Intelligent vehicle control system for special operation and intelligent vehicle

Abstract

The invention discloses an intelligent vehicle control system for special operation and an intelligent vehicle, which support various types of sensors and executive devices, and avoid the adoption of expensive or customized special modules for expanding functions through highly integrated and standardized hardware design, so that the cost is reduced, and the working efficiency is improved. And on the premise of ensuring high performance and high expansibility of the system, the hardware manufacturing cost is effectively controlled and the remarkable advantage of low cost is realized through optimized mainboard interface layout and core components with economical model selection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent control, in particular to an intelligent vehicle control system for special operation. BACKGROUND

[0002] With the development of intelligent control technology, intelligent vehicles are increasingly applied to various scenarios, especially in special operation scenarios such as fire scene rescue, earthquake scene rescue, and other dangerous scenarios.

[0003] However, the intelligent vehicles currently used for special operations have the following problems:

[0004] 1. Vehicle power and control scheme has power consumption and volume defects;

[0005] The vehicle control unit (VCU) of the intelligent vehicle currently used for special operation mostly adopts traditional control strategies such as PID, which is difficult to balance high dynamic response and energy efficiency in FOC (field-oriented control) applications, especially when the load suddenly changes (such as obstacle climbing and slope climbing), which is prone to current shock and energy waste, resulting in high overall system power consumption and seriously affecting the vehicle's endurance;

[0006] Although the conventional FOC algorithm can achieve basic motor control, the current loop PI parameters are usually fixed and cannot be adjusted online according to complex working conditions such as sudden changes in road impedance and dynamic changes in load, resulting in a significant decline in control performance in unstructured terrain, which is manifested as insufficient torque response or oscillation, affecting the passability and energy efficiency;

[0007] At the same time, the overall vehicle structure is bulky and difficult to adapt to narrow, complex or unstructured operating environments, limiting its deployment and maneuverability in specific scenarios;

[0008] 2. Low integration of perception system, insufficient expansion capability;

[0009] The existing control unit generally does not integrate a camera module dedicated interface, which cannot directly expand camera devices on the mainboard, and needs to rely on external conversion or other communication interfaces to achieve image acquisition;

[0010] Such indirect connection not only occupies limited system resources, increases system complexity and data delay, and seriously restricts the deployment of multi-sensor tight coupling algorithms (such as VIO-SLAM), affecting the efficiency of real-time perception and response in special operating environments;

[0011] 3. Weak environmental adaptability, difficult to cope with harsh working conditions;

[0012] In the operation scene of insufficient light, closed space or visual interference such as dust, smoke and the like (such as underground mining, tunnel, night rescue and the like), it is difficult to realize effective environment perception by relying on visible light camera, and simple image switching or primary fusion algorithm almost fails in the scene of thick smoke, high dust and the like, which limits the application range and operation reliability of the intelligent vehicle.

[0013] Therefore, it is necessary to improve the existing special operation intelligent vehicle.

[0014] The above content is only used to assist in understanding the technical solutions of the present application and does not represent the acknowledgement of the above content as prior art. SUMMARY

[0015] The main purpose of the present application is to provide an intelligent vehicle control system for special operation and an intelligent vehicle. While supporting various types of sensors and actuators, the highly integrated and standardized hardware design avoids the use of expensive or customized special modules for expansion, and instead uses optimized mainboard interface layout and economically selected core components to effectively control the hardware manufacturing cost under the premise of ensuring high performance and high expandability of the system, thereby achieving significant cost advantage.

[0016] To achieve the above purpose, in a first aspect, the present application provides a control system for a special operation intelligent vehicle, comprising:

[0017] an integrated interface module, a vehicle control unit and a motor drive module connected electrically

[0018] The integrated interface module is provided with a plurality of hardware interfaces for connecting functional modules and external devices of the special operation intelligent vehicle.

[0019] The vehicle control unit comprises a real-time control kernel and a task coprocessor connected electrically.

[0020] The real-time control kernel is used to collect target data through the integrated interface module, and to perform real-time task analysis according to the target data to obtain real-time task analysis results.

[0021] The task coprocessor is used to perform non-real-time task analysis according to the target data to obtain non-real-time task analysis results.

[0022] The motor drive module is used to generate control instructions according to the real-time task analysis results and the non-real-time task analysis results, and to send the control instructions to target motors.

[0023] In an embodiment, the real-time control kernel further comprises an adaptive magnetic field oriented control unit.

[0024] The adaptive magnetic field oriented control unit is configured to predict a future dynamic of the system within a preset limited time domain through a preset prediction model according to the target data.

[0025] With the minimum total energy consumption of the intelligent vehicle system as a target, a set of optimal voltage vector sequences for representing optimal torque output is solved online according to preset hard constraint conditions and the future dynamic.

[0026] In an embodiment, the real-time control core is a heterogeneous platform combining an ARM advanced reduced instruction set machine and an FPGA field programmable gate array.

[0027] In an embodiment, the adaptive magnetic field oriented control unit further includes an online parameter identification subunit configured to update key parameters in the prediction model in real time according to operation data of each motor of the intelligent vehicle.

[0028] In an embodiment, the task coprocessor further includes a cross-modal attention fusion perception unit.

[0029] The cross-modal attention fusion perception unit is configured to directly acquire two pieces of image data, i.e., visible light image data and infrared image data, through the real-time control core.

[0030] According to data qualities of the two pieces of image data, feature weights of the two pieces of image data are dynamically adjusted at a feature fusion layer to obtain a dynamic feature distribution of the image data.

[0031] In an embodiment, the task coprocessor is an SoC system-on-a-chip integrating a CPU central processing unit and a GPU graphics processing unit or a CPU central processing unit and an NPU neural network processing unit.

[0032] The task coprocessor further integrates a multi-path high-speed camera module interface, so that a high-bandwidth sensor for collecting visible light image data or infrared image data is directly connected to a mainboard of the vehicle control unit.

[0033] In an embodiment, the task coprocessor further includes a simultaneous localization and mapping optimization unit.

[0034] The simultaneous localization and mapping optimization unit is configured to perform five-modal tight coupling of geometric, inertial, kinematic, semantic, and ranging data in a unified back-end optimizer according to preset constraint factors, optimize a simultaneous localization and mapping result, and obtain a simultaneous localization and mapping optimization result.

[0035] In an embodiment, the preset constraint factors include:

[0036] A common constraint factor of the VIO visual-inertial odometer and the wheeled odometer, specifically, the count of the VIO visual-inertial odometer is used to correct the heading drift of the wheeled odometer caused by wheel slip, and the count of the wheeled odometer is used to provide an absolute scale constraint for the VIO visual-inertial odometer;

[0037] A geometric plane constraint factor, specifically, the environmental information recognized by the task coprocessor is used as a high-level prior to introduce a geometric plane constraint in the mapping optimization;

[0038] A one-dimensional height above ground constraint factor, specifically, a constraint factor with a confidence reaching a preset value is introduced according to the precise ranging sensor data in the target data to anchor the Z-axis coordinate and eliminate the vertical accumulated drift.

[0039] In an embodiment, the real-time control core and the task coprocessor are connected through a PCIe peripheral component interconnect express high-speed serial bus.

[0040] In addition, in a second aspect, the application further provides an intelligent vehicle provided with the control system for a special operation intelligent vehicle according to any one of the first aspect, and further comprising:

[0041] A pair of longitudinal beams arranged in parallel;

[0042] A front cross beam and a rear cross beam, two ends of the front cross beam are respectively rigidly fixedly connected to the bottom of the front end of the pair of longitudinal beams, and two ends of the rear cross beam are respectively rigidly fixedly connected to the bottom of the rear end of the pair of longitudinal beams;

[0043] The front cross beam and the rear cross beam are respectively connected with two independent wheel hubs through a swing beam, and the front cross beam and the swing beam and the rear cross beam and the swing beam respectively form a door-shaped bearing frame, and the swing beam is configured to swing up and down to realize suspension adjustment of the two independent wheel hubs connected therewith;

[0044] Each of the independent wheel hubs is respectively connected with an independent motor, and each of the independent wheel hubs is further respectively provided with a wheel hub brake;

[0045] A battery and a power control module, the battery and the power control module are used to supply power for the control system for a special operation intelligent vehicle, the independent motors and the wheel hub brakes.

[0046] The intelligent vehicle control system for special operation and the intelligent vehicle provided by the application support multiple types of sensors and actuators, and through highly integrated and standardized hardware design, the use of expensive or customized special modules for expansion is avoided, and instead, through optimized mainboard interface layout and selection of economical core components, the hardware manufacturing cost is effectively controlled under the premise of ensuring high performance and high expansibility of the system, and significant low-cost advantage is achieved. Attached Figure Description

[0047] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0048] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 A schematic structural diagram of an embodiment of the control system for a special-operation intelligent vehicle in this application;

[0050] Figure 2 This is a schematic diagram of an embodiment of an intelligent vehicle according to this application;

[0051] Figure 3 for Figure 2 A schematic diagram of the control system.

[0052] Explanation of icon numbers:

[0053] 1. Longitudinal beam; 2. Swing beam; 3. Rear crossbeam; 4. Drive-by-wire brake pump; 5. Battery and power control module; 6. Independent motor; 7. Front crossbeam; 8. Control system for special operation intelligent vehicle; 801. First communication interface; 802. First CAN interface; 803. Second CAN interface; 804. First GPS interface; 805. Second GPS interface; 806. Ethernet interface; 807. I2C interface; 808. First power interface; 809. Second power interface; 810. SPI interface; 811. USB interface; 812. AD & IO; 813. Digital spectrum modulation receiver interface; 814. Second communication interface; 815. Third communication interface; 816. UART4 & I2C interface; 9. Independent wheel hub.

[0054] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0055] The technical solutions in the present application will be described clearly and completely in the present application with reference to the drawings. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. The components of the present application described and shown in the drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of the present application.

[0056] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. Meanwhile, in the description of the present application, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0057] Embodiment 1

[0058] The embodiments of the present application provide a control system for a special operation intelligent vehicle, referring to Figure 1 , comprising an integrated interface module, a vehicle control unit and a motor drive module connected electrically

[0059] The integrated interface module is provided with a plurality of hardware interfaces for connecting functional modules and external devices of the special operation intelligent vehicle;

[0060] The vehicle control unit comprises a real-time control kernel and a task coprocessor connected electrically;

[0061] The real-time control kernel is used to collect target data through the integrated interface module, and to perform real-time task analysis according to the target data to obtain real-time task analysis results;

[0062] The task coprocessor is used to perform non-real-time task analysis according to the target data to obtain non-real-time task analysis results;

[0063] The motor drive module is used to generate control instructions according to the real-time task analysis results and the non-real-time task analysis results, and to send the control instructions to target motors.

[0064] Specifically, the control system for the special operation intelligent vehicle in this embodiment is electrically connected to the main GPS module, auxiliary GPS module, lidar, servo motor, telemetry system, infrared / visible light camera, Ethernet module, radio receiver and other functional modules of the intelligent vehicle, as well as four independent motors 6, through an integrated interface module. The integrated interface module also reserves hardware interfaces such as CAN, UART, I2C, SPI, ADC&GPIO, USB, and Micro SD for connecting external devices. The radio receiver is also used for data interaction with the radio transmitter of the intelligent vehicle. The control system for the special operation intelligent vehicle is powered by the battery and power control module 5 of the intelligent vehicle. The battery and power control module 5 of the intelligent vehicle also powers other functional modules, independent motors 6 and external devices. The target data includes data from the above-mentioned functional modules, data from the external devices with reserved interfaces, data from the independent motors 6 and data from the battery and power control module 5.

[0065] In this embodiment, a hierarchical heterogeneous vehicle control unit architecture is adopted. The real-time control kernel is dedicated to high-frequency, real-time vehicle dynamics control, attitude calculation, IMU data acquisition, CAN bus and I2C / UART I / O management, and other tasks. The task coprocessor is dedicated to high-computing-power-consuming non-real-time tasks, such as AI image recognition, multimodal fusion, SLAM mapping, and other tasks.

[0066] The real-time control kernel and the task coprocessor are interconnected via an internal high-speed PCIe bus, enabling low-latency, high-time-sequence synchronous exchange of IMU, wheel speed, and camera data at the microsecond level. This provides a solid hardware foundation for achieving high-performance algorithm fusion. The low-latency path formed between the real-time control kernel and the coprocessor allows high-frequency dynamic states (such as motor torque and wheel slip ratio) to be quickly fed back to the coprocessor for optimization calculations. The calculation results can then be instantly sent back to the real-time control kernel for execution, achieving a "perception-decision-control" closed-loop effect that traditional discrete controllers cannot achieve.

[0067] The vehicle control unit has a wealth of communication interfaces, which facilitates the connection of various peripheral devices and adapts to various operating scenarios.

[0068] Example 2

[0069] This embodiment is further optimized based on embodiment 1. In this embodiment, the real-time control kernel further includes an adaptive magnetic field orientation control unit.

[0070] The adaptive magnetic field orientation control unit is used to predict the future dynamics of the system within a preset finite time domain based on the target data and a preset prediction model.

[0071] With the goal of minimizing the total energy consumption of the intelligent vehicle system, a set of optimal voltage vector sequences for characterizing the optimal torque output is solved online based on preset hard constraints and the future dynamics.

[0072] This implementation method improves the FOC field orientation control algorithm based on MPC model predictive control, and provides an adaptive field orientation control unit with an online rolling optimization framework.

[0073] Specifically, within each control cycle, based on the high-frequency motor status (current, speed, position) and vehicle dynamics information (such as ground resistance torque estimated by the model) collected by the real-time control kernel, the future dynamics of the system are predicted within a finite time domain. The primary goal is to minimize the total energy consumption of the intelligent vehicle system. At the same time, physical quantities such as phase current, DC bus voltage, and torque output are used as hard constraints to solve a set of optimal voltage vector sequences online. In principle, this avoids the torque / current impact caused by traditional PID control that relies solely on the current error, and achieves the inherent unity of high dynamic response and optimal energy efficiency.

[0074] Compared to traditional PID control, this adaptive FOC-MPC unit can not only predict the optimal torque output in the short time domain, but also greatly reduce the current surge and energy waste during load changes, achieving online rolling optimization and resolving the contradiction between speed and stability / economy in the dynamic operation of special vehicles in traditional control.

[0075] Furthermore, the real-time control kernel is a heterogeneous platform combining an ARM Advanced Reduced Instruction Set Machine and an FPGA Field Programmable Gate Array.

[0076] This implementation method allows ARM to focus on core control logic and FPGA to focus on underlying data and interface management. It can improve the overall computing efficiency of the real-time control kernel while maintaining low power consumption. At the same time, it has strong stability and can resist interference from complex operating conditions.

[0077] Furthermore, the adaptive magnetic field orientation control unit also includes an online parameter identification subunit, which is used to update the key parameters in the prediction model in real time based on the operating data of each motor of the intelligent vehicle.

[0078] Using this implementation method, the online parameter identification subunit can update the key parameters in the prediction model in real time based on motor operating data (such as stator resistance changes and inductor saturation effects), ensuring that the model can accurately reflect the real characteristics of the motor under different temperatures and loads. This overcomes the defects of traditional FOC parameter fixation, improves the accuracy and robustness of control under complex operating conditions, and realizes adaptive parameter updating.

[0079] In this embodiment, the task coprocessor further includes a cross-modal attention fusion perception unit;

[0080] The cross-modal attention fusion perception unit is used to directly acquire two image data streams, visible light image data and infrared image data, through the real-time control kernel.

[0081] Based on the data quality of the two image data streams, the feature weights of the two image data streams are dynamically adjusted in the feature fusion layer to obtain the dynamic feature distribution of the image data.

[0082] Specifically, in this embodiment, for severe visual conditions such as smoke and dust in special operations, the cross-modal attention fusion perception unit adopts a dual-stream CNN architecture combined with an attention mechanism. When the quality of one sensor data (such as visible light) deteriorates, the attention mechanism will automatically reduce the weight of that information in the feature fusion layer and instead rely on the effective information of another channel (such as infrared) to maintain stable perception performance. This ensures that the VCU still has reliable visual perception capabilities under extreme conditions, breaking through the limitations of traditional multi-sensor fusion in terms of bandwidth, synchronization and robustness, and significantly enhancing the target recognition capability in severe environments such as smoke and dust.

[0083] Furthermore, the task coprocessor is a SoC (System-on-a-Chip) that integrates a CPU and a GPU or integrates a CPU and an NPU (Neural Processing Unit).

[0084] The task coprocessor also integrates multiple high-speed camera module interfaces, enabling high-bandwidth sensors used to acquire visible light or infrared image data to be directly connected to the mainboard of the vehicle control unit.

[0085] By adopting this implementation method, the SoC on-chip system is dedicated to undertaking high computing power non-real-time tasks, allowing the ARM+FPGA to focus on the core control logic, further avoiding resource conflicts between control and computing, and thus further reducing the response latency of key tasks such as vehicle dynamics control and CAN bus communication.

[0086] By natively integrating multiple high-speed camera module interfaces, such as MIPI / Ethernet interfaces, on the task coprocessor, high-bandwidth sensors such as visible light and infrared cameras can be directly connected to the VCU motherboard, realizing a one-stop interface integration solution. This fundamentally avoids the cost of adding extra adapter boards or modules to connect different peripherals, making the overall hardware a low-cost and cost-effective solution when supporting many types of devices. It solves the problem of traditional solutions requiring external converters and occupying system resources. At the same time, it avoids the latency and bandwidth bottlenecks of traditional solutions via USB or serial port conversion, achieving true zero-latency perception.

[0087] At the same time, it also provides highly available perceptual input for subsequent semantic SLAM algorithms.

[0088] In this embodiment, the task coprocessor further includes a synchronous localization and mapping optimization unit;

[0089] The synchronous positioning and mapping optimization unit is used to optimize the synchronous positioning and mapping results in a unified back-end optimizer by performing five-modal tight coupling of geometric, inertial, kinematic, semantic and ranging data according to preset constraint factors, so as to obtain the synchronous positioning and mapping optimization results.

[0090] Specifically, this embodiment provides a robust autonomous localization and mapping algorithm deeply integrated into the hardware architecture of the vehicle control unit (VCU). This algorithm runs on the task coprocessor of the VCU, achieving high cohesion and low latency. All data inputs come from within the VCU. It tightly couples the visual and semantic information processed by the task coprocessor through the native high-speed camera module interface, the high-frequency IMU data transmitted by the real-time control kernel through the internal high-speed bus, and the wheel odometer and precision ranging data from the CAN / I2C bus. This enables unified optimization of multi-source information at the back end, solving the problem of single sensor failure in unstructured environments such as mine tunnels, tunnels, and indoor environments. At the same time, the introduction of semantic information to assist localization greatly improves the autonomous localization accuracy and anti-interference capability of the vehicle under extreme conditions.

[0091] Furthermore, the preset constraint factors include:

[0092] The common constraint factor for VIO visual inertial odometry and wheel odometry is to use the count of VIO visual inertial odometry to correct the heading drift of wheel odometry caused by wheel slippage, and to use the count of wheel odometry to provide absolute scale constraint for VIO visual inertial odometry.

[0093] The geometric plane constraint factor specifically uses the environmental semantic information identified by the task coprocessor as a high-level prior to introduce geometric plane constraints in the graph optimization process.

[0094] The one-dimensional ground height constraint factor is specifically introduced based on the precision ranging sensor data in the target data, with a constraint factor having a confidence level that reaches a preset value, in order to anchor the Z-axis coordinate and eliminate vertical cumulative drift.

[0095] By adopting this implementation method, the common constraint factors of VIO visual inertial odometer and wheeled odometer are combined, and the advantages of both are combined, making it more suitable for the complex working scenarios of special operation intelligent vehicles.

[0096] By using environmental semantic information such as "ground" and "walls" identified by the task coprocessor as high-level priors, geometric plane constraints are introduced into graph optimization. This greatly enhances the robustness of IMU pose in weakly textured environments and can automatically remove dynamic obstacles to prevent map pollution.

[0097] By utilizing the precision ranging sensor data accessed through the VCU's I2C / CAN interface, a high-confidence one-dimensional ground altitude constraint factor is introduced to completely anchor the Z-axis coordinate and eliminate vertical cumulative drift.

[0098] In summary, by tightly coupling five types of information—geometric, inertial, kinematic, semantic, and ranging—within the VCU, the positioning robustness of the VCU in special operating environments such as those without satellite signals, weak textures, and high dynamic interference is improved.

[0099] Example 3

[0100] This embodiment provides an example of an intelligent vehicle, referring to... Figure 2 The intelligent vehicle is equipped with a control system 8 for special operation intelligent vehicles as shown in Embodiment 1 or Embodiment 2, and further includes:

[0101] A pair of parallel longitudinal beams 1;

[0102] The front crossbeam 7 and the rear crossbeam 3 are respectively rigidly fixed at both ends to the bottom of the front end of the pair of longitudinal beams 1, and the two ends of the rear crossbeam 3 are respectively rigidly fixed to the bottom of the rear end of the pair of longitudinal beams 1.

[0103] The front crossbeam 7 and the rear crossbeam 3 are respectively connected to two independent wheel hubs 9 via a swing beam 2. The front crossbeam 7 and the swing beam 2, and the rear crossbeam 3 and the swing beam 2 respectively form a portal-shaped load-bearing frame. The swing beam 2 is configured to swing up and down to achieve suspension adjustment of the two connected independent wheel hubs 9.

[0104] Each of the independent wheel hubs 9 is connected to an independent motor 6, and each of the independent wheel hubs 9 is also equipped with a wheel hub brake;

[0105] The battery and power control module 5 is used to supply power to the control system 8 for the special operation intelligent vehicle, the independent motor 6 and the wheel hub brake.

[0106] Specifically, in this embodiment, a miniaturized vehicle body structure is adopted, equipped with four low-power 100W independent motors 6, each of the four wheels is equipped with an independent motor 6, directly driving the wheel hub. The motor drive module can realize independent control of the speed of the four wheels, which significantly reduces the total power consumption of the system while ensuring sufficient driving force and obstacle crossing ability, making it more suitable for flexible and long-term operation in complex and restricted environments.

[0107] The swing beam 2 is connected to the front crossbeam 7 and the rear crossbeam 3 to form a portal-shaped load-bearing frame. The up and down swing of the swing beam 2 realizes the suspension adjustment, realizing the structural integration of differential speed function and suspension function.

[0108] Preferably, an integrated aluminum alloy frame design is adopted, which integrates the drive system with the longitudinal beam 1 through casting, resulting in significant weight reduction compared to the traditional split structure. High-polymer composite materials are used in components such as the swing beam 2, achieving a balance between "load-weight reduction-differential speed".

[0109] More preferably, a helical spring is installed between the swing beam 2 and the front crossbeam 7 / rear crossbeam 3 to absorb road impact. A load-bearing shaft is added at the connection between the drive shaft of each independent motor 6 and the wheel hub, and is bolted to the mounting base of the independent motor 6. This can prevent the drive shaft from deforming due to downward pressure and limit lateral displacement. The front crossbeam 7 / rear crossbeam 3 is integrally forged. The longitudinal beam 1 is rigidly connected to the front crossbeam 7 / rear crossbeam 3 by high-strength bolts to form a stable triangular support structure, thereby optimizing the force transmission path.

[0110] In one specific implementation, the two longitudinal beams 1, the front crossbeam 7, the rear crossbeam 3, and the independent motor 6 mounting base form the core load-bearing unit. They are rigidly connected by high-strength bolts with a preload of up to 200kN, forming a stable configuration with triangular support characteristics. This structure eliminates the stress concentration and elastic connection gaps of traditional welding, and optimizes and strengthens the force transmission path. Its torsional stiffness is increased by 40% compared with the traditional welded frame, and the rated axle load reaches 8t / axle. Under large span conditions, the bending deformation is ≤3mm. This rigid design makes the vehicle body structure and working platform extremely stable after the vehicle moves to the target position and stops. It can effectively suppress structural vibrations caused by load changes or slight start and stop of the motor, and provide an extremely stable physical platform for high-precision operations.

[0111] Meanwhile, the adaptive magnetic field orientation control unit, the hierarchical heterogeneous vehicle control unit architecture, and the rigid body structure together constitute a deeply collaborative hardware and software system, which can better solve the contradiction between speed and stability / economy in the dynamic operation of special vehicles by traditional control.

[0112] For the adaptive field-oriented control unit, the high rigidity and low hysteresis characteristics of mechanical transmission brought by the rigid connection body can be fully utilized. Since the nonlinearity and deformation uncertainty of the elastic element are eliminated, the prediction model of the adaptive field-oriented control unit is more accurate, and the control of wheel-end torque is more direct and rapid. This allows the energy-efficient control target to be achieved more accurately in the actual physical system. In addition to online rolling optimization and parameter adaptation, hardware characteristic synergy is also realized, further amplifying the effect of the algorithm in reducing current surge and improving range.

[0113] For the synchronous localization and mapping optimization unit, the rigid and low-vibration characteristics of the vehicle body provide cleaner and less interference-free raw data input for the visual and inertial sensors in the SLAM algorithm, which indirectly improves the positioning accuracy from a physical perspective.

[0114] In addition, a brake pump 4 is added at the connection point between each wheel hub and the independent motor 6 to achieve precise control of braking performance.

[0115] Reference Figure 3 In this embodiment, the control system 8 for the special operation intelligent vehicle is mounted on the front crossbeam 7, and is externally equipped with a first communication interface, a first CAN interface, a second CAN interface, a first GPS interface, a second GPS interface, an Ethernet interface, an I2C interface, a first power interface, a second power interface, an SPI interface, a USB interface, AD&IO, a digital spectrum modulation receiver interface, a second communication interface, a third communication interface, and a UART4&I2C interface.

[0116] In summary, the intelligent vehicle in this embodiment adopts a hardware and software collaboration based on a hierarchical heterogeneous architecture. By organically integrating the VCU architecture, differential suspension chassis, and cross-modal perception algorithm, a real-time, robust, and adaptive intelligent vehicle system is formed. Structurally, this system achieves synchronous optimization of the control and execution layers. The integration of the VCU achieves low cost, the rigid connection of the chassis ensures low vibration, and the algorithm achieves tight coupling of multimodal information, demonstrating significant synergistic gains at the system level.

[0117] It can achieve stable movement and precise operation in complex environments (such as mines, disaster areas, fields, and utility tunnels), reduce human intervention and operational risks, and improve the efficiency and safety level of special missions. Due to its high system integration, the number of hardware and wiring complexity are significantly reduced, thereby reducing manufacturing and maintenance costs, forming a cost-effective solution. At the same time, the low vibration characteristics of the rigid body directly improve the success rate and quality of precision operations when the vehicle is parked.

[0118] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A control system for a special-operation intelligent vehicle, characterized in that, include: Electrically connected integrated interface module, vehicle control unit and motor drive module The integrated interface module is equipped with a variety of hardware interfaces for connecting the functional modules and external devices of the special operation intelligent vehicle; The vehicle control unit includes an electrically connected real-time control kernel and a task coprocessor; The real-time control kernel is used to collect target data through the integrated interface module, and to perform real-time task analysis based on the target data to obtain real-time task analysis results. The task coprocessor is used to perform non-real-time task analysis based on the target data and obtain non-real-time task analysis results. The motor drive module is used to generate control commands and send them to the target motor based on the real-time task analysis results and the non-real-time task analysis results.

2. The control system for a special-operation intelligent vehicle as described in claim 1, characterized in that: The real-time control kernel also includes an adaptive magnetic field orientation control unit; The adaptive magnetic field orientation control unit is used to predict the future dynamics of the system within a preset finite time domain based on the target data and a preset prediction model. With the goal of minimizing the total energy consumption of the intelligent vehicle system, a set of optimal voltage vector sequences for characterizing the optimal torque output is solved online based on preset hard constraints and the future dynamics.

3. The control system for a special-operation intelligent vehicle as described in claim 2, characterized in that: The real-time control kernel is a heterogeneous platform combining an ARM Advanced Reduced Instruction Set Machine and an FPGA Field Programmable Gate Array.

4. The control system for a special-operation intelligent vehicle as described in claim 2, characterized in that: The adaptive magnetic field orientation control unit also includes an online parameter identification subunit, which is used to update the key parameters in the prediction model in real time based on the operating data of each motor of the intelligent vehicle.

5. The control system for a special-operation intelligent vehicle as described in claim 1, characterized in that: The task coprocessor also includes a cross-modal attention fusion perception unit; The cross-modal attention fusion perception unit is used to directly acquire two image data streams, visible light image data and infrared image data, through the real-time control kernel. Based on the data quality of the two image data streams, the feature weights of the two image data streams are dynamically adjusted in the feature fusion layer to obtain the dynamic feature distribution of the image data.

6. The control system for a special-operation intelligent vehicle as described in claim 5, characterized in that: The task coprocessor is a SoC (System-on-a-Chip) that integrates a CPU and a GPU or a CPU and an NPU (Neural Processing Unit). The task coprocessor also integrates multiple high-speed camera module interfaces, enabling high-bandwidth sensors used to acquire visible light or infrared image data to be directly connected to the mainboard of the vehicle control unit.

7. The control system for a special-operation intelligent vehicle as described in claim 1, characterized in that: The task coprocessor also includes a synchronous localization and mapping optimization unit; The synchronous positioning and mapping optimization unit is used to optimize the synchronous positioning and mapping results in a unified back-end optimizer by performing five-modal tight coupling of geometric, inertial, kinematic, semantic and ranging data according to preset constraint factors, so as to obtain the synchronous positioning and mapping optimization results.

8. The control system for a special-operation intelligent vehicle as described in claim 7, characterized in that, The preset constraint factors include: The common constraint factor for VIO visual inertial odometry and wheel odometry is to use the count of VIO visual inertial odometry to correct the heading drift of wheel odometry caused by wheel slippage, and to use the count of wheel odometry to provide absolute scale constraint for VIO visual inertial odometry. The geometric plane constraint factor specifically uses the environmental semantic information identified by the task coprocessor as a high-level prior to introduce geometric plane constraints in the graph optimization process. The one-dimensional ground height constraint factor is specifically introduced based on the precision ranging sensor data in the target data, with a constraint factor having a confidence level that reaches a preset value, in order to anchor the Z-axis coordinate and eliminate vertical cumulative drift.

9. The control system for a special-operation intelligent vehicle as described in any one of claims 1 to 8, characterized in that: The real-time control kernel and the task coprocessor are interconnected via a high-speed serial bus connected by PCIe peripheral components.

10. A smart vehicle, characterized in that, The system comprises a control system for a special-operation intelligent vehicle as described in any one of claims 1 to 9, and further includes: A pair of parallel longitudinal beams (1); The front crossbeam (7) and the rear crossbeam (3) are rigidly fixed at both ends of the front crossbeam (7) to the bottom of the front end of a pair of longitudinal beams (1), and the rear crossbeam (3) is rigidly fixed at both ends to the bottom of the rear end of a pair of longitudinal beams (1). The front crossbeam (7) and the rear crossbeam (3) are respectively connected to two independent wheel hubs (9) through a swing beam (2). The front crossbeam (7) and the swing beam (2), the rear crossbeam (3) and the swing beam (2) respectively form a gate-shaped load-bearing frame. The swing beam (2) is configured to swing up and down to achieve suspension adjustment of the two connected independent wheel hubs (9). Each of the independent wheel hubs (9) is connected to an independent motor (6), and each of the independent wheel hubs (9) is also provided with a wheel hub brake; The battery and power control module (5) is used to supply power to the control system (8) for the special operation intelligent vehicle, the independent motor (6) and the wheel hub brake.

Images