32results about How to "Avoid overestimation" patented technology

An object recognition apparatus performs a sweep of a scanned region by transmitting scanning wave beams at respective scan angles and successive timings, derives a received-wave signal strength value and reflection location corresponding to each scan wave beam during the sweep, and assigns a set of mutually adjacent reflection locations as a segment. A corresponding range value of the segment is calculated, expressing an estimated distance of a detected object. A threshold value is derived in accordance with the segment range, a region of the segment in which the signal strength values exceed the threshold value is extracted, and the width of the extracted region is designated as the width of the detected object.

A safe driving support apparatus in a vehicle is used in a manner that prevents an over-reliance of a driver on a safety device by an application of a technique. That is, for example, an evaluation of driving operation is lowered when the vehicle has been under control of an avoidance control for avoiding an unsafe situation of the vehicle, and the lowered evaluation of driving operation is recorded on a memory medium as well as being notified for the driver after evaluation is being changed. A provision of the avoidance control may result in a suppression of a vehicular function that contributes to a comfort in the vehicle among functions provided by in-vehicle devices.

A method of determining parameters of a sample by X-ray scattering comprising the steps of exposing the sample to X-rays and measuring scattered X-ray intensity, generating a parameterized model of the sample which is used for numerical simulation of scattered X-ray intensity on the basis of a physical scattering theory, comparing the experimental and simulated X-ray scattering data to generate an error value and modifying the parameters of the model by means of a genetic algorithm involving an amount of individuals each with an equal number N of encoded parameters forming a generation and applying the genetic operators of “selection”, “crossover” and “mutation” used for composing successive generations of evolving individuals, is characterized in that after a given number k of successive generations the genetic operator of “mutation” is no longer applied in evolution of further generations. The inventive method improves the genetic algorithm such that it can approximate the true sample parameters faster and with a higher accuracy.

The invention discloses a fast detection method for large-scale targets on the sea surface based on the median value of energy, which mainly solves the problem that the prior art is not suitable for detecting large-scale targets on the sea surface in a high-resolution mode. The implementation steps are: 1) acquiring echo data; 2) incoherently processing the echo data along the pulse dimension to obtain an energy average matrix z; 3) performing logarithmic transformation on the matrix Z to obtain a logarithmic space matrix 4) using the matrix Find the detection statistic ξ corresponding to the unit d to be detected; 5) Use the matrix to find the detection threshold T corresponding to the unit d to be detected; 6) Compare the size of the test statistic ξ and the detection threshold T to determine whether the target exists 7) Use the detection result to estimate the length and size. The invention improves the detection performance of the radar for large targets and the estimation accuracy of the length and size, and can be used for fast detection and processing of large targets on the sea surface.

The invention discloses a six-degree-of-freedom fixed-time rendezvous docking control method for a spacecraft. Installation errors of a booster are taken into consideration. The method comprises the steps that under the situation that the installation errors of the booster existing in a service spacecraft are taken into account, a six-degree-of-freedom rendezvous docking motion model of the service spacecraft and a rolling target spacecraft is constructed, wherein the model comprises the motion descriptions of the relative position and relative attitude between the two spacecrafts; based on the established rendezvous docking motion model in which the installation errors of the booster are taken into account, and by using related motion variations, a new fixed time sliding surface is obtained; a fixed time rendezvous docking controller is obtained based on the sliding surface, and an installation error angle of the booster is estimated by constructing new adaptive rules. The method hasthe advantages of fixed time characteristics, high controlling precision, parameter overestimation preventing and the like, and the method is suitable for a six-degree-of-freedom spacecraft rendezvousdocking control mission with higher timeliness demands.

The present application discloses stability-indicating potency assays for influenza vaccines.

The invention relates to a computer image processing and visual detecting genecology. In order to provide a method which can exactly estimate the noise function of different CCD (charge coupled device) images and further can not increase the computing complexity, the technical scheme adopted by the invention is as follows: a signal-correlated noise estimating method for an image sensor comprises the following steps of: searching smooth blocks of an image by adopting a high pass operator-based image structure analyzer; sampling the blocks according to the arrangement mode of CFA (colorful filter array), to obtain a set which corresponds to each gray value, and computing the mean value and variance of the set, to obtain a set of noise estimating sample points; according the smoothness degree of the set of the image block of each sample point, computing the reliability of the sample point; and finally in combination with the sample points, the reliability of the sample points and the base obtained by using PCA (principal component analysis) in advance, rebuilding a noise horizontal curve by a sparse representation technology. The method is mainly used for processing the image.

The invention discloses a sparseness adaptive variable step-size matching tracking method based on compressed sensing. The estimated initial value is adaptively processed in the sparseness initial value estimation part, so as to avoid over-estimation and under-estimation of the sparseness initial value; In the construction part, the Dice coefficient is used to accurately select atoms and the step size is set to a fixed constant under large number of stages, so it has high reconstruction accuracy; in the initial value estimation part of the sparsity, the initial value of the sparsity is initialized, and in the reconstruction part, the initial value of the sparsity is initialized. The step size is adapted to be set, so that the large step size obtained by the exponential estimation method is used to approximate the true sparsity, and then the small step size obtained by the weak matching method is used to approximate the true sparsity. Therefore, the present invention has lower computational complexity; The invention realizes a high degree of consideration of reconstruction accuracy and computational complexity, and meets the use requirement of accurately reconstructing the original signal in a complex electromagnetic environment.

The invention discloses a deep reinforcement learning algorithm based on temporal difference, which is used to solve the mixed flow shop scheduling problem of related parallel machines. The algorithm combines convolutional neural network with TD learning in reinforcement learning, and according to the input state characteristics Behavior selection is more in line with the scheduling decision-making process of the actual order-responsive manufacturing system. By transforming the scheduling problem into a multi-stage decision-making problem, using the convolutional neural network model to fit the state-value function, inputting the processing state characteristic data of the manufacturing system into the model, using the temporal difference method to train the model, and using the heuristic algorithm or allocation rule as the scheduling decision Candidate behavior, combined with reinforcement learning reward and punishment mechanism, selects the optimal combined behavior strategy for each scheduling decision. Compared with the prior art, the algorithm proposed by the invention has the advantages of strong real-time performance and high flexibility.