Tokyo Institute of Technology is a top university in Japan and a world-class university specializing in engineering technology and natural science. It is one of the most important top universities in Japan's super International University program and an important member of the association of East Asian research universities. In the 2021 QS World University ranking, it ranks 56th and 14th in engineering technology.
Tokyo Institute of Technology cooperates with a local parking company to study and promote the intelligent parking project as a key industry university research one. Based on the parking lot, they are planning to update the old equipment and upgrading the current parking lot management.
At present, the management system used in the parking industry in Japan is still based on the traditional vehicle detection with ground sensing coil, and there is still a large space for technology promotion. In recent years, due to the rapid development in Artificial Intelligence field, edge computing technology which includes image recognition through convolution neural network has been widely used in various industries, and parking lot management system is gradually developing towards camera vehicle detection, near-end high-speed data calculation and so on.
1. High cost and low efficiency
At present, how to detect the entering and leaving of the vehicles mainly depends on the traditional detector with ground sensor coil, and the parking fees can only be paid in cash. Due to the high construction cost brought by the complexity of the ground sensing coil, the operation and maintenance management cost of regular detection equipment aging state and the resulting labor cost, most of the current parking lot management faces the bottleneck of high cost and low efficiency.
2. Limited deployment space
Because the space of the parking lot, where the clients need to deploy the servers, is relatively limited for background equipment when maintaining and managing the parking lot, it brings challenges to the placement of a series of equipment, including computers.
3. High computing performance
If the more advanced convolution neural network is used to recognize the vehicles to judge whether there are vehicles entering and leaving the warehouse, process real-time image for license plate recognition, and to complete the charging of the car owner at the same time, this series of operations need fast transmission and accurate calculation of the data at the near end, which requires a lot the efficiency of sharing and processing the front and rear data.
In order to help to cope with these challenges, IEIT proposed EIS series edge intelligent computing station, mainly promoting EIS200 for the parking lot scene. Firstly, compared with the universal server, the edge server has smaller volume and is easy to store, which means that it can be placed in a limited space and has strong environmental adaptability. Besides, it can work normally outside without temperature control. Secondly, both the super image analysis and AI computing capabilities enable it to make rapid reasoning and decision-making at the near end to manage single parking lot independently, saving the time of data remote transmission and processing. In this way, the background data center only needs to manage EIS200, which greatly reduces the pressure of background computing, improves the daily management efficiency of a single parking lot, and promotes the overall management of different parking lots.
1. Effective vehicle management
Unlike the traditional ground sense coil, the solution directly grabs the license plate number and other information through the installed camera to generate the image data. Then it will be transmitted to EIS200 immediately. Through matching the vehicle information in database, it can quickly match the name of the owner, and then associate it with its bank account to deduct the parking fee. The whole computing process relies on edge server EIS200 analyzing the data at the near end where image data is generated. Besides, whether parking cost is correct is based on the accurate calculation of time. After the camera takes the picture of entering and exiting the parking lot, EIS200 can calculate the parking time immediately, and associate it with the relevant data of the bank account. The use of camera and near-end edge server can not only facilitate the installation of ground coil, but also simplify the management and maintenance, finally, the solution greatly shortens the time of vehicle identification and electronic withholding, and improves the accuracy.
2. Vehicle identification and parking management
Car safety is also an important part of parking lot management. With the help of cameras, the EIS200 can "record" the site conditions of the parking lot. If a dangerous event or suspicious person is found, an alarm alert will be sent automatically. EIS200 can also monitor the parking environment. For example, when there is a fire in the parking lot, the dangerous situation can be found and danger warning can be timely detected through the video transmitted by the camera and EIS200 in real time; moreover, the overall use of the parking lot can be evaluated to see whether there are lane occupation and retrograde to enhance the availability of internal space. In this way, with the efficiency of parking space be improved, the congestion outside the parking lot in peak period also be alleviated.
Through the solution of IEIT edge intelligent computing station and high-quality server products, IEIT guarantees the computing resources for the project saving more than 40% of the construction cost of the parking lot, which greatly reduces the operation and maintenance and management pressure and realizes the convenient management of multiple parking lots all over Japan, therefore, it is not necessary to carry out the regular aging test of the ground induction coil. IEIT's excellent solution of edge intelligent computing station also helps the parking company successfully improve customer satisfaction and loyalty.