ID 原文 译文
8444 传统的多核学习需要解复杂的最优化问题,很难直接应用到目标跟踪中,因此提出一种基于boosting学习框架的多核学习算法,使目标跟踪在复杂场景下可以保持跟踪的实时性和准确性。 Traditional multicore learning need to solve complex optimization problems, it is difficult to directly applied to target tracking, therefore proposes a multi-core learning algorithm based on boosting learning framework, the target tracking under complex scene can keep track of real-time and accuracy.
8445 为了进一步减少计算量和提升分类性能,采用极限学习机(extreme learning machine,ELM)作为基分类器,ELM结构简单,训练速度非常快,并且比支持向量机有更好的泛化能力。 In order to further reduce the amount of calculation and improve classification performance, the extreme learning machine (extreme learning machine, ELM) as the base classifier, ELM has simple structure, training speed is very fast, and has better generalization ability than support vector machine (SVM).
8446 最后,将本文算法与其他先进的跟踪算法在多个公开视频序列中进行比较,验证了本文算法性能的有效性。 Finally, the algorithm in this paper and other advanced tracking algorithm in multiple public video sequence comparison, this paper is verified the effectiveness of the algorithm performance.
8447 为了提高接收机自主完好性监测(receiver autonomous integrity monitoring,RAIM)算法对微小缓变伪距偏差的检测能力,提出一种基于抗差扩展卡尔曼滤波和外推-积累的RAIM方法。 In order to improve the receiver autonomous integrity monitoring (receiver autonomous integrity monitoring, RAIM) algorithm for small graded pseudorange deviation detection ability, put forward a kind of extended kalman filtering based on the differential resistance and extrapolation - accumulation RAIM method.
8448 该方法结合了新息外推法和累积历元法在检测微小缓变伪距偏差上的优势,即在新息外推法的基础上,累加多个历元的检验统计量,来更有效地检测微小缓变伪距偏差; This method combines the new rate extrapolation method and cumulative epoch method in detection of tiny graded pseudorange deviation on the advantage, namely the new rate extrapolation method, on the basis of the accumulation of multiple epoch test statistics, to more effectively detect tiny graded pseudorange deviation;
8449 同时,利用抗差扩展卡尔曼滤波对偏差进行抗差处理,提高了定位精度。 At the same time, the deviation to make use of extended kalman filtering resistance is poor poor resistance to treatment, to improve the positioning accuracy.
8450 仿真结果表明,相比较于传统RAIM方法、新息外推法以及累积历元法,新的RAIM方法均提高了对微小伪距偏差正确检测的概率,缩短了对缓变伪距偏差的检测延时,且经过伪距偏差修正后定位精度提高。 Simulation results show that compared with the traditional RAIM method, a new rate extrapolation and accumulative epoch, new RAIM method were raised on small pseudorange deviation correct detection probability, shorten the testing time delay of graded pseudorange deviation, after pseudorange deviation correction and improve location accuracy.
8451 空间站转位是一个长周期过程,需要考虑环境以及自身系统的不确定性因素对整个过程的影响。 Transfer station is a long cycle, need to consider the uncertainty of the environment and its system factors on the impact of the entire process.
8452 使用Kane法建立了空间站系统转位的动力学模型,采用广义混沌多项式方法分析了在转位过程中不确定性因素对空间站姿态的影响。 Using Kane method to establish the dynamic model of transfer station system, the generalized polynomial chaos method analyzes the uncertainty in the process of transfer factor influence on the attitude of the station.
8453 仿真结果表明,gPC方法的计算效率和精度优于蒙特卡罗抽样法,并且空间站若从"零状态"开始转位,则不确定性因素会对转位过程造成较大影响。 The simulation results show that the calculation efficiency and accuracy of the gPC method is better than monte carlo sampling method, and the space station if started from "zero state" transposition, uncertainty factors will counter rotating process caused great influence.