Accurate mirror-based camera pose estimation with explicit geometric meanings

An accurate closed form solution is proposed to estimate camera pose by several mirrored reference object images acquired via a planar mirror under different unknown poses. Compared with state-of-the-art methods, our method is more accurate when there are more than three images and has explicit geometric meanings. This method also properly handles cases in which some of the mirror poses are parallel. The central idea is to minimize an error metric based on all reflections of rotation, which enables the camera rotation to be estimated directly by SVD of sum of mirrored camera rotations. After that, the camera translation is computed by solving a large system of linear equations to minimize object space collinearity error. Both synthesized data and real data experiments show the advantages of our approach.     

规范状态空间系统辨识方法

因为状态空间模型既包含了未知状态,又包含了未知参数,且二者是非线性乘积关系,使得辨识问题变得复杂.针对这一问题,详细研究了规范状态空间系统的状态与参数联合估计方法.采用交互估计理论,即采用递推方法或迭代方法实现系统状态与参数的交互估计.基本思路是在计算参数估计时,辨识算法信息向量中的未知状态用其估计值代替,然后利用获得的参数估计,设计基于参数估计的状态观测器或基于参数估计的Kalman滤波算法估计系统的状态,二者形成一个交互计算过程(递阶计算过程).沿着这条思路,分别从递推方案和迭代方案,研究和提出了基于状态观测器和基于Kalman滤波状态估计的随机梯度辨识算法、递推最小二乘辨识算法、多新息随机梯度辨识算法、多新息最小二乘辨识算法,以及模型分解的辨识算法,并给出了几个典型算法的计算步骤、流程图和计算量.     

电子执法环境下交通违法行为影响因素分析

为了甄别电子执法环境下影响交通违法行为的主要因素,在交通调查基础上,借助广义估计方程分析了有无电子执法设施、交通标线是否齐全、车辆类型和交叉口服务水平4个独立因素对交通违法行为的影响.结果表明:有无电子执法设施和交通标线是否齐全对交通违法行为的影响显著;这两个显著影响因素的交叉作用分析显示,在交通标线齐全时电子执法对交通违法行为具有明显的抑制作用.齐全的交通标线对抑制电子执法环境下的交通违法行为、提高道路交通安全水平具有重要作用,它是实现电子执法设施建设初衷的基本保证.     

Novel method for frequency hopping signals blind identification in the compressed domain

A novel method to blindly identify frequency hopping signals is presented in the compressed domain. The samples obtained by compressive sampling effectively maintain the structure of and the information on the original signal, so the task of identification of the original signal could be done by directly processing the sampling values. The method is based on the difference in numerical characteristic between sampling values. According to the different characteristics of the expectation of sampling values under different hypotheses, identification is accomplished by using the deviation of the actual sampling values from the expectations under the corresponding hypothesis as the criterion. Without reconstructing the frequency hopping signal itself,hopping frequencies can be estimated through a small number of measurements by the compressive sampling algorithm. Simulation results have proved that the proposed method is adequate to the environments in which the signal-to-noise ratio is higher than -2dB. Meanwhile, compared with other traditional methods, the proposed algorithm greatly reduces the amount of data, the computational complexity, and the identification time.     

2-D DOA estimation of coexisting uncorrelated and coherent signals

The common two dimensional (2-D) direction of arrival (DOA) estimation algorithms for coexisting uncorrelated and coherent signals are based on the complex array structure, such as the uniform rectangular array, so the computational complexity is high and the array aperture is not utilized efficiently. By taking advantage of the L-shape array and adopting an efficient method to eliminate the Gaussian noise, a new 2-D DOA estimation method is proposed. Firstly, the DOAs of the uncorrelated signals are estimated and the influence of the coherent signals is eliminated by utilizing its characteristics. Then, the data covariance matrix containing the coherent information only is obtained by exploiting the Toeplitz property of the uncorrelated signals, and the DOAs of the coherent signals are estimated by the direction finding method based on the compressed sensing theory. Theoretical analysis and simulation results show that the proposed method has a small computational load, high array aperture as well as excellent estimation performance.     

Propagator method for multiple-parameter estimation of LFM signal for conformal array antenna

A multiple parameter estimation method for the LFM signal for the conformal array antenna is presented. By employing the energy focusing property of certain-ordered-FRFT to the LFM signal, multiple components blended in received data are separated in the FRFT domain, and hence the simplified data model is acquired, because of which the complicated parameter match is avoided, and then the DOA and polarization estimation is performed by using the PM algorithm. Benefiting from the avoiding of parameter matching and high dimensional eigen-value decomposition, the complexity and computational burden of the algorithm is remarkably controlled. Simulation results show that the proposed algorithm has an approximate performance with a much lower computation burden than ESPRIT.     

Motion compensation approach for wide-swath missile-borne SAR imagery

In wide-swath missile-borne synthetic aperture radar (SAR) imaging, the range-variant residual phase modulation, which is introduced by the presence of trajectory deviations, will seriously degrade the focusing quality of the final image. To deal with this problem, a motion compensation approach for missile-borne SAR is proposed. In the approach, the signal model of motion error for diving movement is established. Then, according to the inertial navigation system (INS) information, the range-variant residual Doppler rates of the echoes are extracted by using the Doppler rate estimation, on the basis of which the corresponding phase error can be calculated and compensated. Simulation results show that, by using the proposed approach, the influences of the motion error on focusing are greatly reduced with the well focused image obtained.     

线性模型中回归系数的c-F(K)型估计的研究和应用

针对引起线性回归模型最小二乘估计性能变坏的根本原因,在线性回归模型中提出了一种新的有偏估计β^(F(K))=(cX'V-1X+TF(K)T')-1X'V-1Y,它将众多经典的有偏估计结合在一起,对有偏估计进行了改进.首先利用Stine式压缩技术说明在均方误差意义下β(F(K))=(cX'V-1X+TF(K)T')-1X'V-1Y,它将众多经典的有偏估计结合在一起,对有偏估计进行了改进.首先利用Stine式压缩技术说明在均方误差意义下β^(F(K))具有优于广义岭估计的优良性质,然后将β(F(K))具有优于广义岭估计的优良性质,然后将β^(F(K))应用到多元线性回归模型的可容许性上.     

一种改进的基于秩检测的信道有效阶数估计算法

信道阶数估计是SOS类盲辨识算法的前端步骤。针对Liavas阶数估计算法对信噪比要求高的问题,通过对接收序列自相关矩阵进行"降噪"处理和改进秩检测目标函数,提出一种改进算法,有效降低了对信噪比的要求。分析了在不同信道条件下,各种阶数估计算法的性能,阐明随着信噪比降低,阶数估计误差不可避免,该结论是进一步研究对阶数估计误差鲁棒的盲辨识算法的重要依据。     

电液振动台随机振动控制算法

为提高电液振动台随机振动的控制精度,分析传统频率响应函数H1估计法,并针对其存在的欠估计问题进行改进,提出基于EV模型的频率响应函数估计方法。此方法能对传统频率响应函数H1估计法的估计结果进行修正,使修正后的频率响应函数更接近系统真实频率响应函数,并通过加速度功率谱再现实验对算法的有效性进行验证,同时给出基于EV模型的随机振动控制算法的实现流程。