The hottest big data helps the transportation indu

Big data helps the transportation industry enter the era of 4.0

recently, the 2017 Guizhou smart transportation big data innovation competition on cloud in China was closed in Guiyang. 535 projects were fiercely contested, and 16 projects finally stood on the podium. In terms of transportation, these projects involve bus route optimization, green traffic, travel behavior prediction, transportation engineering big data and other fields; From the perspective of big data, big data forces such as IOT, data mining and in-depth learning are infiltrating and applying in all fields and dimensions of transportation

it can be seen from the competition that the transportation industry has begun to use massive data to create value, which means that the transportation industry has begun to enter the 4.0 era. Pengzhongren, Professor of intelligent transportation and UAV application research center of Shanghai Jiaotong University and Changjiang Scholar, told China Science Daily

mining history, big data makes highways easier to manage

in June 2016, Yunnan Police used the operation inspection system based on highway charging big data to realize threading, cracked a major case of tampering with the entry information of pass cards to evade fees, and killed 4 card printing gangs, 1652 vehicles involved, and the amount involved was 22.5 million yuan

as early as 2015, the daily traffic flow at the entrance and exit of Yunnan Province has exceeded 1.6 million, and the traditional inspection has long been inadequate. Contestant sunxiuzhen said, so we designed a new high-speed operation inspection system based on big data. Through stream processing, Hadoop distributed parallel data processing, memory computing and other advanced big data technologies, the system integrates high-speed full-dimensional data, uses neural network technology, accurately portrays the vehicles of evasion fees, and trains and optimizes more than 30 kinds of evasion fee models

in addition to solving major cases, the competition also solved the problem of green traffic fee evasion. Green passage is a policy adopted by the state to exempt the toll of vehicles transporting agricultural and sideline products. However, some people mixed other goods in agricultural and sideline products to avoid tolls. The inspection of green traffic not only consumes a lot of human and material resources, but also reduces the traffic capacity of toll stations. If it is not inspected, it will face a massive loss of high-speed charges. Yang Ying of Guizhou Expressway group told. The integrity management platform for the aerial inspection of Expressway green channel, a competition item in this competition, used a historical backtracking method to let the toll station staff understand the operation history of green traffic in a full dimension, trying to solve this problem

green traffic drivers need to download an app with anti-counterfeiting photography technology. The rule of recording the loaded goods of vehicles as required can be used in a wide range of scenarios. The app will automatically record and identify multidimensional data such as loading time, driving path, goods category, etc; When the vehicle arrives at the toll station, the toll station staff will combine the historical portrait of the vehicle, use big data technology to check the green traffic in the air, accurately perceive the vehicle type, and quickly release the real green traffic vehicles

grasp the present, big data makes accident monitoring real-time and accurate

in March 2014, two methanol transport vehicles rear ended in Shanxi jinchengyan rear tunnel of Jinji expressway, causing a series of accidents, resulting in 40 deaths and 12 injuries

it took us 20 minutes from the accident to its discovery. If we can find the accident at the first time and accurately grasp the situation of people and vehicles in the tunnel, we may formulate a reasonable solution, and the tragedy of 2014 will not be so painful. Contestant Lu Chao said that our project insight traffic was developed after the accident. Together with the team, Lu Chao developed a set of deep learning algorithm and multi task neural network model. The system only needs 4 months of training, and it can alarm tunnel traffic accidents, traffic violations, congestion and other abnormal events at the second level, and confirm the location of abnormal events, vehicle types, tunnel pedestrians and other conditions. The recognition accuracy is as high as 99.6%

with insight into the traffic, the high-speed inspector will monitor the tunnel conditions in seconds, react in real time, and find and locate the accident at the first time. Said Lu Chao. Another participating project, huiyanda and huiyanzhi traffic, are both Expressway abnormal event monitoring platforms that use video data, but huiyanda is more secure, which is an investment, and is good at real-time monitoring of the whole road and scene

according to the introduction, huiyanda uses deep learning technology and entropy mutation model to monitor and predict the operation situation of each road, accurately identify abnormal events and give real-time alarm. Even in extreme environments and blind areas of highway cameras, real-time monitoring of road operation situation can be achieved through the target recognition technology of adaptive scene switching and the road operation room (3) the sliding surface of the inlaid steel plate in contact with the lining plate and the dovetail groove surface on the lining plate should be kept clean

predict the future, big data makes urban rail easy to dispatch

the urban rail dispatcher of Guangzhou Metro may be the happiest dispatcher in the world, because they can see the future. Using the visual screen of the multi-dimensional intelligent prediction platform of urban rail passenger flow, the urban rail dispatcher can easily see the incoming and outgoing volume, passenger flow distribution between stations, passenger flow and transfer volume of each subway exit in the next 5 minutes, 10 minutes, 1 hour, even 1 week and 1 month. Even in case of holidays, extreme weather, etc., the operation of the new metro line will not affect the accuracy at all

we can run faster than time because we integrate and analyze a large amount of historical data. Ying Ning, the contestant and the project leader of the multi-dimensional intelligent prediction platform for urban rail passenger flow, said that we have gathered the card swiping data, urban rail operation data, travel behavior data, POI data, transfer data, and test meteorological data mainly used for tensile, contraction, bending, shearing, peeling, tearing and other mechanical performance indicators of various metal, non-metal and composite materials, and established nine prediction models, It realizes the accurate prediction of urban rail passenger flow with multiple granularity, multiple scenarios and multiple indicators

in addition to the prediction platform, new prediction algorithms also frequently appeared in this competition. The research project on abnormal road flow prediction algorithm has designed a new traffic flow prediction algorithm. This algorithm is based on the characteristics that the change of abnormal traffic flow under the same attribute shows a highly repetitive trend. The road flow is divided into two parts: the benchmark reflecting the trend of the data and the deviation of the data from the benchmark sequence, which are predicted respectively based on similar patterns, and then superimposed. After verification, its accuracy is improved by 3%-5% compared with the traditional algorithm

this competition is a mirror of the development of China's intelligent transportation. In this mirror, there is the future of China's transportation. Xiexiaoyao, vice president of Guizhou Normal University, said