For Realizing an Intelligent Transport Infrastructure Utilizing 5G Ultra-low Latency

Highlights

Real-time recognition of roadway situations by using wireless communications assumed as the 5th generation (5G) ultra-low latency communications.
Acquisition of location, speed and so on for autonomous mobilities such as vehicles by using “Wireless Electronic Traffic Mirrors (WETMs)” built-in sensors
Expectation of an intelligent transport infrastructure to avoid mobility collisions and to predict movement.

Summary

NICT, Wireless Networks Research Center has built the test-bed intelligent transport infrastructure in the Yokosuka Research Park (YRP) utilizing the 5th generation wireless communication system (5G) ultra-low latency, which is purposed to support autonomous mobility in the intersection of roadway environment.

In addition, we have developed a wireless electronic traffic mirror (WETM) built-in camera and location measuring sensors that enables to collect the various situations on the roadway such as road construction, vehicle congestion, etc. Such roadway information collected by the multiple WETMs will be reflected in the dynamic map (DM), which is a database maintaining the roadway information, and will be transmitted to the autonomously moving vehicles. Under the test-bed infrastructure, we have confirmed that it is possible to grasp the roadway situations in real time, especially an area in the vicinity of an intersection with poor visibility. From the achievements, we expect to construct an intelligent transport infrastructure to avoid mobility collisions and to predict movement.

Background

In the near future, it is expected that the diverse, autonomous mobilities of vehicles, drones, tractor, etc., will be popularized, and a highly reliable intelligent transport infrastructure will be required to support the safe autonomous movement. In addition, various mobilities may autonomously move on the various road situations with the combinations of congestion, construction, collision and so on. In order to realize the autonomous transport system, it is necessary to grasp the roadway situations accurately in real time. For this reason, it is required for the establishment of a roadway acquisition technology that a large number of sensors with wireless communications collect the road information and a dynamic map maintains the various road information. To verify the roadway acquisition technology, it is necessary to develop the test-bed infrastructure of acquiring the roadway situations by using infrastructure sensors, and to measure quantitatively the performance of wireless communications on a real roadway.

Achievements

NICT has developed a test-bed infrastructure of acquiring the roadway situations by using the infrastructure sensors, namely as WETMs, which simulates autonomous mobility in the intersection of roadway environment in the Yokosuka Research Park (YRP). We have confirmed the following contents.

● The developed WETM built-in a stereo camera and an LRF (laser range finder) recognizes the position, speed, mobility type, etc. of moving objects and obstacles in real time.

● After reducing the amount of sensing information such the techniques as image compression and clipping, the sensing information is transmitted to an edge server by using a wireless system assumed as 5G.

● On the assumption that the sensing information is reflected in the dynamic map, the edge server recognizes the change of the road environment by extracting the features like position, speed, type, etc. obtained the sensing information from the multiple sensors

● In order to absorb the difference in transmission time caused by wireless communications, the recognized sensing information transmitted by multiple sensors must be synchronized by time stamp. In the edge server, the sensor information is integrated, and a snapshot of the road environment is generated.

Future Prospects

In the future, we will conduct the 5G wireless communication system on the test-bed infrastructure, and plan to evaluate the performance with various wireless systems in case the various roadway conditions such that a large number of moving objects are located on a roadway and the moving speed of each is different or the area is wide. We will confirm the functional requirements to be aimed at establishing the technology to realize a more advanced autonomous transport system.

Appendix

Figure 1 shows an overview of the intelligent transportation infrastructure for real-time recognition of roadway situations by using the developed test-bed infrastructure. We will place the multiple WETMs in a location such as the shade of the building or the intersection. Each WETM is equipped with a stereo camera and LRF (hereinafter referred to as a sensor) and recognizes the position, speed, mobility type, etc. of the moving objects such that are vehicles and obstacles. Both the camera image and the recognized object information are collected to the edge server over the wireless communications. Generally, multiple WETMs are installed on a roadway, and they transmit the road information by wireless communications to the edge server. However, the communication errors will occur between the time recognized the road situation and the time collected the road situation at the edge server. In the edge server, therefore, sensor information at a certain time is integrated based on the time stamp to synchronize the collected information time, and a snapshot of the road situations at the synchronized time is generated. This information is reflected in the dynamic map and is expected to be finally distributed to the autonomous moving users.

In addition, although IEEE 802.11ac wireless systems are used to simulate, it is also designed to be able to operate in place of other wireless systems such as 5G and LTE. We are planning to confirm its performance using various wireless systems.

Figure 1: An overview of intelligent transport infrastructure

Figure 2 shows the appearance of the developed WETM. The WETM contains a stereo camera and LRF to detect the moving objects on the roadway. Also, the detected moving objects are processed at the object recognition processing device to classify the object type, position, speed, and so on. Finally, the processed information is transmitted to the edge server on the wireless communication system.

Figure 2: An appearance of the developed Wireless Electronic Traffic Mirror (WETM)

Figure 3 shows the results of the field test conducted in YRP. The above figure of Figure 3 shows the installation position of the WETMs in YRP and their appearances. The lower part of Figure 3 shows the screen shot when a car is traveling from the left side of the T-junction and turns right and moves down. We integrate the images transmitted from the WETMs at the edge server and grasp the road situation around the intersection. From the field test by using the developed test-bed infrastructure, we have confirmed that the actual positions of the moving objects are grasped in real time. Moreover, by extracting only image information being changed from the image information acquired by the WETMs, it is possible to reduce the amount of transmitting data.