Recent Patents on Computer Science (v.7, #2)

Distributed video coding (DVC) is considered to be a potential solution for various applications, e.g. wirelessmultimedia sensor networks, mobile devices, owing to the feature of easy encoding/complex decoding. However, practicalDVC systems still have a rate distortion (RD) performance gap with respect to the traditional video coding standardswith complex encoding/easy decoding. Side information, created at the DVC decoder as the temporal prediction for thecurrent Wyner-Ziv (WZ) frame, is of great importance for improving the RD performance of DVC. This paper firstlycompares the SI used by DVC and the traditional video coding standards respectively. The comparison reveals that the informationand the way to exploit the information used in the SI generation are the most important factors affecting thesystem's performance. According to the types of information used in SI creation, the related papers and patents areroughly classified into three classes: using the external-info (the previous decoded frames at the decoder), using the extrainfo(the information extracted at the encoder and sent to the decoder), and the main-info (the parity bits of the current WZframe and the main signal of transmission). The various methods of exploiting the three types of information to generateSI are surveyed and discussed. Finally some future potential developments are concluded.

A Transfer Optimization Method for Urban Passenger Hub based on Multiple Objectives by Wei Cheng, Yiming Bie, Zhiyuan Liu, Yan Jiang (110-118).
To improve the transfer efficiency of urban passenger hub, the platforms, stations and transfer routes were expressedas variables based on Graph theory. The calculation methods for passenger transfer time of the routes were developedbased on regression model. Conflict level of traffic flow was measured by a weighting model. A bi-level programmingmodel was brought forward, where the upper level was transfer time minimum and the lower level was traffic conflictlevel minimum. The optimization plan of transfer organization was established based on a compromise model. TakingChuangchun railway station hub as an example, the required traffic data were collected. The current transfer plan wasevaluated and existing problems were found. Then a new transfer plan was proposed based on the optimization model.The linear weighting of transfer time and conflict level was selected as the evaluation index, and the two plans were compared.The results show that the new plan can decrease total conflict by 89.4% and transfer time by 51.5%.

Research on Path Following Control Strategies for Lunar Rover under Loose Soil Terrains by Tao Zhang, Cheng Shao, Pingshu Ge, Kun W. Yang (119-127).
The path following control strategies for a six-wheel lunar rover under loose soil terrains were proposed afterbuilding the rover sideslip model, as well as the sideslip and slippage model of each wheel. The design of controller wasdone using the prior layered hierarchical architecture. The path following control level only considered the sideslip ofrover. The inner control level took the sideslip and slip of each wheel as corrections to the inverse kinematics of the rover,which is helpful to prevent the rover from losing traction performance caused by excessive wheel slippage. Several pathfollowing control simulations were conducted under different kinds of soft terrain conditions. Simulation patent resultsshow that the proposed path following control method can guarantee the traction performance and provide favorable controleffects.

Water Hazard Detection for Intelligent Vehicle Based on Vision Information by Yibing Zhao, Jining Li, Lie Guo, Yunxiang Deng, Cross Raymond (128-136).
Water hazard is one of the most dangerous obstacles for unmanned vehicle to travel in off-road environment.The vision information for water hazard body includes relatively high brightness, low saturation and smooth texture. Inthis paper, static image and sequence image are captured by CCD. Color and texture information extracted from imagepixels would be utilized to form a feature matrix to train the SVM classifier. While, with the assistance of RBF kernelfunction, low-dimensional sample space is projected to high-dimensional space. Based on large amount of experiments,RBF kernel function parameters are optimized by using grid method. The optimized RBF kernel function parameters areproved satisfactory when detecting in the static water image. Notice that there is certain relationship between target positionand its scale for the adjacent frame. SURF feature detection and matching method can be used to match feature pointbetween adjacent frames in image sequences. The searching window size and position in new frame could be updated intime, which allows to detect water hazard in a relatively small region. Water obstacle tracking experiment with patentsproved the satisfactory performance of the SURF method in this paper.

Research on Motor Vehicle Travel Trip Test for Low Carbon Transportation in ITS by Haiwei Wang, Huiying Wen, Liyou Tian, Feng You, Hailin Kui, George Georgiou (137-149).
Today more and more people have concerned about how to reduce the energy consumption of vehicles on roadin urban city. This paper uses the petrol car as research object, researching the energy saving of vehicles in conditions ofreal urban transportation. The best travel route of the energy saving is found by comparing different travel routes, and it isfinally proved that there exists a travel route that may not be the shortest one but is the one which has the least fuel consumption.Firstly, two experimental lines are selected which have the same origin and destination but have different travelroutes. By using the geographic information of Google Earth, we collected the information of latitude and longitude basedon the manually labeled, and combined MATLAB to generate the road map of city with intersection identifies. We chosetwo trial routes, and used separately fuel consumption instrument to collect fuel consumption data of vehicles by doingvehicle test on two different paths of the trial route. Secondly, test database is built and relative data is manipulated. Accordingto the changes of city's traffic flow in one day, test time is divided into four different periods: morning peak period,morning off-peak period, evening off-peak period and evening peak period, the data of every period is divided intoleaving trip and returning trip, and database is built by MATLAB. The vehicle fuel consumption is contrastively analyzedfrom different angles, such as period and direction. According to the experiment data, the average speed and percentage ofidle time were calculated, and fuel consumption model per 100 kilometers was established based on these two parameters.Results of calculation with model are in agreement with the fuel consumption of test to a certain extent, therefore, it isalso proved with patents that the fuel consumption model has high reliability.

Patent Selections: (150-154).