Study on Microscopic Simulation Technology of Operation Process of Unmanned Vehicle Based on Virtual Track

With the continuous development of unmanned driving technology, more and more unmanned vehicles are on the road to start testing, under this circumstance, how to control these unmanned vehicles has become an urgent issue. Based on the basic idea of virtual track and cellular automata model, this paper researches and designs a microscopic simulation system of the operation process of unmanned vehicles. The aim of this paper is to use the idea of strong controllability of railways to control unmanned vehicles in the form of train diagram to ensure their orderly safety on the road. The stability and correctness of the system are verified through case analysis. At the end of the article, the future research direction is prospected.


Introduction
The popularization of cars makes our travel more convenient and faster, but the ensuing traffic congestion and traffic accidents also cause more loss of life and property. With the progress of science and technology and the development of artificial intelligence, unmanned vehicles [1] [2] have appeared. Compared with human beings, unmanned vehicles are faster, more responsive, less affected by the environment, and they can avoid traffic accidents caused by human factors. In addition, unmanned vehicles can accurately identify traffic signals and road markings and strictly follow the preset procedures [3] . Thus unmanned vehicles make driving more standardized and traffic jams be avoided.
However, the operation of unmanned vehicles also faces some difficulties. Unmanned vehicles need to deal with the irregular driving behavior of other vehicles during the process of driving. In addition, how to obtain environmental information more accurately under the changeable and unpredictable traffic conditions [4]~ [6] also brings great challenges to unmanned vehicles.
In view of the above problems, virtual track is considered to be applied to the management of unmanned vehicles, and the idea of strong control of railway is used to carry out the control of unmanned vehicles in the way of operation diagram, so as to ensure the order and safety of unmanned vehicles more effectively. The actual road network on the ground is used as a virtual track to guide the unmanned vehicles. A number of virtual points are created with the safe distance as the interval, and the line length is the safe distance. According to OD requirements, the path is arranged for the unmanned vehicles, and then the vehicle operation is studied through microscopic simulation. Therefore, the microscopic simulation system of unmanned vehicle running process based on virtual track needs to be designed by software program.

The system structure and function division
This system uses the structural design diagram shown in Figure 1, including three parts: input module, simulation module and output module. It is not difficult to find out from the above Figure 1 that the microscopic simulation system of unmanned vehicle running process needs to have the following functions: (1) The function of reading the road network in 'shp' format, reading the composition of the vehicle, and reading the mutual attraction between the communities; (2) The function of generating OD demand between traffic zones through the gravity model; (3) The function of using reasonable traffic distribution procedures to distribute the obtained OD traffic to the network, and generating the physical path of each agent based on the distribution result; (4) The function of using a reasonable car following model to simulate the path of an individual vehicle; (5) The function of outputting the trajectory information of individual vehicles and outputting the overall statistical data of vehicles;

The design of road network format
According to KW (T) model, the G and W between two adjacent vehicles are shown in Figure 2, where G refers to the minimum space interval and W refers to the minimum time interval. decomposed. Since the length of each link calculated may not be an integer, the length of each cell is a noninteger approximate to 7m.

2.3Applied technology and programming environment
Based on the requirement of future application in largescale road network test, the language used in the design process of the system is C++ language.

2.4Algorithm design
According to the division of system mechanism and function, the algorithm flow of microscopic simulation program is designed as shown in Figure 3.
Step1: Initialize label_cost first, set the label_cost of all time and space state points to infinity, and set the predecessor of node, time, and state to -1. This item is generally done by dp, in order to avoid the influence of the other agents.
Step2: Set the label_cost of the selected starting point to 0.
Step3: For each t: Clear the SElist (scan eligible list), and put the starting point in the SElist.
Step4: For each node: Get the starting point number from the SElist.
Step5: For each m_outgoing_node: Get the number of to_node and to_link.
Step6: First judge whether cost is sufficient to meet the requirements; Second, judge whether arrival time and travel time meet the requirements; Finally, if temporary_label_cost is less than m_label_cost and the time also meets the requirements, update the label_cost of this point and record the predecessor of node, time and state. And go back to Step 5 to put the to_node into SElist.
Step7: Trace the shortest path according to label_cost of the ending point.

3.1The Basic test network
The basic test network for unmanned vehicles contains 10 nodes and 9 links (the naming rule for links is that the link between node 0 and node 1 is 1, and so on.), as shown in Figure Figure 5.

The Basic test network with merging behavior
Assume that the merging behavior is introduced in the basic test network of unmanned vehicles, as shown in Figure 6 below.

Actual road network
Malian Road is located in the west of Xuanwu District, Beijing, with West Third Ring Road in the west, Guanganmenwai Street in the north, across the street from Beijing West Railway Station, and Lize Road in the south. The road network after opening and cutting in NEXTA is shown in Figure 8.

Conclusion
In the context of the rapid development of unmanned vehicles, this paper combines the basic ideas of virtual track and railway control to conduct microscopic simulation research on the operation process of unmanned vehicles. At the same time, related functions are realized through the simulation system, and then the planning and control suggestions for unmanned vehicles are proposed.
It will be a long process for unmanned cars to replace drivers. The popularization of unmanned cars will make cities more intelligent. In the future, there is an opportunity to have a centralized scheduling system to optimize the population of the same destination and promote Shared travel.