A method for realizing intelligent driving based on Carla simulator

The map and bicycle model are built through Carla simulator, and real-time communication is achieved by combining driving strategies and ros nodes. This solves the problem of limited simulation environment in the existing technology, improves the flexibility and debugging efficiency of the intelligent driving system, and meets high-efficiency simulation Test requirements.

CN114968782BActive Publication Date: 2024-11-08CHONGQING CHANGAN TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210576595.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-05-25
Publication Date
2024-11-08
Estimated Expiration
2042-05-25

AI Technical Summary

Technical Problem

The simulation environment and perfection built by the existing technology are relatively limited, with high requirements for operating equipment and insufficient real-time performance, making it difficult to effectively support the simulation testing of intelligent driving systems.

Method used

Carla simulator is used to build map models and bicycle models, subscribe and publish control signals through multiple node modules, and combine driving strategy startup modules to make bicycle models drive in the map, and use ros nodes for real-time communication, independently developed and debugged. Intelligent driving system.

Benefits of technology

It improves the flexibility and debugging efficiency of the system, realizes independent development and real-time communication between the simulation system and the intelligent driving system, and meets the needs of efficient simulation testing.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114968782B_ABST
    Figure CN114968782B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for realizing intelligent driving based on Carla simulator, comprising the following steps: Step 1: Based on Carla simulator, a map model and a self-vehicle model are constructed, and information sent by a Carla client is subscribed through multiple node modules, and a control signal to the self-vehicle model is issued to realize control of the self-vehicle model; Step 2: The driving strategy is imported into the Carla simulator, and multiple modules are started in sequence according to the driving strategy, so that the self-vehicle model can travel in the map model according to the driving strategy. Based on the Carla simulator, the present invention can not only isolate the simulation system from the intelligent driving system, so that the two can be developed and debugged independently, but also communicate in real time through ros nodes; for the intelligent driving system, its modules have signal dependence on each other, but can also run independently in different command windows, which undoubtedly improves the overall flexibility of the system and the efficiency of debugging programs and locating problems.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of intelligent driving technology, and in particular to a method for realizing intelligent driving based on a Carla simulator. Background Art

[0002] Self-driving cars, also known as driverless cars, computer-driven cars, or wheeled mobile robots, are intelligent cars that are unmanned by computer systems. They have been around for decades in the 20th century and are approaching practical use in the early 21st century.

[0003] Self-driving cars rely on artificial intelligence, visual computing, radar, monitoring devices and global positioning systems to work together, allowing computers to automatically and safely operate motor vehicles without any active human operation. To build a truly self-driving vehicle, the current technology research and development road is still long. When a new technology or algorithm is proposed, it is necessary to conduct simulation tests to confirm its feasibility. On the other hand, safe driving is not limited to vehicle control, but also includes understanding complex social interactions and predicting human behavior, and solving many difficult and complicated problems that have been discovered. At the hardware level, most self-driving systems rely on high-speed cameras and sensors. In addition to real-car testing of vehicle performance, the necessity and importance of simulation testing are self-evident. However, the simulation environment built by existing technologies is still limited, and the degree of perfection is also relatively limited. These not only have certain requirements on the computing speed of the computer, but also have certain requirements on the signal transmission efficiency within the simulation system. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the purpose of the present invention is to provide a method for realizing intelligent driving based on the Carla simulator, so as to solve the problems that the simulation environment constructed by the prior art and its perfection are relatively limited, the requirements for operating equipment are high, and the real-time performance is insufficient.

[0005] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0006] A method for realizing intelligent driving based on the Carla simulator comprises the following steps:

[0007] Step 1: Based on the Carla simulator, build a map model and a self-vehicle model, subscribe to the information sent by the Carla client through multiple node modules, and publish control signals for the self-vehicle model to achieve control of the self-vehicle model;

[0008] Step 2: Import the driving strategy into the Carla simulator and start multiple modules in sequence according to the driving strategy so that the ego-vehicle model can drive in the map model according to the driving strategy.

[0009] Compared with the prior art, the present invention has the following beneficial effects:

[0010] The present invention is based on the Carla simulator, which can not only isolate the simulation system from the intelligent driving system so that the two can be developed and debugged independently, but also can communicate in real time through the ROS node; for the intelligent driving system, its modules have signal dependencies on each other, but can also run independently in different command windows, which undoubtedly improves the overall flexibility of the system and the efficiency of debugging programs and locating problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 It is a schematic diagram of the process flow of the present invention. DETAILED DESCRIPTION

[0012] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present invention. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present invention.

[0013] The present invention provides a method for realizing intelligent driving based on the Carla simulator, such as Figure 1 As shown, the following steps are included:

[0014] Step 1: Based on the Carla simulator, build a map model and a self-vehicle model, subscribe to the information sent by the Carla client through multiple node modules, and publish control signals for the self-vehicle model to achieve control of the self-vehicle model. Among them, the map model and the self-vehicle model are imported into the Carla simulator through the Carla client, and the initial position and orientation of the self-vehicle model and the map model are adjusted through the python code. The map model or the self-vehicle model can also be edited through the Carla client using the python code to generate other vehicles, pedestrian environment objects, etc. The Carla simulator provides a variety of maps and vehicle models. The map comes with the entire simulation scene, but there is no initial vehicle, which can be set through the client. The topological structure of the current map can be directly obtained in the function interface library of the Carla simulator. For each lane, its road id, center reference line, left and right boundary lines, and road type can be obtained. Based on the map information, the current coordinates of the self-vehicle model and the coordinates of the target point are obtained, and the global shortest path is calculated through the AI ​​algorithm. The lane centerline trajectory of the self-vehicle model from the current position to the target position can be drawn in the current map, and the road id can be changed by color switching.

[0015] Step 2: Import the driving strategy into the Carla simulator, and start multiple modules in sequence according to the driving strategy, so that the ego vehicle model can drive in the map model according to the driving strategy. In actual operation, the entire intelligent driving simulation system can be started by starting the ros module nodes, Carla server, and Carla client in sequence, where the publishing and subscription frequency of the ros nodes is 50Hz. After receiving the corresponding control signal, the ego vehicle will drive from the initial position to the preset target point and then stop. In the Carla simulator, there are several ros nodes that subscribe to the messages sent by the Carla client at the same time, but only the lateral control and longitudinal control generate the final vehicle control signal and publish it. Among them, VehicleControl is used to control the driving state of the ego vehicle model in the next frame. VehicleControl is used to control the signals of various control components in the next frame of the ego vehicle model, including signals of vehicle control components such as throttle, steering, brake, handbrake, gear, etc. Among them, the steering is provided by the lateral control module, and the throttle and brake signals are provided by the longitudinal module. In the present invention, a variety of vehicle-mounted sensors are included, and GNSS sensors and IMU sensors are set in the vehicle model to obtain the latitude and longitude information and speed / acceleration information of the vehicle. At the same time, the driving strategy includes a neural network model or collected driving data, which makes the method of the present invention not only able to test the constructed intelligent driving neural network model, judge its operating status, and adjust the neural network model in time according to the test results; it can also be used to form multiple virtual sample data based on the collected driving data for the neural network model to be used for training, thereby expanding the application of the Carla simulator.

[0016] As described above, the reminder system of the present invention is not limited to the configuration described above, and other systems that can implement the embodiments of the present invention may fall within the scope of protection of the present invention.

[0017] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit the technical solution. Those skilled in the art should understand that those modifications or equivalent substitutions of the technical solution of the present invention that do not depart from the purpose and scope of the technical solution should be included in the scope of the claims of the present invention.

Claims

1. A method for realizing intelligent driving based on the Carla simulator, characterized in that: The steps include: Step 1: Based on the Carla simulator, build a map model and a self-vehicle model, subscribe to the information sent by the Carla client through multiple node modules, and publish control signals for the self-vehicle model to achieve control of the self-vehicle model; set GNSS sensors and IMU sensors in the self-vehicle model to obtain the latitude and longitude information and speed / acceleration information of the vehicle; import the map model and the self-vehicle model into the Carla simulator through the Carla client, and adjust the initial position and orientation of the self-vehicle model and the map model through the python code; use the python code to edit the map model or the self-vehicle model through the Carla client; Step 2: Import the driving strategy into the Carla simulator, and start multiple modules in sequence according to the driving strategy, so that the ego vehicle model can drive in the map model according to the driving strategy; Among them, after starting the ros module node, Carla server, and Carla client in sequence, the entire intelligent driving simulation system is started, and the publishing and subscription frequency of the ros node is 50Hz; after receiving the corresponding control signal, the vehicle will drive from the initial position to the preset target point and then stop; in the Carla simulator, there are several ros nodes that subscribe to the messages sent by the Carla client at the same time, but only the lateral control and longitudinal control generate the final vehicle control signal and publish it; the driving strategy includes a neural network model or collected driving data; VehicleControl is used to control the driving state of the next frame of the vehicle model; VehicleControl is used to control the signals of each control component in the next frame of the vehicle model; the steering is provided by the lateral control module, and the throttle and brake signals are provided by the longitudinal module; The topological structure of the current map can be directly obtained in the function interface library of the Carla simulator; based on the map information, the current coordinates of the vehicle model and the target point coordinates are obtained, the global shortest path is calculated through the AI ​​algorithm, and the lane centerline trajectory of the vehicle model from the current position to the target position can be drawn in the current map.