CN219172410U

CN219172410U - Vehicle-mounted intelligent computing device, control system and intelligent driving vehicle

Info

Publication number: CN219172410U
Application number: CN202223140749.5U
Authority: CN
Inventors: 马琎劼; 许洪华; 周显威; 王德顺; 耿瑶; 杨楠; 曹远志; 王开毅; 张鑫熔; 薛金花
Original assignee: China Online Shanghai Energy Internet Research Institute Co ltd; Nanjing Branch Of China Electric Power Research Institute Co ltd; State Grid Jiangsu Comprehensive Energy Service Co ltd; State Grid Jiangsu Electric Power Co ltd Innovation And Innovation Center; State Grid Corp of China SGCC; State Grid Jiangsu Electric Power Co Ltd; Nanjing Power Supply Co of State Grid Jiangsu Electric Power Co Ltd
Current assignee: China Online Shanghai Energy Internet Research Institute Co ltd; Nanjing Branch Of China Electric Power Research Institute Co ltd; State Grid Jiangsu Comprehensive Energy Service Co ltd; State Grid Jiangsu Electric Power Co ltd Innovation And Innovation Center; State Grid Corp of China SGCC; State Grid Jiangsu Electric Power Co Ltd; Nanjing Power Supply Co of State Grid Jiangsu Electric Power Co Ltd
Priority date: 2022-11-25
Filing date: 2022-11-25
Publication date: 2023-06-13
Anticipated expiration: 2032-11-25

Abstract

The embodiment of the utility model provides a vehicle-mounted intelligent computing device, a control system and an intelligent driving vehicle, wherein the vehicle-mounted intelligent computing device comprises: the system comprises an external interface module, an AI calculation module, a fusion calculation module and a regulation positioning module; the external interface module comprises a plurality of interfaces; the fusion calculation module is connected with the vehicle-mounted sensor equipment through a first interface; the gauge control positioning module is connected with the vehicle-mounted positioning equipment and the vehicle control unit through a second interface; the fusion calculation module is respectively connected with the AI calculation module and the regulation positioning module. The intelligent driving calculation efficiency can be improved, and the hardware cost is reduced.

Description

Vehicle-mounted intelligent computing device, control system and intelligent driving vehicle

Technical Field

The utility model relates to the technical field of intelligent driving, in particular to vehicle-mounted intelligent computing equipment, a control system and an intelligent driving vehicle.

Background

Electric locomotives traveling on the basis of rails are important carriers for industrial and mining enterprises to transport goods in a park. After the traditional locomotive is electrically modified, the whole process of zero pollution of the operation of the locomotive in industrial and mining enterprises can be realized, and the novel rail electric locomotive is subjected to system upgrading and optimization in aspects of a battery system, whole vehicle control, a motor driving system and the like. The electric locomotive needs to observe the running environment in a manual mode and performs manual driving control according to the environment condition and the running task, and the problems of manual misjudgment, high running cost of the vehicle driving, low running efficiency and the like exist.

In order to solve the above problems, the intelligent driving technology of the electric locomotive is one of the research hotspots in the industry. Because the intelligent driving technology needs to process mass data and perform complex logic operation, the intelligent driving technology needs to be supported with great effort. The vehicle-mounted intelligent computing equipment needs to process all sensor data simultaneously, decision making and execution are carried out based on the processing result, the requirement on the chip computing power is high in the mode, and the vehicle-mounted intelligent computing equipment needs to carry out excellent structural design and functional design.

Therefore, how to provide a vehicle-mounted intelligent computing device, a control system and an intelligent driving vehicle, so as to improve the intelligent driving computing efficiency and reduce the hardware cost, becomes a problem to be solved urgently.

Disclosure of Invention

Aiming at the defects in the prior art, the embodiment provides a vehicle-mounted intelligent computing device, a control system and an intelligent driving vehicle.

The utility model provides a vehicle-mounted intelligent computing device, which comprises: the system comprises an external interface module, an AI calculation module, a fusion calculation module and a regulation positioning module;

the external interface module comprises a plurality of interfaces;

the fusion calculation module is connected with the vehicle-mounted sensor equipment through a first interface;

the gauge control positioning module is connected with the vehicle-mounted positioning equipment and the vehicle control unit through a second interface;

the fusion calculation module is respectively connected with the AI calculation module and the regulation positioning module.

Optionally, the vehicle-mounted intelligent computing device provided by the utility model further includes: a redundancy calculation module;

the redundancy calculation module is respectively connected with the AI calculation module and the regulation positioning module.

Optionally, the vehicle-mounted intelligent computing device provided by the utility model further includes: a monitoring management module;

the monitoring management module is respectively connected with the external interface module, the AI calculation module, the fusion calculation module and the regulation positioning module.

Optionally, the vehicle-mounted intelligent computing device provided by the utility model, the external interface module includes: at least one of LVDS interface, ETH interface, CAN bus interface, UART interface, DI interface and AI interface.

Optionally, in the vehicle-mounted intelligent computing device provided by the utility model, the AI computing module adopts an AI chip;

the fusion calculation module and the regulation positioning module adopt RK3588 chips.

Optionally, the vehicle-mounted intelligent computing device provided by the utility model has the advantage that the monitoring management module adopts a TC397 chip.

The utility model also provides a vehicle-mounted intelligent control system which comprises the vehicle-mounted intelligent computing equipment, the vehicle-mounted sensor equipment, the vehicle-mounted positioning equipment and the vehicle control unit.

Optionally, the vehicle-mounted intelligent control system provided by the utility model, the vehicle-mounted sensor device comprises: at least one of a camera, a laser radar sensor and a millimeter wave radar sensor.

Optionally, the vehicle-mounted intelligent control system provided by the utility model has the advantage that the vehicle-mounted positioning equipment adopts GNSS/IMU integrated navigation equipment.

The utility model also provides an intelligent driving vehicle, and the intelligent driving vehicle is loaded with the vehicle-mounted intelligent control system.

According to the vehicle-mounted intelligent computing device, the control system and the intelligent driving vehicle, the three computing modules with different functions, namely the AI computing module, the fusion computing module and the regulation positioning module, are designed in the vehicle-mounted intelligent computing device, the data sources are controlled through the external interface module, the data required to be computed in intelligent driving are respectively delivered to the different computing modules for processing, the demand on computing power of a single module is effectively reduced, the hardware cost is reduced, and the mode that the whole computing task is divided into a plurality of subtasks and simultaneously processed by the different computing modules is adopted, so that the intelligent driving computing efficiency can be effectively improved.

Drawings

In order to more clearly illustrate the embodiments of the present utility model or the technical solutions of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are some embodiments of the present utility model, and other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic structural diagram of a vehicle-mounted intelligent computing device according to an embodiment of the present utility model;

FIG. 2 is a diagram of an overall design architecture of an intelligent computing unit according to an embodiment of the present utility model;

FIG. 3 is a schematic diagram of the core structure of an AI chip according to an embodiment of the utility model;

FIG. 4 is a diagram of an overall design architecture of a fusion computing module according to an embodiment of the present utility model;

FIG. 5 is a diagram of an overall design architecture of a gauge control positioning module according to an embodiment of the present utility model;

FIG. 6 is a diagram of an overall design architecture of a redundant computing module according to an embodiment of the present utility model;

FIG. 7 is a diagram of an overall design architecture of a monitoring management module according to an embodiment of the present utility model;

FIG. 8 is a schematic diagram of a communication architecture of a vehicle-mounted intelligent computing device according to an embodiment of the present utility model;

FIG. 9 is a schematic diagram of intelligent driving data flow of an electric locomotive according to an embodiment of the present utility model;

fig. 10 is a schematic diagram of a multi-sensor fusion positioning architecture according to an embodiment of the present utility model.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the embodiments of the present utility model more apparent, the technical solutions of the embodiments of the present utility model will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present utility model, and it is apparent that the described embodiments are some embodiments of the present utility model, but not all embodiments of the present utility model. All other embodiments, which can be made by those skilled in the art based on the embodiments of the utility model without making any inventive effort, are intended to be within the scope of the utility model.

Fig. 1 is a schematic structural diagram of a vehicle-mounted intelligent computing device according to an embodiment of the present utility model, where, as shown in fig. 1, the present utility model provides a vehicle-mounted intelligent computing device, including: the system comprises an external interface module, an AI calculation module, a fusion calculation module and a regulation positioning module;

the external interface module comprises a plurality of interfaces;

Specifically, the vehicle-mounted intelligent computing device provided by the embodiment of the utility model can be divided into an external interface module, an AI computing module, a fusion computing module and a regulation positioning module according to a hardware composition structure.

The external interface module comprises a plurality of interfaces; can be connected with the vehicle-mounted equipment according to different buses. It will be appreciated that, in order to implement intelligent driving, the interface needs to connect the vehicle-mounted sensor device with the vehicle-mounted positioning device and the vehicle control unit, and the specific device type and the bus type required to be used by the interface may be set according to actual requirements, which is not limited by the present utility model.

And the external interface module transmits the data acquired by the vehicle-mounted sensor to the fusion calculation module through the first interface. The fusion calculation module comprises middleware and a driver for processing the vehicle-mounted sensor data, and after the vehicle-mounted sensor data are acquired, preprocessing, sensing, fusion, prediction and other processes (such as camera preprocessing and laser radar preprocessing) can be performed on the data through a preset operation sensing algorithm.

The fusion calculation module is connected with the AI calculation module and can directly schedule the AI calculation module to perform AI calculation. The AI calculation module comprises an AI chip which is used for carrying out AI related calculation and can run a target detection algorithm based on video analysis to detect and identify target objects of the rail, the front tail, the obstacle and the pedestrian of the electric locomotive.

The external interface module transmits the data acquired by the vehicle-mounted positioning device to the regulation positioning module through the second interface. The regulation and control positioning module can run a preset regulation and control algorithm and a preset positioning algorithm after acquiring data acquired by the vehicle-mounted positioning device. In addition, the regulation positioning module is connected with the fusion calculation module and can be used as a time synchronization master to realize synchronization of different sensor data (for example, to execute camera and laser radar frame synchronization).

It should be noted that, the "first" and "second" in the first interface module and the second interface module are merely used as distinction between the text portions and the interfaces, and do not include actual meaning.

It can be understood that the utility model only limits the specific connection modes of the external interface module, the fusion calculation module, the AI calculation module and the regulation positioning module, so that the intelligent driving related development can be realized more conveniently by technicians under the hardware configuration.

In addition, the vehicle control unit is a unit for executing intelligent decision control (such as starting, accelerating, braking, parking avoidance, steering, light conversion and the like) in the vehicle. In practical application of the present utility model, the specific equipment type and control method included in the vehicle control unit may be set according to the type of the vehicle and the actual requirement, which is not limited in the present utility model.

For example, after the data collected by the vehicle-mounted sensor is obtained according to the external interface module, the intelligent decision is determined to be that the left turn lamp is turned on and the left turn is performed according to the regulation and positioning module, the fusion calculation module and the AI calculation module, at the moment, the vehicle control unit activates the left turn lamp and enables the steering gear to work, so that the front wheels of the vehicle generate angles to achieve the effect of turning left.

The specific implementation steps of the method mentioned in the embodiments are only used as a specific example to assist in describing the practical application of the present utility model, the present utility model only provides reference, and the specific algorithm and implemented functions adopted by the modules in the practical application may be set according to the actual requirements, and in addition, the chip selection type used by the modules may be selected according to the actual requirements, which is not limited by the present utility model.

Taking a rail electric locomotive as an example, the vehicle-mounted computing equipment provided by the utility model can realize intelligent perception, intelligent positioning, route navigation planning and intelligent decision control (such as starting, accelerating, braking, parking avoidance, steering, light conversion and the like) of the electric locomotive by combining a software algorithm, and solves the problems of insufficient real-time, accuracy and intelligence of peripheral access, environment perception, regulation positioning, decision control and the like existing in the manual driving of the current electric locomotive by reasonably collocating and designing each device module and module, thereby realizing the unmanned intelligent driving of the electric locomotive with high efficiency and low cost. Under the scene such as iron and steel factory molten iron transportation, can liberate the people from high strength, high dangerous operational environment, reduce comprehensive cost such as human input 30%, improve cargo transportation efficiency, can wholly improve the competitiveness of relevant enterprise.

According to the vehicle-mounted intelligent computing device provided by the embodiment of the utility model, three computing modules with different functions, namely the AI computing module, the fusion computing module and the regulation positioning module, are designed in the vehicle-mounted intelligent computing device according to the principle of edge computing, the data sources are controlled through the external interface module, the data required to be computed by intelligent driving are respectively delivered to the different computing modules for processing, the demand on computing force of a single module is effectively reduced, the hardware cost is reduced, and the mode of dividing the whole computing task into a plurality of subtasks and simultaneously processing the subtasks by the different computing modules is adopted, so that the computing efficiency of intelligent driving can be effectively improved.

Specifically, the utility model can also add a redundant calculation module, provide more redundant calculation force and can share the pressure of the fusion calculation module and the regulation positioning module, such as the sensor pretreatment, the fusion calculation and the like.

According to the vehicle-mounted intelligent computing device provided by the embodiment of the utility model, the redundant computing module is arranged to supplement the performance, so that the computing pressure of the fused computing module and the rule control positioning module is shared, the computing capacity of the vehicle-mounted intelligent computing device is further enhanced, and the computing efficiency is improved.

Specifically, the utility model can also be additionally provided with a monitoring management module which is responsible for power management and state monitoring in the whole vehicle-mounted intelligent computing equipment, and the running state of the monitoring module, such as the voltage and temperature information of the monitoring module, can be monitored in real time to determine whether hardware faults exist.

The vehicle-mounted intelligent computing device provided by the embodiment of the utility model also monitors the modules in the vehicle-mounted intelligent computing device by arranging the monitoring management module, so that each module can normally operate and can alarm in time when abnormal conditions occur.

Specifically, fig. 2 is a schematic diagram of the overall design of the intelligent computing unit according to the embodiment of the present utility model, as shown in fig. 2, considering that in practical application, interfaces used by each sensor device, each positioning device, and each vehicle control unit are inconsistent.

The external interface module of the present utility model includes: at least one of LVDS interface, ETH interface, CAN bus interface, UART interface, DI interface and AI interface.

The external interface module supports the access to the electric locomotive camera equipment through an LVDS (Low Voltage Differential Signaling, low voltage differential signal) interface; supporting access to the laser radar device through a standard ETH (ethernet) interface; supporting access to millimeter wave radar equipment through a CAN bus interface (Controller Area Network ) or a standard ETH interface; support access to GNSS (Global Navigation Satellite System )/IMU (Inertial Measurement Unit, inertial navigation measurement Unit) integrated navigation equipment through CAN bus interface or UART (Universal Asynchronous Receiver/Transmitter, universal asynchronous receiver Transmitter interface); the vehicle control unit device of the electric locomotive is supported to be in butt joint with the vehicle control unit device of the electric locomotive through a CAN bus interface and a DI (switching value input)/AI (direct current analog value input) interface.

It will be appreciated that, in practical application of the present utility model, the specific type and number of interfaces included may be set according to practical requirements, which is not limited by the present utility model.

According to the vehicle-mounted intelligent computing device provided by the embodiment of the utility model, the external interface module is provided with a plurality of interfaces, the number and the types of the interfaces can be set, the peripheral access capability is provided, and the data transmission of different computing modules and peripherals is convenient to realize.

Specifically, fig. 3 is a schematic diagram of the core structure of an AI chip provided by the embodiment of the present utility model, as shown in fig. 3, the AI computing module, i.e. the AI chip with a high energy efficiency ratio, is a domestic autonomous controllable integrated memory computing chip, and the computing power 28tops@int8, the memory GDDR6, the memory bandwidth 128GB/s,16 real-time video streaming codec, pcie4.0x8, X4, EPMode, and TDP power consumption 15-30 w. In addition, other AI chips may be employed, as the utility model is not limited in this regard.

Fig. 4 is a diagram of an overall design architecture of a fusion calculation module provided by an embodiment of the present utility model, and fig. 5 is a diagram of an overall design architecture of a gauge control positioning module provided by an embodiment of the present utility model, where, as shown in fig. 4-5, the fusion calculation module and the gauge control positioning module are implemented based on a chip RK 3588. The fusion calculation module chip is referred to herein as RK3588-A, and the regulatory localization module is referred to herein as RK3588-B. The english letter suffix is used herein as a distinction between the chips only for the letter portion, and does not include the actual meaning.

In practical application, the fusion calculation module (RK 3588-A) chip is designed with 8 ARMCortex ACores, wherein 4 ARMCortex ACores are cortex A76, the main frequency can reach 2256MHz, the remaining 4 ARMCortex ACores are cortex A55, and the main frequency can reach 1800MHz. The total calculated power of 8 cores can reach about 100 KDMIPS. In addition, RK3588 also had 3 ARMCortex-M0 MCUs, 4 ARMMali-G610MP4 GPUs, and ISPs that could process 48MPixels image data.

The fusion calculation module (RK 3588-A) is mainly responsible for running perception algorithm applications, including camera preprocessing, lidar preprocessing, perception, fusion, prediction, and middleware and drivers on which these applications depend. And the peripheral part is connected with a plurality of GMSL2 cameras and the laser radar. Meanwhile, the fusion calculation module (RK 3588-A) is the only master chip capable of directly dispatching the AI chip with high energy efficiency ratio to perform reasoning operation, and the interaction mode is through the PCI-E channel.

The fusion calculation module (RK 3588-A) uses the following for the external communication interface:

1. and the monitoring management module (TC 397) is used for carrying out monitoring data transmission through one UART. The monitoring management module (TC 397) monitors the operating state of the fusion calculation module (RK 3588-a), such as hardware failure, voltage, temperature, etc.

2. The access PCI-Eswitch and the high energy efficiency ratio AI chip perform the related data transmission of the algorithm model operation through the PCI-E, and the method comprises the following steps: algorithm model, sensor data, operation result.

3. The possible fault reasons can be obtained when the working state of the high energy efficiency ratio AI chip, especially the PCI-E communication is abnormal, is read through SPI communication with the high energy efficiency ratio AI chip.

4. Sensor data and traffic data can be transceived within the domain by RGMII connection to an ethernet switch.

The fusion calculation module (RK 3588-A) is externally connected with a memory: 1. and the external connection is 16GBDDR4, and the external connection is used as a program running memory. 2. And the external 64GBeMMC is used for storing system software and algorithm programs. 3. And the NVMe solid state disk is externally connected with an M.2 interface and is used for storing sensor data and log files.

The chip of the regulation and positioning module (RK 3588-B) is designed with 8 ARMCortex ACores, wherein 4 ARMCortex ACores are cortex A76, the main frequency can reach 2256MHz, the remaining 4 ARMCortex ACores are cortex A55, and the main frequency can reach 1800MHz. The total calculated power of 8 cores can reach about 100 KDMIPS. In addition, RK3588 also had 3 ARMCortex-M0 MCUs, 4 ARMMali-G610MP4 GPUs, and ISPs that could process 48MPixels image data.

The regulation and positioning module (RK 3588-B) is mainly responsible for running applications of pure CPU operations, including positioning, planning and control, as well as middleware and drivers on which these applications depend. The peripheral part is connected with a plurality of laser radars and 1 integrated navigation.

The gauge positioning module (RK 3588-B) uses the following for the external communication interface:

1. and the monitoring management module (TC 397) is used for carrying out monitoring data transmission through one UART. The monitoring management module (TC 397) monitors the operating conditions of the regulatory location module (RK 3588-B), such as hardware faults, voltage, temperature, etc.

2. The access PCI-Eswitch can carry out large-flow data transmission with other RK3588, such as laser radar point cloud data.

3. And acquiring positioning sensor data through UART connection integrated navigation.

4. And acquiring GPS time through PPS connection integrated navigation, and providing timebase for intra-domain time synchronization.

5. Sensor data and traffic data can be transceived within the domain by RGMII connection to an ethernet switch.

The gauge positioning module (RK 3588-B) is externally connected with a memory: 1. and the external connection is 16GBDDR4, and the external connection is used as a program running memory. 2. And the external 64GBeMMC is used for storing system software and algorithm programs. 3. And the NVMe solid state disk externally connected with the 512GBM.2 interface is used for storing sensor data and log files.

Fig. 6 is a schematic diagram of an overall design of a redundant computing module according to an embodiment of the present utility model, as shown in fig. 6, it can be understood that, on the basis of this schematic diagram, the redundant computing unit is also optionally RK3588, and is denoted as RK3588-C.

In practical application, the redundant calculation module (RK 3588-C) chip is designed with 8 ARMCortex ACores, wherein 4 ARMCortex ACores are cortex A76, the main frequency can reach 2256MHz, the remaining 4 ARMCortex ACores are cortex A55, and the main frequency can reach 1800MHz. The total calculated power of 8 cores can reach about 100 KDMIPS. In addition, RK3588 also had 3 ARMCortex-M0 MCUs, 4 ARMMali-G610MP4 GPUs, and ISPs that could process 48MPixels image data.

A redundancy calculation module (RK 3588-C) is used for the Max version of the intelligent computing unit for performance replenishment. The computation pressure of the fusion computation module (RK 3588-A)/B can be shared, such as the post-fusion and laser radar point cloud preprocessing are split to run on the redundant computation module (RK 3588-C).

The redundant computing module (RK 3588-C) uses the following for the external communication interface:

1. and the monitoring management module (TC 397) is used for carrying out monitoring data transmission through one UART. The monitoring management module (TC 397) monitors the operating status of the redundant computing module (RK 3588-C), such as hardware faults, voltages, temperatures, etc.

3. Sensor data and traffic data can be transceived within the domain by RGMII connection to an ethernet switch.

The redundant computing module (RK 3588-C) is externally connected with a memory: 1. and the external connection is 16GBDDR4, and the external connection is used as a program running memory. 2. And the external 64GBeMMC is used for storing system software and algorithm programs. 3. And the NVMe solid state disk externally connected with the 512GBM.2 interface is used for storing sensor data and log files.

According to the vehicle-mounted intelligent computing device provided by the embodiment of the utility model, three computing modules with different functions, namely an AI computing module (AI chip), a fusion computing module and a regulation positioning module (RK 3588), are designed in the vehicle-mounted intelligent computing device according to the principle of edge computing, the data sources are controlled through the external interface module, the data required to be computed by intelligent driving are respectively delivered to different computing modules for processing, the demand on the computing force of a single module is effectively reduced, the hardware cost is reduced, and the mode of dividing the whole computing task into a plurality of subtasks and simultaneously processing the subtasks by the different computing modules is adopted, so that the computing efficiency of intelligent driving can be effectively improved.

Specifically, fig. 7 is a schematic diagram of an overall design of a monitoring management module according to an embodiment of the present utility model, and as shown in fig. 7, the monitoring management module uses a TC397 chip.

In practical application, the monitoring management module (TC 397) is respectively connected with the high energy efficiency ratio AI chip, the fusion calculation module (RK 3588-A), the regulation positioning module (RK 3588-B) and the redundancy calculation module (RK 3588-C) through 1-path UART, transmits monitoring data, and monitors state information of each module, such as hardware faults, voltage, temperature and the like.

The vehicle-mounted intelligent computing device provided by the embodiment of the utility model also adopts the TC397 chip to set the monitoring management module to monitor the modules in the vehicle-mounted intelligent computing device, so that each module can be ensured to normally operate and alarm in time when abnormal conditions occur.

In combination with the specific example of the utility model, the intelligent driving data flow of the electric locomotive and the communication architecture of the vehicle-mounted intelligent computing equipment are described when the utility model is practically applied.

Fig. 8 is a schematic diagram of a communication architecture of a vehicle-mounted intelligent computing device according to an embodiment of the present utility model, as shown in fig. 8, 1) a monitoring management module communicates with an AI computing module, a fusion computing module, a regulation positioning module, and a redundant computing module by using UART interfaces, so as to obtain working state information of each module.

2) The intelligent computing unit is communicated with the camera of the electric locomotive by using a FAKRA interface, is communicated with the laser radar by using an Ethernet1000baseT1/Tx interface, and is communicated with the GNSS & IMU integrated navigation by using a UART or CAN bus to acquire the perception data of each sensor.

3) The intelligent computing unit communicates with the vehicle control unit of the electric locomotive by using a CAN bus or an Ethernet1000baseT1/Tx interface, and transmits the perception, regulation and decision information of the intelligent computing unit to the vehicle control unit in real time.

4) The AI calculation module, the fusion calculation module, the regulation positioning module and the redundant calculation module are communicated through PCI-ESwitch and Ethernet switch.

Fig. 9 is a schematic diagram of intelligent driving data flow of an electric locomotive according to an embodiment of the present utility model, as shown in fig. 9, 1) sensor data acquisition: the fusion calculation module obtains a camera video stream through a GMSL interface and obtains laser radar data through an Ethernet interface.

2) Sensor data preprocessing: the fusion calculation module respectively carries out preprocessing (denoising, smoothing, brightness conversion and the like) on the camera video stream and the laser radar point cloud.

3) Target perception recognition: and (3) invoking an AI chip with high energy efficiency ratio, and detecting and identifying the front vehicle tail, the track, the pedestrians, the obstacles and the like by using a model algorithm and AI calculation force of the AI chip.

4) Positioning an electric locomotive: and the gauge control positioning module performs multi-sensor fusion positioning calculation on the GNSS and IMU integrated navigation data to acquire high-precision positioning data of the motor vehicle.

5) Intelligent driving prediction: the fusion calculation module analyzes the behavior patterns of other traffic participation objects based on the self-vehicle position and the surrounding environment model provided by the target perception identification data and the high-precision positioning data of the motor vehicle, and evaluates the track trend in a future time range of the traffic participation objects, so that the decision planning capability of the intelligent driving system in a complex scene is improved.

6) Intelligent driving planning control: the path planning receives the upper layer perception prediction result, which comprises three parts of route searching, behavior decision and motion planning. The electric locomotive control module combines the path planning information, the vehicle body attribute and the dynamic calculation of external physical factors, converts the path planning information, the vehicle body attribute and the dynamic calculation of external physical factors into decision data for locomotive control, and transmits the relevant control decision data to a vehicle control unit of the electric locomotive through a CAN bus to realize specific control of throttle, brake, direction and the like of the electric locomotive.

The foregoing embodiment is merely illustrative of a specific example of the present utility model, and is not intended to limit the present utility model.

Specifically, the vehicle-mounted intelligent control system provided by the utility model is loaded on a vehicle, environmental data and position data are acquired through the vehicle-mounted sensor equipment and the vehicle-mounted positioning equipment, the vehicle-mounted intelligent computing equipment processes the acquired data and performs route navigation planning and intelligent decision control (such as starting, accelerating, braking, parking avoidance, steering, light transformation and the like), and the intelligent control on the vehicle is realized through the vehicle control unit. The specific structure and function of the vehicle-mounted intelligent computing unit are as described above, and are not described in detail herein.

The vehicle-mounted intelligent control system provided by the embodiment of the utility model comprises vehicle-mounted intelligent computing equipment, vehicle-mounted sensor equipment, vehicle-mounted positioning equipment and a vehicle control unit, wherein three computing modules with different functions, namely an AI computing module, a fusion computing module and a regulation positioning module, are designed in the vehicle-mounted intelligent computing equipment, data of the vehicle-mounted sensor equipment and the vehicle-mounted positioning equipment are obtained through an external interface module, data sources are controlled, data required to be calculated for intelligent driving are respectively delivered to the different computing modules for processing, and intelligent driving is realized through the vehicle control unit. The method effectively reduces the demand on calculation force of a single module, reduces hardware cost, and can effectively improve intelligent driving calculation efficiency by adopting a mode of dividing an integral calculation task into a plurality of subtasks and simultaneously processing the subtasks by different calculation modules.

Specifically, the in-vehicle sensor apparatus includes: at least one of a camera, a laser radar sensor and a millimeter wave radar sensor. The sensing data CAN be obtained by interfacing with the vehicle-mounted sensor through the external interface module and through standard interfaces such as a CAN bus interface, a standard ETH interface, a UART interface, an LVDS interface and the like.

In practical application, the specific type and number of the vehicle-mounted sensors can be set according to practical requirements, and the utility model is not limited to the specific type and number.

According to the vehicle-mounted intelligent control system provided by the embodiment of the utility model, the acquisition of the sensing data around the vehicle is realized by arranging at least one vehicle-mounted sensor device, the data is transmitted to the calculation module through the external interface module, the data processing is carried out, the intelligent sensing recognition of the surrounding environment of the vehicle is realized, and the data support is provided for intelligent driving.

Specifically, fig. 10 is a schematic diagram of a multi-sensor fusion positioning architecture provided by an embodiment of the present utility model, as shown in fig. 10, a vehicle-mounted positioning device adopts a GNSS/IMU integrated navigation device, obtains positioning satellite signals through a GNSS, and implements high-precision positioning according to Real-Time Kinematic (RTK) technology, assisted by differential signals of a ground reference standard value station. And measuring the triaxial acceleration and the triaxial angular velocity of the electric locomotive by an INS (Inertial Navigation System) and performing dead reckoning to realize high-precision positioning.

In practical application, the high-precision positioning of the motor vehicle can be realized through SLAM (Simultaneous Localization and Mapping, instant positioning and map construction) technology for performing feature matching on the laser radar real-time point cloud and a pre-stored map. And the positioning data of the three sensors are fused and calculated based on a gauge control positioning module (RK 3588-B), so that the higher-precision positioning of the electric locomotive is realized. And combining the acquired environment sensing data, performing intelligent sensing recognition and high-precision fusion positioning on the surrounding environment of the electric locomotive, and realizing vehicle intelligent control through a reliable decision control algorithm.

According to the vehicle-mounted intelligent control system provided by the embodiment of the utility model, the acquisition of the vehicle position information is realized by arranging the GNSS/IMU integrated navigation equipment, the high-precision positioning is realized, and the data support is provided for intelligent driving.

The specific implementation manner is consistent with the control manner, and will not be repeated here.

According to the vehicle-mounted intelligent computing device provided by the embodiment of the utility model, the three computing modules with different functions, namely the AI computing module, the fusion computing module and the regulation positioning module, are designed in the vehicle-mounted intelligent computing device, the data sources are controlled through the external interface module, the data required to be computed by intelligent driving are respectively delivered to the different computing modules for processing, the demand on the computing force of a single module is effectively reduced, the hardware cost is reduced, and the whole computing task is divided into a plurality of subtasks and is processed by the different computing modules, so that the intelligent driving computing efficiency can be effectively improved.

Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present utility model, and are not limiting; although the utility model has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present utility model.

Claims

1. An in-vehicle intelligent computing device, comprising: the system comprises an external interface module, an AI calculation module, a fusion calculation module and a regulation positioning module;

wherein the external interface module comprises a plurality of interfaces;

and the fusion calculation module is respectively connected with the AI calculation module and the regulation positioning module.

2. The in-vehicle intelligent computing device of claim 1, further comprising: a redundancy calculation module;

and the redundancy calculation module is respectively connected with the AI calculation module and the regulation positioning module.

3. The in-vehicle intelligent computing device of claim 1, further comprising: a monitoring management module;

and the monitoring management module is respectively connected with the external interface module, the AI calculation module, the fusion calculation module and the regulation positioning module.

4. The vehicle-mounted intelligent computing device of claim 1, wherein the external interface module comprises: at least one of LVDS interface, ETH interface, CAN bus interface, UART interface, DI interface and AI interface.

5. The vehicle-mounted intelligent computing device of any of claims 1-4, wherein the AI computing module employs an AI chip;

and the fusion calculation module and the regulation positioning module adopt RK3588 chips.

6. The vehicle-mounted intelligent computing device of claim 3, wherein the monitoring management module employs a TC397 chip.

7. An on-board intelligent control system comprising the on-board intelligent computing device of any one of claims 1-6, the on-board sensor device, the on-board positioning device, and the vehicle control unit.

8. The vehicle-mounted intelligent control system of claim 7, wherein the vehicle-mounted sensor apparatus comprises: at least one of a camera, a laser radar sensor and a millimeter wave radar sensor.

9. The vehicle-mounted intelligent control system of claim 7, wherein the vehicle-mounted positioning device employs a GNSS/IMU integrated navigation device.

10. An intelligent drive vehicle, characterized in that it is equipped with an on-board intelligent control system according to any one of claims 7-9.