International Journal of Information Acquisition

Crowd density estimation is very important for intelligent surveillance systems in public places. This paper presents an automatic method of estimating crowd density using texture analysis and machine learning. First the crowd scene is modeled as a series of multi-resolution image cells based on perspective projection. The cell size is normalized to obtain a uniform representation of texture features. Then the feature vectors of textures are extracted from each input image cell and the support vector machine (SVM) method is utilized to solve the regression problem for calculating the crowd density. In order to diminish the instability of texture feature measurements, a technique of searching the extrema in the Harris–Laplacian space is applied. Finally, the SVM method is used again to detect some abnormal situations caused by the changes in density distribution. Experiments on real crowd videos show the effectiveness of the proposed system.

The ability of cognition and recognition for complex environment is very important for a real autonomous robot. A new scene analysis method using kernel principal component analysis (kernel-PCA) for mobile robot based on multi-sonar-ranger data fusion is put forward. The principle of classification by principal component analysis (PCA), kernel-PCA, and the BP neural network (NN) approach to extract the eigenvectors which have the largest k eigenvalues are introduced briefly. Next the details of PCA, kernel-PCA and the BP NN method applied in the corridor scene analysis and classification for the mobile robots based on sonar data are discussed and the experimental results of those methods are given. In addition, a corridor-scene-classifier based on BP NN is discussed. The experimental results using PCA, kernel-PCA and the methods based on BP neural networks (NNs) are compared and the robustness of those methods are also analyzed. Such conclusions are drawn: in corridor scene classification, the kernel-PCA method has advantage over the ordinary PCA, and the approaches based on BP NNs can also get satisfactory results. The robustness of kernel-PCA is better than that of the methods based on BP NNs.

Human gait is a dynamic biometrical feature which is complex and difficult to imitate. It is unique and more secure than static features such as passwords, fingerprints and facial features. In this paper, we present intelligent shoes for human identification based on human gait modeling and similarity evaluation with hidden Markov models (HMMs). Firstly we describe the intelligent shoe system for collecting human dynamic gait performance. Using the proposed machine learning method hidden Markov models, an individual wearer's gait model is derived and we then demonstrate the procedure for recognizing different wearers by analyzing the corresponding models. Next, we define a hidden-Markov-model-based similarity measure which allows us to evaluate resultant learning models. With the most likely performance criterion, it will help us to derive the similarity of individual behavior and its corresponding model. By utilizing human gait modeling and similarity evaluation based on hidden Markov models, the proposed method has produced satisfactory results for human identification during testing.

Event detection is important in many sensor network applications. Traffic in these applications exhibits a many-to-one pattern, in which multiple source nodes send sensing data to a single sink node. Since bandwidth and storage capacity is highly constrained in sensor networks, packet loss is inevitable when a great deal of traffic rushes to sink. Based on the observation that the packets are not equal in the serving of event detection, we propose to employ priority-based queue management method (PBQM) to improve packet delivery ratio of more important packets. We put forward three priority mapping schemes and evaluate the performance of PBQM method by simulation with Ns-2.

An intelligent soft measurement and information processing method for predicting parameters of process control system was proposed. Process neural network (PNN) is a new configuration of artificial neural network put forward in recent years. Some algorithms of PNN were discussed, and these algorithms were based on function orthogonal basis expansion, yet the convergence rate was comparatively low. An improved algorithm for raising training speed based on function orthogonal basis expansion in PNN for soft measurement was researched. After increasing the normalizing rule on the original algorithm, and introducing function momentum adjustment item and learning rate automatically adjustment method for network weight function, the training time of learning algorithm for PNN was reduced, and a good result was represented by simulation in wastewater treatment system.

Multiple moving targets detection is one of the fundamental problems in information acquisition. In this paper, the use of a transformable period and symmetrical linear frequency modulated (TPS-LFM) waveform for microwave surveillance sensor multiple moving targets identification, is proposed. In order to accurately estimate target's true position and velocity, a relatively unknown yet powerful technique, the so-called fractional Fourier transform (FrFT), is applied to estimate the moving target parameters. By mapping a target's signal onto a fractional Fourier axis, the FrFT permits a constant-velocity target to be focused in the fractional Fourier domain thereby affording orders of magnitude improvement in signal-clutter-ratio. Moving target velocity and position parameters are derived and expressed in terms of an optimum fractional angle and a measured fractional Fourier position, allowing a target to be accurately located. Moreover, to resolve the problem whereby weak targets are covered by the sidelobes of strong ones, the CLEAN technique is also applied. Simulation results show that the method is effective in estimating target velocity and position parameters for microwave surveillance sensors.

Iris recognition technology relates to computer vision, pattern recognition, statistical inference, and optics etc. Because the randomness of iris patterns has very high dimensionality, it is difficult to find an efficient approach to iris recognition. The intersecting cortical model (ICM) is good at directly extracting important information from image. A new method of iris recognition is proposed for the first time based on the ICM. With the method, a series of binary images are firstly produced from iris image through the ICM. Entropy sequence can be gained from these binary images. And then phase information is obtained from entropy sequence. This phase information is taken as feature vector because of its uniqueness. The results show that our method is feasible, potential and effective in obtaining feature from iris image.

Global Positioning System (GPS) is often used as a main information source for robot localization and navigation. However, it cannot be used in room or in field complex environment because of the bad signal there. To solve this problem, the authors designed and implemented a specific wireless sensor network (WSN) to provide information about the environment and indicate path for robot navigation. A two-stage auto-adaptive route selecting mechanism of the WSN was proposed to facilitate data relaying in localization and the robot's navigation. A low complexity localization algorithm was used to localize both the nodes and the robot. An indirect communication method was designed to make the communication between the WSN and the robot possible. In addition, a robot navigation method was proposed based on the wireless sensor network. In this method, the robot did not need to obtain the environment information; the wireless sensor nodes collected and fused the distributed information and then indicate a path for the robot. Experiments showed that the wireless sensor network can result in obstacle avoidance navigation, and can implement the online navigation.