Bao Rong Chang

Forecast approach using neural network adaptation to support vector regression grey model and generalized auto-regressive conditional heteroscedasticity

In order to reduce the volatility clustering effect that deteriorate the efficiency and effectiveness of time series prediction and gives rise to large residual errors, a composite method, which is SVRGM/GARCH model with neural network adaptation, is introduced to improve the predictive accuracy of the complex time series, e.g. stocks price index or futures trading index. A support vector regression (SVR) is employed to improve the control and environment parameters of grey model (GM) denoted by SVRGM. Thus, SVR learning functions to highly reduce the overshoot effect when GM is applied to time series prediction. Moreover, a generalized auto-regressive conditional heteroscedasticity (GARCH) is utilized to resolve the problem of volatility clustering in time series so as to best fit the model. Incorporating GARCH model into SVRGM prediction is scheming to effectively and efficiently tackle two crucial problems, the overshoot and volatility clustering effects, simultaneously. This composite system is adapted optimally by back-propagation neural network (BPNN).

Intelligent data fusion system for predicting vehicle collision warning using vision/GPS sensing

In this study, fuzzy approach with fault-tolerance has proposed to fuse heterogeneous sensed data and overcome the problem of imprecise collision warning due to perturbed input signal when processing the pre-crash warning. Meanwhile, another problem relevant to the danger in drowsy driving, involving fatigue level, carbon monoxide concentration, and breath alcohol concentration, was considered and has approximately reasoned to an extra reaction time to modify NHTSA algorithm. A vision-sensing analysis cooperating with global-positioning system is applied for lane marking detection and collision warning, particularly exchanging the dynamic and static information between neighboring cars via inter-vehicle wireless communications. In addition to pre-crash warning, event data recording very useful for accident reconstruction on scene is also established here. In order to speed up data fusion on both quantum-tuned back-propagation neural network (QT-BPNN) and adaptive network-based fuzzy inference system (ANFIS), a distributed dual-platform DaVinci+XScale_NAV270 has been employed. Several tests on systems reliability and validity have been done successfully, and the comparison of system effectiveness showed that our proposed approach outperforms two current well-known collision-warning systems (AWS-Mobileye and ACWS-Delphi).

Cooperative Collision Warning Based Highway Vehicle Accident Reconstruction

Highway safety is always an important issue for automobile industry, so many researches have been conducted to prevent from or reduce the accidents. Cooperative Collision Warning (CCW), which provides an active safety mechanism for vehicles on highways, is implemented by exchanging static and dynamic vehicle parameters with neighboring vehicles through inter-vehicle wireless communications. Received information is not only used for calculating the relative safety distance between neighboring vehicles, but also stored in a Motor Vehicle Event Data Recorder (MVEDR) for future accident investigation. A CCW-based MVEDR can easily rebuild the trajectory and interaction between the host and neighboring vehicles after an accident. This device saves time, labor and cost required for accident analysis and reconstruction.

A highway traffic simulator with Dedicated Short Range Communications based cooperative collision prediction and warning mechanism

Highway safety is always a critical issue. A variety of sensors are employed in a vehicle for reducing traffic accidents, but it is useful only within the sensor range of this single vehicle. The earlier the driver notices the dangerous situation, the longer the response time is available to the driver. Dedicated Short Range Communications (DSRC) is a good medium for inter-vehicle communications to periodically exchange static and dynamic parameters, and for each vehicle to maintain a map of its surrounding vehicles. The cooperative collision prediction and warning mechanism (CCPWM) will inform the driver of any potential risk to prevent accidents. A Highway Traffic Simulator (HiTSim) with DSRC-based CCPWM equipped on each vehicle was developed for simulating highway traffic and evaluating the probability of accidents. The results demonstrate that the vehicle with DSRC-based CCPWM will increase the time available to the driver to respond and therefore the vehicular safety.

Improving network traffic analysis by foreseeing data-packet-flow with hybrid fuzzy-based model prediction

Forecast of the flow of data packets on a computer network gives valuable information about the change of data-packet-flow to the website at the next upcoming period, which is a way to enhance the capability of network traffic analysis. Thousands of web-smart businesses depend on network traffic analysis to improve network conversions, reduce marketing costs, facilitate network optimization, speed-up network monitoring and provide a higher level of service to their customers and partners. In this study, an intelligent-based hybrid model prediction is introduced for foreseeing data-packet-flow on a network. This is to combine adaptive neuro-fuzzy inference system (ANFIS) with nonlinear generalized autoregressive conditional heteroscedasticity (NGARCH), tuned optimally by adaptive support vector regression (ASVR). The hybrid model is chosen for resolving the problems of the overshoot and volatility clustering simultaneously so as to improve the predictive accuracy and we denote it as ASVR-ANFIS/NGARCH in this paper. Once we start on the scheme of foreseeing data-packet-flow on a network, the throughput ratio of foreseeing and non-foreseeing data-packet-flow is increased roughly up to 20%. We thereby drew the conclusion that the proposed scheme above can aid webmaster to improve network bandwidth allocation effectively and efficiently and then help web analytics to optimize their website, maximize online marketing conversions, and lead campaign tracking.

Highway vehicle accident reconstruction using Cooperative Collision Warning based Motor Vehicle Event Data Recorder

There is often uncertainty or negligence in measurements and calculations by collecting relevant evidence and site information at the scene for the conventional vehicle accident analysis and reconstruction. Cooperative Collision Warning (CCW) mechanism exchanging static and dynamic vehicle information with neighboring vehicles through inter-vehicle wireless communications provides an active safety mechanism for vehicles on highways. Received information is not only used for calculating the relative safety distance between neighboring vehicles, but also preserved in a Motor Vehicle Event Data Recorder (MVEDR) for future accident reconstruction. A CCW-based MVEDR can easily rebuild the trajectory and interaction between the host and neighboring vehicles within communication range. This device is a supplementary tool for conventional accident reconstruction, and it saves time, labor and cost required for investigations, and eliminates uncertainty regarding accident analysis.

Simulation and Implementation of High-Performance Collision Warning System for Motor Vehicle Safety Using Embedded ANFIS Prediction

A good driver-vehicle interface needs two functions: (a) how to effectively prevent traffic crash and (b) how to record event data precisely for facilitating the traffic crash investigation after accident. This study is to explore how to realize high-performance collision warning system (CWS), providing the precaution against traffic crash in transit. An embedded adaptive network-based fuzzy inference system (ANFIS) built in XScale-NAV270 platform was employed to realize collision warning system and we also installed motor vehicle event data recorder (MVEDR). Finally, simulation and verification of the proposed approach were successfully done to achieve better accuracy and more effectiveness on warning operation and event data record to motor vehicle.

Embedded System for Inter-Vehicle Heterogeneous Wireless-Based Real-Time Multimedia Streaming and Video/Voice over IP

Inter-platform streaming multimedia between embedded systems has encountered the problem of head-of-line blocking and non-real-time transmission/receiving. When a mobile device in an embedded system moves between subnets, the node will face the handover problem. In this paper, we utilize Stream Control Transmission Protocol (SCTP) features to develop the multimedia communications including streaming multimedia and video/voice over IP based on Linux built in an embedded platform. In SCTP, the multi-homing feature enables the mobile node to handover seamlessly between networks. Also the multi-streaming enables data to be sent in multiple, independent streams in parallel, so that data loss in one stream does not affect the delivery of data in other streams to eliminate head-of-line blocking. Another feature Partially Reliable SCTP (PR-SCTP) is used to achieve the realtime video/audio services.

ID-Based PPM for IP Traceback

A modified scheme in probabilistic packet marking (PPM) for IP traceback against distributed denial-of-service attack is presented. This method, ID-based PPM, is proposed to improve the original PPMs complexity of fragments combination through clustering these fragments in advance. Furthermore, by reducing the time of fragments combination, the attacking path reconstruction and attack response time can also be speeded up

A Fast and Effective Approach to Lane Marking and Neighboring Vehicles Detections Based on Vision/GPS Sensing Together with Vehicle-to-Vehicle Communication

A fast warning response to an impending crash based on vision/GPS sensing together with v2v communication has been realized to achieve line marking recognition, neighboring vehicles detection, and headway estimation, where a distributed embedded dual-platform DaVinci+XScale-NAV270 was built to improve system effectiveness as well as a data fusion QT-BPNN/ANFIS was applied to dissolving the heterogeneous data. GPS information for each car was exchanged and disseminated each other via dedicated short range communication (DSRC). It performs at the least response time 0.5972 seconds and gets reduced accident rate about 70~82%, superior to two alternative collision warning systems.