Zhu Xiao

A Nonlinear Framework of Delayed Particle Smoothing Method for Vehicle Localization under Non-Gaussian Environment

In this paper, a novel nonlinear framework of smoothing method, non-Gaussian delayed particle smoother (nGDPS), is proposed, which enables vehicle state estimation (VSE) with high accuracy taking into account the non-Gaussianity of the measurement and process noises. Within the proposed method, the multivariate Student’s t-distribution is adopted in order to compute the probability distribution function (PDF) related to the process and measurement noises, which are assumed to be non-Gaussian distributed. A computation approach based on Ensemble Kalman Filter (EnKF) is designed to cope with the mean and the covariance matrix of the proposal non-Gaussian distribution. A delayed Gibbs sampling algorithm, which incorporates smoothing of the sampled trajectories over a fixed-delay, is proposed to deal with the sample degeneracy of particles. The performance is investigated based on the real-world data, which is collected by low-cost on-board vehicle sensors. The comparison study based on the real-world experiments and the statistical analysis demonstrates that the proposed nGDPS has significant improvement on the vehicle state accuracy and outperforms the existing filtering and smoothing methods.

Analytical Study on Multi-Tier 5G Heterogeneous Small Cell Networks: Coverage Performance and Energy Efficiency

In this paper, we investigate the coverage performance and energy efficiency of multi-tier heterogeneous cellular networks (HetNets) which are composed of macrocells and different types of small cells, i.e., picocells and femtocells. By virtue of stochastic geometry tools, we model the multi-tier HetNets based on a Poisson point process (PPP) and analyze the Signal to Interference Ratio (SIR) via studying the cumulative interference from pico-tier and femto-tier. We then derive the analytical expressions of coverage probabilities in order to evaluate coverage performance in different tiers and investigate how it varies with the small cells’ deployment density. By taking the fairness and user experience into consideration, we propose a disjoint channel allocation scheme and derive the system channel throughput for various tiers. Further, we formulate the energy efficiency optimization problem for multi-tier HetNets in terms of throughput performance and resource allocation fairness. To solve this problem, we devise a linear programming based approach to obtain the available area of the feasible solutions. System-level simulations demonstrate that the small cells’ deployment density has a significant effect on the coverage performance and energy efficiency. Simulation results also reveal that there exits an optimal small cell base station (SBS) density ratio between pico-tier and femto-tier which can be applied to maximize the energy efficiency and at the same time enhance the system performance. Our findings provide guidance for the design of multi-tier HetNets for improving the coverage performance as well as the energy efficiency.

A Gaussian mixture framework for incremental nonparametric regression with topology learning neural networks

Incremental learning is important for memory critical systems, especially when the growth of information technology has pushed the memory and storage costs to limits. Despite the great amount of effort researching into incremental classification paradigms and algorithms, the regression is given far less attention. In this paper, an incremental regression framework that is able to model the linear and nonlinear relationships between response variables and explanatory variables is proposed. A three layer feed-forward neural network structure is devised where the weights of the hidden layer are trained by topology learning neural networks. A Gaussian mixture weighted integrator is used to synthesize from the output of the hidden layer to give smoothed predictions. Two hidden layer parameters learning strategies whether by Growing Neural Gas (GNG) or the single layered Self-Organizing Incremental Neural Network (SOINN) are explored. The GNG strategy is more robust and flexible, and single layered SOINN strategy is less sensitive to parameter settings. Experiments are carried out on an artificial dataset and 6 UCI datasets. The artificial dataset experiments show that the proposed method is able to give predictions more smoothed than K-nearest-neighbor (KNN) and the regression tree. Comparing to the parametric method Support Vector Regression (SVR), the proposed method has significant advantage when learning on data with multi-models. Incremental methods including Passive and Aggressive regression, Online Sequential Extreme Learning Machine, Self-Organizing Maps and Incremental K-means are compared with the proposed method on the UCI datasets, and the results show that the proposed method outperforms them on most datasets.

Short-Term Traffic Flow Prediction Based on Ensemble Real-Time Sequential Extreme Learning Machine Under Non-Stationary Condition

Short-term traffic flow forecasting has been a crucial component in the area of intelligent transportation systems (ITS), which plays a significant role in operating traffic management systems and dynamic traffic assignment effectively as well as proactively. In this paper, a novel short-term traffic flow prediction method called Ensemble Real-time Sequential Extreme Learning Machine (ERS-ELM) with simplified single layer feed- forward networks (SLFN) structure under freeway peak traffic condition and non-stationary condition is proposed. By quickly training historical data and incrementally updating model with new arrived data, ERE-ELM has the characteristics of less training time consumption and high prediction accuracy. Ensemble mechanism is also used to improve stability and robustness. Experiment results show that average mean absolute percentage error (MAPE), test root mean square error (RMSE) as well as training time consumption of proposed method is superior to classical Wave-NN, MLP-NN and ELM methods.

Incremental semi-supervised kernel construction with self-organizing incremental neural network and application in intrusion detection

The semi-supervised learning (SSL) problems are often solved by graph based algorithms, semi-definite program- mings etc. These methods always require high space complexities, and thus are not efficient for network intrusion detection systems. Apart from the space complexity challenge, a network intrusion detection system should be able to handle the distribution drifting of data flow as well. A common solution for this concept drift problem is by SSL. In this paper, an incre- mental SSL training framework is proposed to combine the low space complexity advantage of topology learning and SSL for network intrusion detection. First, the unsupervised self-organizing incremental neural network is extended to process labeled and unlabeled information incrementally. Second, a kernel function is constructed from the training results of the previous step. Finally, a kernel machine is trained with the constructed kernel function. The proposed method reduces the space complexity of SSL to the magnitude similar to supervised learning. The experiments are carried out on the NSL-KDD datasets, and the results show that the proposed method outperforms the mainstream methods such as Transductive Support Vector Machine and Label Propagation.

Hybrid Cooperative Vehicle Positioning Using Distributed Randomized Sigma Point Belief Propagation on Non-Gaussian Noise Distribution

This paper proposes a cooperative positioning (CP) algorithm based on distributed modified sigma point belief propagation. Range measurement collected from the raw global navigation satellite systems (GNSSs) and ultra wide-band (UWB) data can be used for vehicle localization in urban areas under non-light-of-sight (NLOS) conditions. In order to alleviate the drawbacks of previous methods in handling non-Gaussian noise in NLOS and particularly in GNSS challenging environments, a newly range measurement error model is developed. The designed range measurement error model is based on a well-tailored non-Gaussian distribution for representing noise error in NLOS conditions. The CP problem is transformed into Bayesian inference on factor graph, which relies on a novel version of sigma point belief propagation (SPBP). The proposed CP algorithm is referred to hybrid cooperative non-Gaussian randomized sigma-point belief propagation (HC-ngR-SPBP). By using a nonlinear approximation based on randomized sigma points and stochastic integration rule (SIR), the HC-ngR-SPBP can: reduce the number of sigma point and solve a low computational cost the nonlinear measurement function exhibited by UWB and GNSS devices. The mean and the covariance calculated by the SIR are used to generate the non-Gaussian ranging probability density functions via the asymmetric generalized Gaussian mixture model. Simulation results conducted using various iterations with with real-time position information and range errors of 2 and 4 m demonstrate that the proposed algorithm outperforms the standard SPBP and nonparametric belief propagation and is very well suited for vehicle localization.

A novel hybrid approach based-SRG model for vehicle position prediction in multi-GPS outage conditions

Abstract Trajectory prediction in autonomous driving system is an important aspect for preventing for instance the multi-vehicle collision. However, predicting accurately the future location of a vehicle is still a delicate task especially in intelligent transport systems. This paper proposes a hybrid approach of solving the position prediction problem of vehicle in multi-GPS outage conditions such as free and partial as well as short and long complete GPS outages. The proposed approach aggregates the advantages of both fuzzy inference system (FIS) and sparse random Gaussian (SRG) models, consequently named FIS-SRG, leading to a significant decrease in position prediction error of vehicle. The aforementioned outages are defined by adjusting the GPS propagation weight monitored by the Gaussian model and updated by fuzzy logic system. Experimental results based on data from GPS and INS and the comparison study with the existing prediction methods illustrate the good performance of the proposed approach, in all considered GPS outage conditions.

A Load-Balancing Energy Consumption Minimization Scheme in 5G Heterogeneous Small Cell Wireless Networks Under Coverage Probability Analysis

Heterogeneous small cell networks (HSCN), as a promising paradigm to increase end-user data rates and improve the overall capacity, is expected to be a key cellular architecture in 5G wireless networks. However, energy consumed in HSCN is considerable due to the massive use of small cells. In this paper, we investigate the energy consumption issue which stems from the enormous number of running small cell base stations (SBSs) deploying in the HSCN. We first propose power consumption models so as to characterize the active state and the idle state of SBSs, respectively. Then two sleep modes for SBSs tier, i.e. random sleep mode and load-awareness dynamic sleep mode, are proposed. The random sleep is designed based on a binomial distribution of the SBS operation probability. Through the analysis on activeness of SBSs, we define the operation probability for the SBS applying the proposed dynamic sleep mode is associated to its traffic load level. The closed-form expressions of success probability for coverag...

Load-awareness energy saving strategy via success probability constraint for heterogeneous small cell networks

In this paper, we investigate the energy consumption issue which stems from the enormous number of running small cells deploying in the heterogeneous networks. We first propose two power consumption models so as to characterize the active state and the idle state of small cells respectively. Then two sleep modes for small cells tier, random sleep mode and load-awareness dynamic sleep mode, are proposed. The random sleep is designed based on a binomial distribution of the small cell operation probability. Through the analysis on activeness of small cell base stations (SBSs), we define the operation probability for the small cell applying the proposed dynamic sleep mode is associated to its traffic load level. The closed-form expressions of success probability, which is used to decide whether an active user can connect to a small cell successfully, are derived for the proposed two sleep modes. Energy consumption minimizations are presented for each of the proposed sleep modes with condition on success probability constraint. Simulation results prove the effectiveness of the proposed two sleep modes. Different energy saving gains can be achieved via using of the concrete sleep mode. The superior of the dynamic sleep mode by comparing the random sleep is also verified in terms of energy consumption, success probability and power efficiency.

A Novel Probabilistic Approach for Vehicle Position Prediction in Free, Partial, and Full GPS Outages

In this paper, a novel framework is developed with the intention of continuously predicting vehicle position even in the challenging environments such as partial and full GPS outages. To achieve this, the Bayesian-Sparse Random Gaussian Prediction (B-SRGP) approach is proposed where the sparse random Gaussian matrix which obeys the restricted isometry property with high probability is adopted to handle the measurement model. During the full GPS outages, the proposed method fuses all available INS measurements to improve the vehicle position nprediction whereas in free outages only the GPS data are processed. Besides, the Bayesian inference is used to specifically deal with the vehicle position prediction in partial GPS outages where data from both GPS and INS are taken as inputs. In all cases, measurement noises are nassumed to be non-Gaussian distributed and follow the generalized error distribution. The performance of B-SRGP is evaluated with respect to real-world data collected using Smartphone-based vehicular sensing model. The proposed method is tested when measurement noises are both Gaussian and non-Gaussian distributed and also compared with the existing prediction methods. Experimental results confirm that B-SRGP presents higher accuracy prediction and lower mean-squared prediction error for vehicle position when measurement noises are non-Gaussian distributed.