CN112311773B

CN112311773B - Implementation method for intelligent automobile sensor interface system

Info

Publication number: CN112311773B
Application number: CN202011099695.5A
Authority: CN
Inventors: 王亚飞; 周志松; 张鑫; 方凯; 殷承良
Original assignee: Shanghai Jiaotong University
Current assignee: Shanghai Jiaotong University
Priority date: 2020-10-15
Filing date: 2020-10-15
Publication date: 2021-07-06
Anticipated expiration: 2040-10-15
Also published as: CN112311773A

Abstract

A method for realizing an intelligent automobile sensor interface system virtualizes a hardware interface of a vehicle-mounted computing platform into a virtual interface by using interface classes in a dynamic link library, when a sensor is accessed into the hardware interface, a virtual interface class object is newly established by calling the dynamic link library, CAN messages from the sensor are verified, and the type of the sensor is identified to realize initialization; and calling the dynamic link library to newly create a corresponding sensor class object, storing the address of the corresponding virtual interface in member variables of the sensor class object, completing the matching of CAN message data and an analysis mode, and realizing the plug-and-play of the sensor through a uniform data output format.

Description

Implementation method for intelligent automobile sensor interface system

Technical Field

The invention relates to a technology in the field of intelligent automobiles, in particular to an implementation method for an intelligent automobile sensor interface system.

Background

The existing method for accessing the sensor into the intelligent automobile electronic system has certain defects. For different types or models of sensors and different manufacturers or different vehicle types, an autopilot application development engineer usually needs to consume a large amount of resources to develop and test different communication interfaces, not only needs to allocate a specified hardware interface for a certain sensor, but also needs to specially analyze and select output information of the communication matrix of the sensor, and the sensor signal intervention method generally has no universality, portability, maintainability and reliability. Therefore, a sensor interface system design method capable of conveniently and robustly connecting a sensor signal to an intelligent automobile is urgently needed in the development of the automatic driving technology.

Disclosure of Invention

Aiming at the defects in the prior art, the invention provides the implementation method for the intelligent automobile sensor interface system, and after the information is obtained through the CAN, the information of different sensors and vehicles CAN be automatically analyzed and uniformly output, so that the development of application layer software is greatly facilitated. In the aspect of CAN analysis, the packaged library CAN directly output road or obstacle information required to be used in a uniform format, and CAN detect the types and models of the sensors under the condition that different sensors simultaneously send information.

The invention is realized by the following technical scheme:

the invention relates to a method for realizing an intelligent automobile sensor interface system, which virtualizes a hardware interface of a vehicle-mounted computing platform into a virtual interface by utilizing an interface class in a dynamic link library, when a sensor is accessed into the hardware interface, a virtual interface class object is newly established by calling the dynamic link library, a CAN message from the sensor is verified, and the type of the sensor is identified to realize initialization; and calling the dynamic link library to newly create a corresponding sensor class object, storing the address of the corresponding virtual interface in member variables of the sensor class object, completing the matching of CAN message data and an analysis mode, and outputting through a uniform data format so as to realize the plug and play of the sensor.

The interface class in the dynamic link library, namely the CAN interface class, is a virtual interface directly faced by an application layer developer.

The virtual interface class object traverses the hardware interface of the computing platform, tries to establish CAN message communication, and establishes a CAN interface class for each hardware interface establishing CAN message communication.

The virtual interface comprises an uninitialized state and an initialized successful state, wherein: the uninitialized virtual interface only executes the task identified by the sensor, and discards the received CAN message after judging the received CAN message without storing the CAN message; the virtual interface that is initialized successfully will store the CAN message until it is parsed by the corresponding sensor object.

The verification specifically includes: and selecting a unique message ID as a retrieval target in different CAN communication protocols, performing traversal retrieval in a message group received by the virtual interface during initialization, and completing verification after retrieving a corresponding target.

The types of sensors described include, but are not limited to: the system comprises millimeter waves, a laser radar and a camera, wherein the identification is realized through matching of unique ID numbers contained in CAN protocols of different sensors in a CAN message, and a virtual interface completes initialization after the matching is passed.

The data format comprises: Timestamp-Timestamp, Header-id, Frameid-coordinate system, Lane-road information, type-road type, Quality-perceived Quality, Model-road Model, Viewrange-perceived length, Intercept-road Model Intercept, Curvature in Curvature-road Model, derivative of Curvature in dCurvature-road Model.

The corresponding sensor class object refers to: and for sensors with different CAN communication protocols, corresponding class objects are arranged in the dynamic link library, and message analysis, original data processing and organization and data format output are completed.

The member variables of the sensor class object comprise: virtual interface, data publisher.

Technical effects

The invention integrally solves the problems that the existing sensor signal access method has no universality, portability, maintainability and reliability, namely, the existing software developer needs to consider the hardware interface accessed by the sensor and specially analyze and select the output information of the communication matrix of the sensor, thereby influencing the independence of software algorithm development and the portability when in actual scale deployment.

Compared with the prior art, the invention automatically identifies the type of the sensor accessed by the system hardware interface through the sensor interface system and completes the analysis of the corresponding communication matrix without allocating the hardware interface of the sensor in advance and completing the corresponding software analysis work, the interface system design method has universality in different vehicles and different controllers, can be directly transplanted in different control systems through one-time development without repeated development, can effectively improve the universality, transportability and reliability of the intelligent automobile electronic system, realizes the plug-and-play target when the sensors in the intelligent automobile electronic system frame are added, greatly shortens the software development period and saves the repeated development cost of the interface function.

Drawings

FIG. 1 is a schematic diagram of the system of the present invention;

FIGS. 2(a) and (b) are a message body structure and visualization diagram of the present invention; the shaded portion in the figure is optional.

Detailed Description

As shown in fig. 1, the present embodiment relates to an interface system for an intelligent automobile sensor, which includes: interface virtualization unit, sensor identification unit and data output unit, wherein: the interface virtualization unit establishes communication towards the hardware interface, and the acquired original message data is output to the sensor identification unit. The sensor identification unit carries out retrieval and matching according to the received original message data to obtain the result of sensor identification and outputs the result to the corresponding sensor data output unit. The data output unit analyzes the message data, organizes the message data according to the data format mentioned in the invention, and outputs the message data to the application layer developer for use.

And when the sensor is accessed to any hardware interface of the vehicle-mounted computing platform, a dynamic link library is called to generate a corresponding uninitialized virtual interface for the hardware interface.

The interface virtualization unit comprises: CAN communication module and data storage module, wherein: the CAN communication module is connected with a specific hardware interface to establish communication with the sensor, and the data storage module is connected with the sensor identification unit and transmits message data.

The sensor identification unit includes: an identification module and a data storage module, wherein: the identification module receives the message data transmitted by the interface virtualization unit, processes the message data and outputs an identification result, the data storage module starts to store the message data after the identification is completed and is connected with the identification module, the data is discarded after the identification, and the storage is closed after the identification is successful; the data storage module is in butt joint with the data output unit, the sensor starts to store after being successfully identified, and the data is discarded after being analyzed and output.

The data output unit includes: analysis module and output module, wherein: the analysis module receives the output data of the sensor identification unit, is connected with the output module and transmits the sensor data analyzed according to the message protocol, and the output module organizes the sensor data according to the data format shown in fig. 2(b) and transmits the standardized sensor data for the function developer.

As shown in fig. 2(a), the connection relationship of each unit in the sensor interface system achieves the goal of conveniently and robustly connecting the sensor signal to the intelligent automobile electronic system.

As shown in fig. 2(b), it is a sensor data organization form output by the sensor access system of the present invention.

Compared with the prior art, the invention realizes the decoupling of software and hardware in the electronic system of the intelligent automobile, virtualizes the hardware interface of the vehicle-mounted computing platform, ensures that the deployment of the multiple sensors is not influenced by the positions of the interfaces, and can automatically identify the types of the sensors accessed by the interfaces and finish the analysis of a communication matrix. The invention unifies the data output formats of different sensors, and an upper application developer can start function development only by calling the standard data output in the invention without knowing the data formats of different sensors. In addition, analog data can be generated according to the data output format in the invention, and the simulation test can be carried out on the function.

The foregoing embodiments may be modified in many different ways by those skilled in the art without departing from the spirit and scope of the invention, which is defined by the appended claims and all changes that come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.

Claims

1. A method for realizing an intelligent automobile sensor interface system is characterized in that a hardware interface of a vehicle-mounted computing platform is virtualized into a virtual interface by using interface classes in a dynamic link library, when a sensor is connected into the hardware interface, a virtual interface class object is newly established by calling the dynamic link library, CAN messages from the sensor are verified, and the type of the sensor is identified to realize initialization; calling a dynamic link library to newly create a corresponding sensor class object, storing the address of a corresponding virtual interface in member variables of the sensor class object, completing the matching of CAN message data and an analysis mode, and outputting through a uniform data format so as to realize the plug-and-play of the sensor;

the interface class in the dynamic link library, namely the CAN interface class, is a virtual interface directly faced by an application layer developer;

the data format comprises: timestamp, identity, coordinate system, road information, road type, perceived quality, road model, perceived length, intercept in the road model, curvature in the road model, and derivative of curvature in the road model.

2. The method as claimed in claim 1, wherein the virtual interface object traverses the hardware interface of the computing platform, attempts to establish CAN message communication, and creates a CAN interface class for each hardware interface establishing CAN message communication.

3. The method as claimed in claim 1, wherein the virtual interface comprises an uninitialized state and an initialized successful state, and wherein: the uninitialized virtual interface only executes the task identified by the sensor, and discards the received CAN message after judging the received CAN message without storing the CAN message; the virtual interface that is initialized successfully will store the CAN message until it is parsed by the corresponding sensor object.

4. The implementation method for the intelligent automobile sensor interface system according to claim 1, wherein the verification specifically includes: and selecting a unique message ID as a retrieval target in different CAN communication protocols, performing traversal retrieval in a message group received by the virtual interface during initialization, and completing verification after retrieving a corresponding target.

5. The method as claimed in claim 1, wherein the types of the sensors include: the system comprises millimeter waves, a laser radar and a camera, wherein the identification is realized through matching of unique ID numbers contained in CAN protocols of different sensors in a CAN message, and a virtual interface completes initialization after the matching is passed.

6. The method as claimed in claim 1, wherein the corresponding sensor class objects are: for the sensors with different CAN communication protocols, corresponding class objects are arranged in the dynamic link library, and the analysis of messages, the processing and organization of original data and the output of the data format are completed; the member variables of the sensor class object comprise: virtual interface, data publisher.