrajectory predictionGaussian process(GP)

霍梦真、辛龙、段海滨

展开 >

北京航空航天大学自动化科学与电气工程学院,飞行器一体化控制全国重点实验室,北京 100083

北京宇航系统工程研究所,北京 100076

鸽群归巢 仿生导航 轨迹预测 高斯分布

2024

导航与控制
北京航天控制仪器研究所

导航与控制

CSTPCD
影响因子:0.133
ISSN:1674-5558
年,卷(期):2024.23(4)