A carry equipment line controlled chassis multi-mode execution domain reconstruction system and a coordination control method thereof

Abstract

The application discloses a kind of carrying equipment drive-by-wire chassis multi-mode execution domain reconstruction system and its coordination control method, comprising: state perception unit, for collecting carrying equipment state data, and output state data set;Execution domain reconstruction unit is used to determine the target operation mode of carrying equipment, and output mode information, execution domain allocation result and control authority parameter;Coordination control unit is used to generate whole vehicle control target, and whole vehicle control target is decomposed into each corner module control instruction;Execution unit is used to transition processing and issue to each corner module actuator to the corner module control instruction that coordination control unit outputs.The application can improve the accuracy and foresight of operation mode identification by processing the environment information, working condition information, task information, whole vehicle motion state information and each corner module state information of carrying equipment, and combining the working condition evolution prediction in future preset interval for mode determination, so that drive-by-wire chassis changes from passive response to active adjustment.

Description

Technical Field

[0001] This invention belongs to the field of vehicle drive-by-wire chassis technology, specifically relating to a multi-mode execution domain reconfiguration system for drive-by-wire chassis of transport equipment and its coordinated control method. Background Technology

[0002] With the continuous improvement of the intelligence, electrification, and drive-by-wire capabilities of transportation equipment, traditional chassis systems relying on mechanical connections and fixed function allocation are gradually evolving into drive-by-wire chassis systems centered on sensors, controllers, and actuators. Drive-by-wire chassis achieve decoupled control of steering, driving, and braking functions through electrical signals, offering advantages such as flexible structural layout, high degree of control freedom, and ease of intelligent control. It has become an important development direction for transportation equipment for complex missions.

[0003] For multi-module vehicle equipment, the chassis actuators are numerous and widely distributed, with different modules typically possessing multiple capabilities such as steering, driving, and braking. During actual operation, factors such as gradient variations, adhesion differences, changes in traffic conditions, load transfer, and task mode switching can cause changes in the available capabilities and control requirements of each module at different times. Therefore, how to dynamically adjust the functional division, execution domain affiliation, and control authority of each module according to changes in operating conditions, and achieve coordinated allocation of vehicle objectives, has become a pressing issue in the field of drive-by-wire chassis control.

[0004] Most existing drive-by-wire chassis control methods still employ a fixed control architecture, pre-determining the functional divisions of the steering, drive, and braking domains, as well as the participation methods of each actuator, during the system design phase. While this approach can meet basic requirements under single operating conditions, it lacks adaptability in complex environments and under variable task conditions. When the capabilities of some corner modules decrease due to changes in adhesion conditions, load transfer, tire slippage, or actuator performance degradation, the fixed function allocation method cannot reflect the true usable capabilities of each corner module in a timely manner, failing to fully utilize the control potential of high-capacity corner modules. Furthermore, when the transport equipment needs to switch between different operating modes, existing methods lack a systematic consideration of the timing of mode switching, functional migration relationships, and changes in control authority, easily leading to unreasonable control allocation and discontinuous switching processes.

[0005] Furthermore, existing coordination and control methods often focus on target allocation or actuator control allocation based on the current state, typically employing fixed-weight allocation, single-layer optimized allocation, or local coordination methods. This makes it difficult to simultaneously consider the overall vehicle target requirements, execution domain-level collaboration, and fine-grained allocation at the corner module level. Especially during execution domain reconstruction, control authority reallocation, and functional role switching, the lack of effective transition mechanisms can lead to takeover shocks between control channels, impacting system stability and execution security.

[0006] Therefore, there is an urgent need to propose a multi-mode execution domain reconfiguration system and its coordinated control method for drive-by-wire chassis of transport equipment. This system should be able to predict and determine the operating mode, assess the available capabilities of each corner module, and dynamically reconfigure the execution domain affiliation, functional roles, and control permissions accordingly. At the same time, a unified coordination mechanism should be established in the stages of vehicle target generation, execution domain target decomposition, inter-domain coordinated allocation, corner module-level optimized allocation, and smooth transition of mode switching. This will improve the adaptive capability, control accuracy, resource utilization, and operational stability of the drive-by-wire chassis under complex working conditions. Summary of the Invention

[0007] To address the shortcomings of the existing technologies, the present invention aims to provide a multi-mode execution domain reconfiguration system for a drive-by-wire chassis of a transport vehicle and its coordinated control method, so as to solve the problems in the existing technology where fixed control architectures are unable to fully utilize the execution capabilities of corner modules under complex environments and variable task conditions, and are unable to take into account the hierarchical coordination of execution domains, fine allocation at the corner module level, and smooth transition of mode switching.

[0008] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0009] The present invention provides a multi-mode execution domain reconfiguration system for a drive-by-wire chassis of a transport vehicle, comprising: a state perception unit, an execution domain reconfiguration unit, a coordination control unit, and an execution unit;

[0010] The state perception unit is used to collect environmental information, operating condition information, task requirement information, vehicle motion state information and state information of each module of the transport equipment, and output a state dataset.

[0011] The execution domain reconstruction unit is used to determine the target operating mode of the carrier equipment based on the state dataset output by the state perception unit, and to reconstruct the functional division and control authority of each module according to the target operating mode, and output mode information, execution domain allocation results and control authority parameters.

[0012] The coordination and control unit is used to generate a vehicle control target based on the mode information output by the execution domain reconstruction unit, the execution domain allocation result, and the control authority parameters, and to decompose the vehicle control target into control instructions for each module.

[0013] The execution unit is used to process the corner module control commands output by the coordination control unit and send them to each corner module execution mechanism. At the same time, it collects execution feedback information and returns it to the state perception unit, the execution domain reconstruction unit, and the coordination control unit.

[0014] Furthermore, the state sensing unit includes: a data acquisition module and a data processing module;

[0015] The data acquisition module is used to classify and collect environmental conditions, operating conditions, vehicle motion conditions, and corner module conditions, and send the collected raw data to the data processing module. The data acquisition module includes: an environmental information acquisition submodule, an operating condition information acquisition submodule, a vehicle condition acquisition submodule, and a corner module condition acquisition submodule.

[0016] The environmental information acquisition submodule is used to collect information on ambient temperature, humidity, wind speed, slope, and terrain undulation.

[0017] The working condition information acquisition submodule is used to collect road surface adhesion characteristics, traffic condition characteristics, and task requirement information;

[0018] The vehicle status acquisition submodule is used to collect vehicle speed, longitudinal acceleration, lateral acceleration, yaw rate, pitch angle, roll angle and pose information;

[0019] The corner module status acquisition submodule is used to collect information on the steering angle, steering angular velocity, driving torque, driving speed, braking pressure, and actuator operating status of each corner module.

[0020] The data processing module is used to perform unified preprocessing and feature organization on the raw data sent by the data acquisition module, eliminate the differences in sampling time, unit scale and noise level between different data sources, form a state dataset that can be directly called for pattern determination, capability assessment and target generation and send it to the execution domain reconstruction unit.

[0021] The preprocessing includes time alignment, filtering, outlier removal, and normalization.

[0022] The state dataset is constructed according to environmental features, operating condition features, vehicle state features, and corner module state features.

[0023] Furthermore, the execution domain reconstruction unit includes: a mode determination module, a capability assessment module, and a reconstruction management module;

[0024] The mode determination module is used to perform long-term domain prediction and intelligent decision-making on the evolution of operating conditions within a preset prediction interval based on the state dataset output by the data processing module, determine the target operating mode, and send the target operating mode to the target generation module in the reconfiguration management module and the coordination control unit; the target operating mode is determined based on environmental characteristics, operating condition characteristics, vehicle attitude characteristics, and task requirement characteristics.

[0025] The capability assessment module is used to quantitatively assess the available control capabilities of each corner module under the current operating conditions and target operating mode constraints based on the status dataset output by the data processing module, forming a corner module capability parameter set, and sending the corner module capability parameter set to the control allocation module in the reconfiguration management module and the coordination control unit; the corner module capability parameter set includes steering adjustment capability parameters, drive output capability parameters, braking adjustment capability parameters and allocable weight parameters for each corner module;

[0026] The reconstruction management module is used to reconstruct the execution domain affiliation, functional role, and control authority of each corner module in the current mode according to the target operating mode output by the mode determination module and the corner module capability parameter set output by the capability assessment module, forming execution domain allocation results and control authority parameters, and sending the execution domain allocation results and control authority parameters to the coordination and control unit; the execution domain allocation results represent the execution domain to which each corner module belongs and its functional role, and the control authority parameters represent the control authority and constraint boundaries of each corner module in the target operating mode.

[0027] Furthermore, the coordination and control unit includes: a target generation module, a target decomposition module, and a control allocation module;

[0028] The target generation module is used to generate a vehicle control target that matches the current operating requirements based on the target operating mode output by the mode determination module, the status dataset output by the data processing module, and the task requirement information, and then send the vehicle control target to the target decomposition module; the vehicle control target includes one or more of the following: trajectory tracking target, attitude stabilization target, passability target, and motion response target;

[0029] The target decomposition module is used to decompose the vehicle control target into execution domain control targets based on the vehicle control target output by the target generation module and the execution domain allocation result output by the reconfiguration management module, and send the execution domain control targets to the control allocation module; the execution domain control targets represent the steering adjustment task, drive allocation task and braking allocation task undertaken by the corresponding execution domain.

[0030] The control allocation module is used to calculate and allocate the control tasks of each corner module according to the control targets of each execution domain, the corner module capability parameter set, and the control permission parameters, to form corner module control instructions, and send the corner module control instructions to the transition processing module in the execution unit; the corner module control instructions include the target steering angle, target driving torque, and target braking force corresponding to each corner module.

[0031] Furthermore, the execution unit includes: a transition processing module, an instruction issuing module, and a feedback module;

[0032] The transition processing module is used to continuously process the corner module control commands output by the control allocation module during the operation mode switching, execution domain adjustment, or corner module function migration, and send the processed control commands to the command issuing module; the continuous processing is used to make the target steering angle, target driving torque, and target braking force change continuously before and after reconstruction according to a preset transition law;

[0033] The instruction issuing module is used to receive control instructions output by the transition processing module, and send the control instructions to the steering actuator, drive actuator and braking actuator corresponding to each corner module respectively, so as to drive each corner module to perform corresponding control actions; the control instructions include target steering angle instruction, target drive torque instruction and target braking force instruction;

[0034] The feedback module is used to collect the actual execution results of each corner module and the vehicle motion response, and return the feedback results to the data processing module, the mode determination module and the reconstruction management module to update the status dataset, the current operating mode and the execution domain allocation results; the feedback results include the actual steering response, actual driving response, actual braking response and vehicle motion response of each corner module.

[0035] The present invention provides a coordinated control method for a multi-mode execution domain reconfiguration system of a drive-by-wire chassis for transport equipment. Based on the aforementioned system, the steps are as follows:

[0036] 1) Collect environmental information, operating condition information, mission requirement information, vehicle motion status information, and status information of each module of the transport equipment, and form a status dataset after unifying the time base, denoising correction and feature extraction;

[0037] 2) Based on the state dataset, perform long-term time-domain prediction and intelligent decision-making on the evolution of working conditions within the future preset prediction interval, and determine the target operating mode and the timing of mode switching.

[0038] 3) Based on the state dataset, the available controllability of each corner module under the constraints of the current working condition and the target operating mode is quantitatively evaluated, and an actuator effectiveness model is established to form a corner module capability parameter set;

[0039] 4) Based on the target operating mode and mode switching timing in step 2) and the corner module capability parameter set in step 3), the execution domain affiliation, functional role and control authority of each corner module in the target operating mode are reconstructed to form the execution domain allocation result and control authority parameters;

[0040] 5) Based on the target operation mode in step 2), the state dataset in step 1), and the task requirement information, generate a vehicle control target that matches the requirements of the target operation mode, and perform executability verification and constraint correction on the vehicle control target in conjunction with the corner module capability parameter set in step 3); and decompose the vehicle control target into control targets for each execution domain according to the corrected vehicle control target and the execution domain allocation results in step 4).

[0041] 6) Based on the corner module capability parameter set in step 3), the control authority parameters in step 4), and the control objectives of each execution domain in step 5), construct a hierarchical coordination control allocation model, perform inter-domain coordination allocation and corner module-level optimization allocation of the control tasks of each corner module, and form corner module control instructions;

[0042] 7) Based on the corner module control instructions in step 6), combined with the mode switching timing in step 2) and the execution domain allocation results and control authority parameters in step 4), a transition takeover process is performed on the execution domain affiliation and control authority change process before and after reconstruction. The processed control instructions are sent to the steering actuator, drive actuator and braking actuator corresponding to each corner module. At the same time, according to the actual execution results of each corner module and the vehicle motion response, the state dataset in step 1), the target operating mode and mode switching timing in step 2), the corner module capability parameter set in step 3), and the execution domain allocation results in step 4) are updated to achieve closed-loop coordinated control under multi-mode execution domain reconstruction.

[0043] Furthermore, the steps for forming the state dataset in step 1) are as follows:

[0044] 11) Collect environmental information, operating condition information, mission information, vehicle motion status information, and corner module status information of the transport equipment to form a multi-source heterogeneous original state sequence;

[0045] 12) Perform unified time reference, filtering and denoising, and outlier and missing value correction on the multi-source heterogeneous original state sequence in step 11) to form an effective state sequence;

[0046] 13) Standardize the valid state sequence formed in step 12) and extract features using a sliding time window to form a state dataset.

[0047] Furthermore, the specific steps of step 2) are as follows:

[0048] 21) Based on the state dataset in step 1), establish a knowledge base for the operation mode of the launch vehicle, including the applicable working conditions, task requirements, performance target weights and execution constraint boundaries corresponding to each operation mode;

[0049] 22) A multimodal temporal prediction network based on spatiotemporal attention mechanism is adopted to jointly encode environmental features, working condition features, task requirement features, vehicle motion features and corner module state features and extrapolate long-term trends to obtain the working condition evolution sequence within the future preset prediction interval.

[0050] 23) Using a pattern matching method based on contrastive representation learning, the working condition evolution sequence obtained in step 22) is matched with the feature prototypes of each pattern in the operation mode knowledge base to generate a set of candidate operation modes;

[0051] 24) Using a comprehensive evaluation method based on the mode benefit prediction network and the switching cost assessment model, the task completion benefit, stability benefit, energy consumption benefit and mode switching cost of the candidate operating mode set obtained in step 23) are jointly predicted within the prediction interval to form a comprehensive benefit ranking result for each operating mode.

[0052] 25) Using a rolling time-domain optimization-based mode decision-making method, the comprehensive benefit ranking results obtained in step 24) are used to make time-series decisions to determine the target operating mode and the timing of mode switching.

[0053] Furthermore, the specific steps of step 3) are as follows:

[0054] 31) Based on the state dataset formed in step 1), extract the road surface adhesion conditions, wheel end normal load, tire slip state, and driving, braking, and steering execution boundary information to construct the wheel end constraint state set of each corner module;

[0055] 32) Based on the tire force constraint analysis and friction circle correction method, the wheel end constraint state set obtained in step 31) is processed to solve the available boundaries of the longitudinal and lateral forces at the wheel end of each corner module under the current working condition and to perform mode constraint correction, so as to obtain the steering execution capability, driving execution capability and braking execution capability of each corner module under the target operating mode.

[0056] 33) Based on the generalized force feasible domain construction and control efficiency matrix modeling method, the execution capabilities of each corner module obtained in step 32) are mapped to the actuator effectiveness model, generating the corner module capability parameter set and channel weight parameters.

[0057] Furthermore, step 4) specifically involves the following steps:

[0058] 41) Based on the target operating mode and mode switching timing determined in step 2) and the corner module capability parameter set formed in step 3), construct the mode demand-capability supply coupling relationship to form the set of functions that each corner module can undertake in different execution domains and the candidate set of execution domains;

[0059] 42) An execution domain reconstruction method based on pattern-capability coupling graph matching and priority constraint optimization is adopted to make reconstruction decisions on the set of achievable functions and the candidate set of execution domains formed in step 41), and to determine the execution domain affiliation of each module in the target operating mode, as well as the main functional role, auxiliary functional role and support functional role.

[0060] 43) Based on the execution domain affiliation and functional role determined in step 42), classify the participation level, control sequence and takeover conditions of each module in the steering, driving and braking execution domains to form control authority parameters;

[0061] 44) Output the execution domain allocation results and control permission parameters.

[0062] Furthermore, the specific steps of step 5) are as follows:

[0063] 51) Based on the state dataset formed in step 1), the target operation mode determined in step 2), and the task requirement information, generate a vehicle control target that matches the operation requirements of the future preset prediction interval.

[0064] 52) Combining the vehicle generalized force feasible boundary corresponding to the corner module capability parameter set formed in step 3), perform executability verification and constraint correction on the vehicle control target generated in step 51) to obtain the executable vehicle control target under the target operation mode;

[0065] 53) Based on the execution domain allocation results and control authority parameters formed in step 4), establish the mapping relationship between the vehicle control target and the control targets of each execution domain, and decompose the executable vehicle control target obtained in step 52) into the control targets of each execution domain, and output the control targets of each execution domain and their priority information.

[0066] Furthermore, the specific steps for controlling the allocation in step 6) are as follows:

[0067] 61) Based on the corner module capability parameter set in step 3), the control authority parameters in step 4), and the control objectives of each execution domain obtained in step 5), construct the control efficiency relationship of each execution domain and the control efficiency relationship of each corner module to form a hierarchical coordination control allocation model;

[0068] 62) Adopt a role-permission constraint-based inter-domain coordination and allocation method to allocate responsibility and contribution ratios for the control objectives of each execution domain, and determine the target responsibilities and coordination weights of each execution domain at the current moment;

[0069] 63) Using a corner module-level coordination allocation method based on control efficiency matrix, consistency constraints and weighted quadratic optimization, the target load of each execution domain determined in step 62) at the current time is solved at the corner module level, and the steering control quantity, drive control quantity and braking control quantity of each corner module are calculated to form corner module control commands.

[0070] Furthermore, step 7) specifically involves the following steps:

[0071] 71) Based on the corner module control instructions in step 6), the execution domain allocation results in step 4), and the mode switching timing in step 2), construct the transition takeover relationship corresponding to the control channel participation relationship and control authority change relationship before and after reconstruction;

[0072] 72) Based on the permission gradual change and instruction continuous processing method, the transition takeover relationship constructed in step 71) is smoothed to form continuous control instructions, which are then sent to the steering actuator, drive actuator and braking actuator corresponding to each corner module.

[0073] 73) Based on the actual execution results of each corner module and the vehicle motion response, update the state dataset in step 1), the target operating mode and mode switching timing in step 2), the corner module capability parameter set in step 3), and the execution domain allocation results in step 4), thereby realizing closed-loop coordinated control under multi-mode execution domain reconstruction.

[0074] The beneficial effects of this invention are:

[0075] This invention improves the accuracy and foresight of operating mode recognition by uniformly sensing and processing environmental information, operating condition information, mission information, vehicle motion status information and status information of each module of the transport equipment, and combining the prediction of operating condition evolution within a preset range in the future to determine the mode, so as to transform the drive-by-wire chassis from passive response to active adjustment.

[0076] This invention evaluates the steering, driving, and braking capabilities of each module and dynamically reconstructs the execution domain affiliation, functional roles, and control permissions according to the target operating mode. This breaks through the limitations of traditional fixed control architecture, improves the utilization of execution resources, and enhances the system's adaptability to complex operating conditions.

[0077] This invention constructs a hierarchical coordinated control allocation mechanism for vehicle targets, execution domain targets, and corner module instructions. This mechanism can meet the vehicle motion control requirements while taking into account the differences in capabilities of each execution domain and the constraints of corner modules, thereby improving the rationality, coordination, and executability of control allocation.

[0078] This invention reduces control abrupt changes and switching shocks by smoothly transitioning control quantities during mode switching and function migration, and by performing closed-loop updates based on execution results, thereby improving the control continuity, operational stability, and engineering application value of the drive-by-wire chassis system. Attached Figure Description

[0079] Figure 1 This is a schematic diagram of the system of the present invention.

[0080] Figure 2 This is a schematic diagram of the method of the present invention. Detailed Implementation

[0081] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to embodiments and accompanying drawings. The content mentioned in the embodiments is not intended to limit the present invention.

[0082] Reference Figure 1 As shown, a multi-mode execution domain reconfiguration system for a drive-by-wire chassis of a transport vehicle according to the present invention includes: a state perception unit, an execution domain reconfiguration unit, a coordination control unit, and an execution unit;

[0083] The state perception unit is used to collect environmental information, operating condition information, task requirement information, vehicle motion state information and state information of each module of the transport equipment, and output a state dataset.

[0084] The state perception unit includes a data acquisition module and a data processing module.

[0085] The data acquisition module is used to classify and collect environmental conditions, operating conditions, vehicle motion conditions, and corner module conditions, and send the collected raw data to the data processing module. The data acquisition module includes: an environmental information acquisition submodule, an operating condition information acquisition submodule, a vehicle condition acquisition submodule, and a corner module condition acquisition submodule.

[0086] The environmental information acquisition submodule is used to collect information on ambient temperature, humidity, wind speed, slope, and terrain undulation.

[0087] The working condition information acquisition submodule is used to collect road surface adhesion characteristics, traffic condition characteristics, and task requirement information;

[0088] The vehicle status acquisition submodule is used to collect vehicle speed, longitudinal acceleration, lateral acceleration, yaw rate, pitch angle, roll angle and pose information;

[0089] The corner module status acquisition submodule is used to collect information on the steering angle, steering angular velocity, driving torque, driving speed, braking pressure, and actuator operating status of each corner module.

[0090] The data processing module is used to perform unified preprocessing and feature organization on the raw data sent by the data acquisition module, eliminate the differences in sampling time, scale and noise level between different data sources, form a state dataset that can be directly called for pattern determination, capability assessment and target generation and send it to the execution domain reconstruction unit.

[0091] The preprocessing includes time alignment, filtering, outlier removal, and normalization.

[0092] The state dataset is constructed according to environmental features, operating condition features, vehicle state features, and corner module state features.

[0093] The execution domain reconstruction unit is used to determine the target operation mode of the carrier equipment based on the state dataset output by the state perception unit, and to reconstruct the functional division and control authority of each module according to the target operation mode, and output mode information, execution domain allocation results and control authority parameters.

[0094] The coordination and control unit is used to generate a vehicle control target based on the mode information output by the execution domain reconstruction unit, the execution domain allocation result, and the control authority parameters, and to decompose the vehicle control target into control instructions for each module.

[0095] The execution unit is used to process the corner module control commands output by the coordination control unit and send them to each corner module execution mechanism. At the same time, it collects execution feedback information and returns it to the state perception unit, the execution domain reconstruction unit, and the coordination control unit.

[0096] The execution domain reconstruction unit includes: a mode determination module, a capability assessment module, and a reconstruction management module.

[0097] The mode determination module is used to perform long-term domain prediction and intelligent decision-making on the evolution of operating conditions within a preset prediction interval based on the state dataset output by the data processing module, determine the target operating mode, and send the target operating mode to the target generation module in the reconfiguration management module and the coordination control unit; the target operating mode is determined based on environmental characteristics, operating condition characteristics, vehicle attitude characteristics, and task requirement characteristics.

[0098] The capability assessment module is used to quantitatively assess the available control capabilities of each corner module under the current operating conditions and target operating mode constraints based on the status dataset output by the data processing module, forming a corner module capability parameter set, and sending the corner module capability parameter set to the control allocation module in the reconfiguration management module and the coordination control unit; the corner module capability parameter set includes steering adjustment capability parameters, drive output capability parameters, braking adjustment capability parameters and allocable weight parameters for each corner module;

[0099] The reconstruction management module is used to reconstruct the execution domain affiliation, functional role, and control authority of each corner module in the current mode according to the target operating mode output by the mode determination module and the corner module capability parameter set output by the capability assessment module, forming execution domain allocation results and control authority parameters, and sending the execution domain allocation results and control authority parameters to the coordination and control unit; the execution domain allocation results represent the execution domain to which each corner module belongs and its functional role, and the control authority parameters represent the control authority and constraint boundaries of each corner module in the target operating mode.

[0100] The coordination and control unit includes: a target generation module, a target decomposition module, and a control allocation module;

[0101] The target generation module is used to generate a vehicle control target that matches the current operating requirements based on the target operating mode output by the mode determination module, the status dataset output by the data processing module, and the task requirement information, and then send the vehicle control target to the target decomposition module; the vehicle control target includes one or more of the following: trajectory tracking target, attitude stabilization target, passability target, and motion response target;

[0102] The target decomposition module is used to decompose the vehicle control target into execution domain control targets based on the vehicle control target output by the target generation module and the execution domain allocation result output by the reconfiguration management module, and send the execution domain control targets to the control allocation module; the execution domain control targets represent the steering adjustment task, drive allocation task and braking allocation task undertaken by the corresponding execution domain.

[0103] The control allocation module is used to calculate and allocate the control tasks of each corner module according to the control targets of each execution domain, the corner module capability parameter set, and the control permission parameters, to form corner module control instructions, and send the corner module control instructions to the transition processing module in the execution unit; the corner module control instructions include the target steering angle, target driving torque, and target braking force corresponding to each corner module.

[0104] The execution unit includes: a transition processing module, an instruction issuing module, and a feedback module;

[0105] The transition processing module is used to continuously process the corner module control commands output by the control allocation module during the operation mode switching, execution domain adjustment, or corner module function migration, and send the processed control commands to the command issuing module; the continuous processing is used to make the target steering angle, target driving torque, and target braking force change continuously before and after reconstruction according to a preset transition law;

[0106] The instruction issuing module is used to receive control instructions output by the transition processing module, and send the control instructions to the steering actuator, drive actuator and braking actuator corresponding to each corner module respectively, so as to drive each corner module to perform corresponding control actions; the control instructions include target steering angle instruction, target drive torque instruction and target braking force instruction;

[0107] The feedback module is used to collect the actual execution results of each corner module and the vehicle motion response, and return the feedback results to the data processing module, the mode determination module and the reconstruction management module to update the status dataset, the current operating mode and the execution domain allocation results; the feedback results include the actual steering response, actual driving response, actual braking response and vehicle motion response of each corner module.

[0108] Reference Figure 2 As shown, the coordinated control method of a multi-mode execution domain reconfiguration system for a drive-by-wire chassis of a transport equipment according to the present invention, based on the above system, comprises the following steps:

[0109] 1) Collect environmental information, operating condition information, mission requirement information, vehicle motion status information, and status information of each module of the transport equipment, and form a status dataset after unifying the time base, denoising correction and feature extraction;

[0110] The steps for forming the state dataset are as follows:

[0111] 11) Collect environmental information, operating condition information, mission information, vehicle motion status information, and corner module status information of the transport equipment to form a multi-source heterogeneous original state sequence;

[0112] 12) Perform unified time reference, filtering and denoising, and outlier and missing value correction on the multi-source heterogeneous original state sequence in step 11) to form an effective state sequence;

[0113] 13) Standardize the valid state sequence formed in step 12) and extract features using a sliding time window to form a state dataset.

[0114] Specifically, the formation of the multi-source heterogeneous original state sequence is as follows:

[0115] The collected environmental information vector Operating condition information vector Task information vector Vehicle motion state information vector and the state information vectors of each corner module According to the unified discrete time The integrated sequence of multi-source heterogeneous original states is represented as:

[0116] ;

[0117] In the formula, To unify discrete time The following is a sequence of multi-source heterogeneous primitive states; superscript Indicates the transpose operation;

[0118] The state information vector of the corner module is represented as follows:

[0119] ;

[0120] In the formula, the first Each corner module status information vector Represented as:

[0121] ;

[0122] In the formula, To achieve a unified discrete time Next Each corner module's steering angle, To achieve a unified discrete time The angular velocity of the lower wheel, To achieve a unified discrete time Lower longitudinal slip ratio, To achieve a unified discrete time Lower tire slip angle, To achieve a unified discrete time Lower wheel end normal load, and At the unified discrete time respectively Downward driving available torque and braking available torque, This represents the number of corner modules.

[0123] The formation of the effective state sequence in step 12) is specifically as follows:

[0124] The multi-source heterogeneous original state sequence formed in step 11) is subjected to unified time reference, filtering and denoising, and outlier and missing value correction processing to obtain the unified discrete time. Next effective state sequence , denoted as:

[0125] ;

[0126] In the formula, This represents a state preprocessing operator used to map raw state variables with different sampling frequencies, communication delays, and quality levels to effective state variables under a unified time base.

[0127] Based on the effective state vector Construct the first The corner module at the unified discrete time Normalized capability-related state vector (To highlight the differences in the capabilities of diagonal modules in multi-mode execution domain reconstruction and the sensitivity to wheel-end state differentiation), as follows:

[0128] ;

[0129] In the formula, and At the unified discrete time respectively Normalized longitudinal slip ratio and tire slip angle and These are the reference maximum values ​​for the driving torque and braking torque, respectively.

[0130] Defined at the unified discrete time Heterogeneity index among the lower corner modules for:

[0131] ;

[0132] In the formula, To achieve a unified discrete time The average value of the state vectors related to the normalization capability of all corner modules; The second norm of a vector; Used to characterize at a unified discrete time. The degree of dispersion of each corner module in terms of attachment utilization, lateral state, vertical load distribution, and drive / brake availability; the heterogeneity index serves as an important criterion for subsequent operation mode prediction, execution domain reconstruction, and control authority adjustment.

[0133] The standardization process and sliding time window feature extraction in step 13) are as follows:

[0134] For the effective state vector formed in step 12) After standardization, in length of The mean state and the state evolution trend are extracted within the sliding time window, and are expressed as follows:

[0135] ;

[0136] ;

[0137] In the formula, Discrete moments within a sliding time window The corresponding valid state vector; This is the effective state vector corresponding to the initial discrete moment of the sliding time window; The length of the sliding time window; Index of discrete moments within a sliding time window; Characterization at unified discrete time The average state level within the next time window. Characterization at unified discrete time The state evolution trend within the next time window To standardize the sampling period;

[0138] By combining the current effective state, time window statistical characteristics, and corner module heterogeneity index, a unified discrete time is constructed. Feature vector at the time of order placement as follows:

[0139] ;

[0140] In the formula, They are respectively , and The transpose of;

[0141] Stacking the feature vectors from consecutive time steps constructs a unified discrete time step. Lower State Dataset as follows:

[0142] ;

[0143] In the formula, The starting discrete time of the time window for the state dataset; The next discrete moment after the start time within the time window; These represent the single-time feature vectors at the corresponding discrete time points. The length of the state dataset is given; the resulting state dataset not only represents the current operating condition and the overall vehicle status, but also the capability differences between corner modules and their temporal evolution characteristics.

[0144] 2) Based on the state dataset, perform long-term time-domain prediction and intelligent decision-making for the evolution of operating conditions within a preset prediction interval to determine the target operating mode and the timing of mode switching; the specific steps are as follows:

[0145] 21) Based on the state dataset in step 1), establish a knowledge base for the operation mode of the launch vehicle, including the applicable working conditions, task requirements, performance target weights and execution constraint boundaries corresponding to each operation mode;

[0146] 22) A multimodal temporal prediction network based on spatiotemporal attention mechanism is adopted to jointly encode environmental features, working condition features, task requirement features, vehicle motion features and corner module state features and extrapolate long-term trends to obtain the working condition evolution sequence within the future preset prediction interval.

[0147] 23) Using a pattern matching method based on contrastive representation learning, the working condition evolution sequence obtained in step 22) is matched with the feature prototypes of each pattern in the operation mode knowledge base to generate a set of candidate operation modes;

[0148] 24) Using a comprehensive evaluation method based on the mode benefit prediction network and the switching cost assessment model, the task completion benefit, stability benefit, energy consumption benefit and mode switching cost of the candidate operating mode set obtained in step 23) are jointly predicted within the prediction interval to form a comprehensive benefit ranking result for each operating mode.

[0149] 25) Using a rolling time-domain optimization-based mode decision-making method, the comprehensive benefit ranking results obtained in step 24) are used to make time-series decisions to determine the target operating mode and the timing of mode switching.

[0150] The specific steps for establishing the knowledge base of the operation mode of the transport equipment in step 21) are as follows:

[0151] Based on the state dataset in step 1) Establish an operation mode knowledge base according to the mode category of historical operation samples. , denoted as:

[0152] ;

[0153] Among them, the The mode entries corresponding to the operating modes Represented as:

[0154] ;

[0155] In the formula, For the first The characteristic prototype of this operating mode For the first The applicable operating conditions range corresponding to each operating mode For the performance target weight vector, To execute the constraint boundary parameter set, Total number of operating modes;

[0156] (To ensure that the predicted features in step 22) and the pattern matching features in step 23) are located in a unified representation space, the state dataset is... Adopting a unified encoding mapping Extracting deep features, then the first The characteristic prototype of this operating mode is defined as follows:

[0157] ;

[0158] In the formula, For belonging to the first A historical sample set of various operating modes For the sample size, For the first A historical state dataset sample, For a unified feature encoder;

[0159] The performance target weight vector is represented as follows:

[0160] ;

[0161] In the formula, , and They represent the first The weighting coefficients of task completion benefits, stability benefits, and energy consumption benefits under different operating modes.

[0162] Specifically, in step 22), the multimodal temporal prediction network based on the spatiotemporal attention mechanism performs joint encoding and long-term trend extrapolation as follows:

[0163] The state dataset obtained in step 1) Input a multimodal temporal prediction network based on a spatiotemporal attention mechanism, and classify the modal features into five categories according to the information source: environmental features, operating condition features, task requirement features, vehicle motion features, and corner module features; for the 1st Modal features, in length of The temporal attention weights are constructed in the time domain as follows:

[0164] ;

[0165] Therefore, we obtain the first... Temporal aggregation features of modal features:

[0166] ;

[0167] In the formula, To achieve a unified discrete time Next Modal features in the first Attention weights at each time location To achieve a unified discrete time The corresponding time aggregation features are as follows. , , and Here, j represents the network parameters; j is the discrete-time position index within the time window of the state dataset.

[0168] To reflect the different contributions of different modal information to the evolution of operating modes, spatial attention weights are introduced among the five types of modal features, as follows:

[0169] ;

[0170] In the formula, To achieve a unified discrete time Spatial attention weights for the u-th modal feature; u and j are both modal feature category indices; , and These are the parameters of the spatial attention network;

[0171] Furthermore, constructing at a unified discrete time... Fusion state characterization as follows:

[0172] ;

[0173] In the formula, To achieve a unified discrete time Next Spatial attention weights for modality features To achieve a unified discrete time The fused global representation vector; due to the feature vector in step 1). Corner module heterogeneity index already included Therefore, the fusion state characterization Simultaneously, it characterizes environmental evolution trends, vehicle motion trends, and execution domain reconstruction sensitive information;

[0174] Based on fusion state representation Using a time series prediction network to predict future intervals The internal working condition evolution is recursively predicted in multiple steps to obtain the hidden state prediction results and working condition evolution prediction vectors for each future prediction step, as expressed as:

[0175] ;

[0176] In the formula, To achieve a unified discrete time The next chapter The hidden state prediction results of the step, To achieve a unified discrete time The next chapter Step-by-step working condition evolution prediction vector, For time series prediction networks, and Output mapping parameters;

[0177] Stacking the load condition evolution vectors within the future prediction interval yields the following load condition evolution sequence:

[0178] ;

[0179] In the formula, Characterization at unified discrete time This study will predict the evolution trends of the environment, tasks, vehicle motion, and angular module heterogeneity in the near future, providing a basis for subsequent mode matching and switching decisions.

[0180] The pattern matching method based on contrastive representation learning in step 23) is as follows:

[0181] Applying a mapping function to the predicted interval sequence Extracting trend representation vectors The working condition evolution sequence obtained in step 22) is used to... Compared with the operation mode knowledge base established in step 21) Placed within a unified discrimination space, as follows:

[0182] ;

[0183] In the formula, To achieve a unified discrete time The trend representation vector corresponding to the current predicted working condition; This is a mapping function used to extract the trend representation of the prediction interval;

[0184] Trend representation vector Prototypes of various operating modes Perform similarity matching as follows:

[0185] ;

[0186] In the formula, To achieve a unified discrete time The current forecast operating conditions and the first The similarity between the prototypes of the various operating modes;

[0187] Similarity constraints and applicable working conditions range When used in combination, the following set of candidate operating modes is constructed:

[0188] ;

[0189] In the formula, For similarity threshold, To achieve a unified discrete time The set of candidate operating modes is used to map future predicted operating conditions from a continuous evolution sequence to a discrete set of candidate modes, providing input for subsequent comprehensive benefit ranking.

[0190] Specifically, step 24) is as follows:

[0191] For each satisfaction obtained in step 23) The The operating mode is based on the operating condition evolution sequence obtained in step 22). The pattern feature prototype established in step 21) and execution constraint boundary parameter set The task completion revenue, stability revenue, and energy consumption revenue of the operating mode within the prediction interval are estimated using a model revenue prediction network as follows:

[0192] ;

[0193] In the formula, The first in the candidate running mode set Various operating modes at a unified discrete time Next for the future The pattern profit prediction vector for each step; , and They represent the unified discrete time. Next This type of operation mode in the future The predicted values ​​of task completion benefits, stability benefits, and energy consumption benefits for each step. For model revenue prediction network;

[0194] Discount factor for each step of the forecast interval Accumulation yields the result at the unified discrete time. Next Cumulative comprehensive benefits of various operating modes for:

[0195] ;

[0196] In the formula, For the prediction interval;

[0197] Considering the changes in execution domain ownership, control authority redistribution, and execution channel takeover transition caused by mode switching, a switching cost evaluation model is constructed to suppress frequent and high-impact switching; as follows:

[0198] ;

[0199] In the formula, To achieve a unified discrete time Next The operating mode at the current moment The cost of switching modes; To achieve a unified discrete time Current operating mode The corresponding pattern feature prototype; To achieve a unified discrete time The predicted future Step angle module heterogeneity index; For the prediction step index; To achieve a unified discrete time The current operating mode is set. For indicator functions, To achieve a unified discrete time The average value of the heterogeneity of the corner modules within the prediction interval. , and For cost weighting coefficients; in the above switching cost evaluation model, the first term represents whether a mode switch occurs, the second term represents the span between the current mode and the candidate mode in the feature space, and the third term represents the reconstruction transition complexity caused by the enhanced heterogeneity of corner modules.

[0200] Therefore, at the unified discrete time Next Net comprehensive income of the operating mode Represented as:

[0201] ;

[0202] For the set of candidate operating modes Net comprehensive income of each model Sort the results in descending order to obtain the overall benefit ranking of the operating modes.

[0203] Specifically, step 25) is as follows:

[0204] At each decision point, the net comprehensive return of the candidate patterns obtained in step 24) is used. As a phase evaluation indicator, within the prediction interval The internal model sequence is optimized by rolling optimization, and the model time series decision model is established as follows:

[0205] ;

[0206] In the formula, It represents the set of feasible pattern sequences consisting of the set of candidate operating patterns within the prediction interval; This is a discount factor for returns; To achieve a unified discrete time Next Each prediction step corresponds to an operating mode. Net comprehensive income; This is the switching suppression coefficient; and The first The operating mode corresponding to each prediction step and the previous prediction step; For the prediction step index; To achieve a unified discrete time The optimal pattern sequence within the next prediction interval. For the first The operating mode corresponding to each prediction step To switch the suppression coefficient; the rolling time-domain optimization satisfies the following constraints:

[0207] ;

[0208] In the formula, To achieve a unified discrete time The next chapter The set of candidate operating modes corresponding to each prediction step; To achieve a unified discrete time The next chapter The predicted operating condition evolution results for each prediction step; For operating mode Corresponding applicable operating conditions range; For pattern Duration, The minimum allowed dwell time for this model is used to avoid frequent jittering near the critical boundary; the first term of the optimal model sequence is taken as the time at a unified discrete moment. The current target operating mode as follows:

[0209] ;

[0210] If the target operating mode With the current operating mode If there is a discrepancy, the mode switching timing is determined by the moment when the first mode change occurs in the optimal mode sequence, as follows:

[0211] ;

[0212] The corresponding mode switching times are:

[0213] ;

[0214] In the formula, To achieve a unified discrete time The number of prediction steps in the optimal pattern sequence when the first pattern change occurs; For the prediction step index; To achieve a unified discrete time The first optimal pattern sequence The operating mode corresponding to each prediction step; To achieve a unified discrete time The current operating mode is set; To achieve a unified discrete time The corresponding mode switching time; To standardize the sampling period;

[0215] And output the target running mode and mode switching timing .

[0216] 3) Based on the state dataset, the available controllability of each corner module under the constraints of the current operating condition and the target operating mode is quantitatively evaluated, and an actuator effectiveness model is established to form a corner module capability parameter set; the specific steps are as follows:

[0217] 31) Based on the state dataset formed in step 1), extract the road surface adhesion conditions, wheel end normal load, tire slip state, and driving, braking, and steering execution boundary information to construct the wheel end constraint state set of each corner module;

[0218] 32) Based on the tire force constraint analysis and friction circle correction method, the wheel end constraint state set obtained in step 31) is processed to solve the available boundaries of the longitudinal and lateral forces at the wheel end of each corner module under the current working condition and to perform mode constraint correction, so as to obtain the steering execution capability, driving execution capability and braking execution capability of each corner module under the target operating mode.

[0219] 33) Based on the generalized force feasible domain construction and control efficiency matrix modeling method, the execution capabilities of each corner module obtained in step 32) are mapped to the actuator effectiveness model, generating the corner module capability parameter set and channel weight parameters.

[0220] Specifically, step 31) is as follows:

[0221] Based on the state dataset formed in step 1) Extract the road surface adhesion conditions, wheel-end normal loads, tire slip states, and driving, braking, and steering execution boundary information of each corner module to construct the wheel-end constraint state set of the corner module. as follows:

[0222] ;

[0223] In the formula, For the first Each module at a unified discrete time The wheel end constraint vector below; To achieve a unified discrete time Next Each corner module corresponds to the wheel end adhesion coefficient.

[0224] Specifically, step 32) is as follows:

[0225] Based on the wheel-end constraint state set, and considering the longitudinal and lateral coupling relationship of the tires, the combined slip effect, and friction constraints, the basic capability boundaries of each corner module under the current operating conditions are solved as follows:

[0226] ;

[0227] In the formula, To achieve a unified discrete time The basic steering capability boundary of the lower corner module. To achieve a unified discrete time The lower basic driving capability boundary, To achieve a unified discrete time Lower basic braking capacity boundary;

[0228] Combined with the target operating mode determined in step 2). and mode switching timing A mode constraint coefficient is introduced for the three control channels of steering, driving, and braking, denoted as . ,in , , and These represent the steering control channel, drive control channel, and brake control channel, respectively. The mode constraint coefficients are used to characterize the first... The corner module in the first Channel in various operating modes The level of permitted participation;

[0229] Consider the current operating mode Target-oriented operating mode The gradual transition, defining the first Each corner module in the channel The mode correction factor is:

[0230] ;

[0231] In the formula, To achieve a unified discrete time Next Each corner module in the channel The mode correction coefficient after the gradual transition from the current operating mode to the target operating mode; To achieve a unified discrete time The corresponding actual time; To achieve a unified discrete time Next mode switching pre-adjustment factor, To achieve a unified discrete time Next mode switching pre-adjustment duration, It is a saturation function;

[0232] The first after pattern constraint correction Each corner module in the channel The effective capability is represented as follows:

[0233] ;

[0234] In the formula, , and They represent the unified discrete time. The effective steering, driving, and braking capabilities of the corner module under the target operating mode.

[0235] Specifically, step 33) is as follows:

[0236] Construct the first Each corner module in the channel The following are the validity parameters, used to characterize the actual capacity of each module to handle different control channels under the target operating mode:

[0237] ;

[0238] In the formula, , and Indicates at the unified discrete time Effectiveness parameters of the lower corner module in steering, driving, and braking channels;

[0239] By combining the effective capabilities and channel effectiveness parameters of each channel, a unified discrete time-varying value is formed. Next Capability parameter vector of each corner module ,as follows:

[0240] ;

[0241] The vector of capability parameters of all angular modules constitutes the unified discrete time. Lower corner module capability parameter set as follows:

[0242] ;

[0243] In the formula, Used to characterize at a unified discrete time. The comprehensive execution capability of each module in different control channels under the current target operating mode;

[0244] Furthermore, based on the corner module capability parameter set The installation positions of each corner module and the interaction between the steering, drive, and braking control channels and the generalized forces of the vehicle are used to construct the feasible domain of the generalized forces of the vehicle at the current moment. It is used to characterize the achievable range of longitudinal force, lateral force and yaw moment of the whole vehicle under the current capability constraints.

[0245] Normalize the effectiveness parameters of each corner module within the same control channel to obtain the results at a unified discrete time. The lower corner modules in the channel Channel weight parameters The following are used for subsequent execution domain candidate construction, functional role division, and control allocation weight generation:

[0246] ;

[0247] In the formula, This refers to the total number of corner modules included in the transport equipment.

[0248] 4) Based on the target operating mode and mode switching timing in step 2) and the corner module capability parameter set in step 3), the execution domain affiliation, functional role, and control permissions of each corner module in the target operating mode are reconstructed to form the execution domain allocation result and control permission parameters; the specific steps are as follows:

[0249] 41) Based on the target operating mode and mode switching timing determined in step 2) and the corner module capability parameter set formed in step 3), construct the mode demand-capability supply coupling relationship to form the set of functions that each corner module can undertake in different execution domains and the candidate set of execution domains;

[0250] 42) An execution domain reconstruction method based on pattern-capability coupling graph matching and priority constraint optimization is adopted to make reconstruction decisions on the set of achievable functions and the candidate set of execution domains formed in step 41), and to determine the execution domain affiliation of each module in the target operating mode, as well as the main functional role, auxiliary functional role and support functional role.

[0251] 43) Based on the execution domain affiliation and functional role determined in step 42), classify the participation level, control sequence and takeover conditions of each module in the steering, driving and braking execution domains to form control authority parameters;

[0252] 44) Output the execution domain allocation results and control permission parameters.

[0253] Specifically, the steps for constructing the pattern demand-capacity supply coupling relationship in step 41) are as follows:

[0254] Combined with the target operating mode determined in step 2). and the corner module capability parameter set formed in step 3). Extract the execution domain baseline requirement vector corresponding to the target execution mode from the execution mode knowledge base. , represented as:

[0255] ;

[0256] In the formula, , and They represent the unified discrete time. The baseline demand intensity of the three execution domains of steering, driving and braking under the target operating mode;

[0257] Consider the corner module heterogeneity index in step 1). The impact of execution domain refactoring demand intensity is constructed by examining the actual demand intensity of each execution domain at the current moment, as follows:

[0258] ;

[0259] In the formula, To achieve a unified discrete time Next target running mode Next The baseline demand intensity for each execution domain; To achieve a unified discrete time Next The actual demand intensity of each execution domain This is the heterogeneity sensitivity coefficient; as the capability differences between corner modules increase, the corresponding execution domain's reconstruction requirements increase accordingly; based on the effectiveness parameters of each corner module obtained in step 3). The demand-capacity supply coupling matrix for the construction mode is as follows:

[0260] ;

[0261] In the formula, To achieve a unified discrete time The following is a model demand-capacity supply coupling matrix; Indicates at the unified discrete time Under the current target running mode, the corner module for the first The ability to match the requirements of each execution domain; subscript This indicates that the coupling matrix consists of 3 control channels and It consists of individual modules;

[0262] Constructing at a unified discrete time Next The set of functions that each module can undertake as follows:

[0263] ;

[0264] And construct at a unified discrete time Next Execution domain candidate set of execution domains as follows:

[0265] ;

[0266] In the formula, The coupling strength threshold, This is the minimum validity threshold for the execution domain; thus forming the set of functions that each module can undertake in different control channels and the candidate set of execution domains.

[0267] Specifically, step 42) is as follows:

[0268] Execution domain node set And corner module node set The construction pattern-capability coupling diagram is as follows:

[0269] ;

[0270] In the formula, at the unified discrete time lower weight set From the coupling matrix Decision; Introduction at the unified discrete time Variable allocation in the execution domain :

[0271] ;

[0272] To balance pattern requirement satisfaction, refactoring stability, and execution domain supply redundancy, the following execution domain refactoring optimization objectives are established:

[0273] ;

[0274] In the formula, Indicates at the unified discrete time Below, the domain allocation matrix is ​​executed. To optimize variables, the optimization objective is to refactor the execution domain. Perform a maximization solution; To achieve a unified discrete time The execution domain allocation matrix, To achieve a unified discrete time The goal of refactoring and optimizing the execution domain is to... To achieve a unified discrete time Execution domain priority weights To reconstruct the switching penalty coefficient, The penalty coefficient for supply and demand matching;

[0275] The execution domain priority weight is obtained by normalizing the current mode demand intensity, as follows:

[0276] ;

[0277] In the formula, To achieve a unified discrete time Next The priority weight of each execution domain at the current moment; To achieve a unified discrete time Next The actual demand intensity of each execution domain; The control channel index is used to traverse the three execution domains: steering, driving, and braking.

[0278] Depend on The resulting revenue streams are used to improve the match between control channel demand and corner module capacity supply; The penalty term is used to suppress frequent abrupt changes in the execution domain allocation results at adjacent time steps; by The penalty term is designed to reduce the discrepancy between the execution domain's supply capacity and the target pattern's demand.

[0279] The execution domain reconstruction optimization satisfies the following constraints:

[0280] ;

[0281] In the formula, For the first The minimum number of guaranteed corner modules in each execution domain; by solving the above optimization problem, the optimal execution domain allocation matrix under the target operating mode is obtained. as follows:

[0282] ;

[0283] In the formula, To achieve a unified discrete time The optimal execution domain allocation variable is used to characterize the first... Is the corner module assigned to the first...? One execution domain;

[0284] The optimal execution domain assignment results for each execution domain, obtained from the optimal execution domain assignment matrix, are as follows:

[0285] ;

[0286] In the formula, , and These represent the optimal execution domains for the steering, drive, and braking control channels, respectively.

[0287] Specifically, step 43) is as follows:

[0288] Construct the first The corner module in the first The role evaluation metrics in each execution domain are as follows:

[0289] ;

[0290] In the formula, To achieve a unified discrete time Next The corner module in the first Relative capacity indicators in each execution domain To prevent the division of tiny positive numbers with a denominator of zero;

[0291] Based on the aforementioned role evaluation index, at a unified discrete time... Next The corner module in the first Functional roles in each execution domain The division is as follows:

[0292] ;

[0293] In the formula, Indicates the main functional role. Indicates a support role. Indicates a safeguard function role. Indicates that it does not participate in the execution domain; , and The thresholds for role classification are divided into primary function, auxiliary function, and support function, and they meet the following requirements: ;

[0294] To determine the control permission parameters for each module within the execution domain, a role weight coefficient is introduced. Constructed at a unified discrete time Next The corner module in the first The permission strengths in each execution domain are as follows:

[0295] ;

[0296] In the formula, To achieve a unified discrete time Next The corner module in the first The strength of permissions in each execution domain; To achieve a unified discrete time Below and functional roles The corresponding role weight coefficient; To achieve a unified discrete time Next The corner module in the first Functional role identifiers in each execution domain; To achieve a unified discrete time Next Each module in the execution domain The validity parameters are as follows;

[0297] The permission strength of each module within the same execution domain is normalized to obtain the following control permission parameters:

[0298] ;

[0299] In the formula, Indicates at the unified discrete time Next The corner module in the first The percentage of control permissions in each execution domain;

[0300] The control order of each module within the execution domain is determined based on control authority parameters, defined at a unified discrete time. Next order indicators for:

[0301] ;

[0302] In the formula, The smaller the value, the higher the priority control order of the corner module in the corresponding execution domain;

[0303] To establish takeover conditions within the execution domain, a unified discrete time is defined. Next The corner module in the first Takeover trigger variables in each execution domain as follows:

[0304] ;

[0305] In the formula, As a safety and effectiveness threshold, To achieve a unified discrete time Next Each corner module fault flag; when When this occurs, it indicates that the corner module has triggered a downgrade or exit condition, and its control authority is transferred to another corner module that is earlier in the same execution domain and has not triggered a takeover condition.

[0306] Specifically, step 44) is as follows:

[0307] Based on the obtained optimal execution domain attribution results and the functional roles obtained in step 43). Control permission parameters Control sequence and takeover conditions Constructed at a unified discrete time Set of execution domain allocation and control permissions output as follows:

[0308] .

[0309] 5) Based on the target operating mode in step 2), the state dataset in step 1), and the task requirement information, generate a vehicle control target that matches the requirements of the target operating mode. Combine this with the corner module capability parameter set in step 3) to perform executability verification and constraint correction on the vehicle control target. Then, based on the corrected vehicle control target and the execution domain allocation results in step 4), decompose the vehicle control target into execution domain control targets. The specific steps are as follows:

[0310] 51) Based on the state dataset formed in step 1), the target operation mode determined in step 2), and the task requirement information, generate a vehicle control target that matches the operation requirements of the future preset prediction interval.

[0311] 52) Combining the vehicle generalized force feasible boundary corresponding to the corner module capability parameter set formed in step 3), perform executability verification and constraint correction on the vehicle control target generated in step 51) to obtain the executable vehicle control target under the target operation mode;

[0312] 53) Based on the execution domain allocation results and control authority parameters formed in step 4), establish the mapping relationship between the vehicle control target and the control targets of each execution domain, and decompose the executable vehicle control target obtained in step 52) into the control targets of each execution domain, and output the control targets of each execution domain and their priority information.

[0313] Specifically, the generation of the vehicle control target in step 51) is as follows:

[0314] Based on the state dataset formed in step 1) The target operating mode determined in step 2) In addition to task requirement information, construct the vehicle task tracking error vector. The original vehicle control objective is formed by adopting a target generation law that matches the target operating mode, and is expressed as:

[0315] ;

[0316] In the formula, To achieve a unified discrete time The original vehicle control target is set below. To achieve a unified discrete time Next target running mode The corresponding feedback gain matrix, To achieve a unified discrete time The error vector between the task requirements and the current state of the vehicle. To achieve a unified discrete time The feedforward term is formed by the preset task trajectory, path curvature, or prior information on operating conditions; thus, the vehicle control objective matching the future preset prediction interval operating requirements is obtained as follows:

[0317] ;

[0318] In the formula, To achieve a unified discrete time The original generalized force target of the whole vehicle that matches the target operating mode; , and At the unified discrete time respectively The original longitudinal force target, the original lateral force target, and the original yaw moment target are determined.

[0319] Specifically, step 52) is as follows:

[0320] Combined with the corner module capability parameter set formed in step 3) The corresponding generalized force feasible boundary of the whole vehicle is used to perform an executability check on the original whole vehicle control target generated in step 51). When the original whole vehicle control target exceeds the current capability boundary, a weighted projection method is used for constraint correction to obtain the following executable whole vehicle control target:

[0321] ;

[0322] In the formula, To achieve a unified discrete time The generalized force feasible boundary of the whole vehicle The target control variable to be optimized in the generalized force space of the whole vehicle. To achieve a unified discrete time The executable vehicle control objective after lower constraint correction. The weighted matrix for the overall vehicle control objectives; , and At the unified discrete time respectively The executable longitudinal force target, lateral force target, and yaw moment target after lower constraint correction.

[0323] Specifically, step 53) is as follows:

[0324] Based on the execution domain allocation and control permission output set formed in step 4). The domain-level comprehensive undertaking capacity of the steering execution domain, drive execution domain, and braking execution domain is calculated separately and defined as follows:

[0325] ;

[0326] In the formula, To achieve a unified discrete time Next The overall capacity of each execution domain To achieve a unified discrete time The first one obtained in step 4) Each execution domain belongs to a set. To achieve a unified discrete time The corresponding control permission parameters are as follows. To achieve a unified discrete time The channel validity parameters obtained in step 3) below;

[0327] Combined with the control channel priority weights formed in step 4) The assumption coefficients of each execution domain for the vehicle control objective are constructed as follows:

[0328] ;

[0329] In the formula, To achieve a unified discrete time Next The target acceptance coefficient of each execution domain; To achieve a unified discrete time Next Priority weights for each control channel; To achieve a unified discrete time Next The comprehensive capacity of each execution domain; This is a control channel index used to traverse the three control channels: steering, drive, and braking.

[0330] To distinguish between traction and braking conditions, a sign assignment factor is introduced for the longitudinal target as follows:

[0331] ;

[0332] ;

[0333] In the formula, To achieve a unified discrete time The vertical target activation factor of the driving execution domain. To achieve a unified discrete time Longitudinal target activation factor in the lower braking execution domain; It is a symbolic function; To achieve a unified discrete time The executable longitudinal force target after lower constraint correction;

[0334] The control objectives for each execution domain are represented as follows:

[0335] ;

[0336] ;

[0337] ;

[0338] In the formula, , and At the unified discrete time respectively The control objectives of the steering execution domain, drive execution domain, and braking execution domain;

[0339] Based on the control objectives and their respective coefficients of each execution domain, a unified discrete time is constructed. Control objectives of each execution domain as follows:

[0340] ;

[0341] Therefore, the output is at a unified discrete time. Information on control targets and priorities of each execution domain That is, the target responsibility coefficient of each execution domain for the overall vehicle control objective.

[0342] 6) Based on the corner module capability parameter set in step 3), the control authority parameters in step 4), and the control objectives of each execution domain in step 5), construct a hierarchical coordination control allocation model, perform inter-domain coordination allocation and corner module-level optimization allocation of the control tasks of each corner module, and form corner module control instructions;

[0343] The specific steps for controlling the allocation in step 6) are as follows:

[0344] 61) Based on the corner module capability parameter set in step 3), the control authority parameters in step 4), and the control objectives of each execution domain obtained in step 5), construct the control efficiency relationship of each execution domain and the control efficiency relationship of each corner module to form a hierarchical coordination control allocation model;

[0345] 62) Adopt a role-permission constraint-based inter-domain coordination and allocation method to allocate responsibility and contribution ratios for the control objectives of each execution domain, and determine the target responsibilities and coordination weights of each execution domain at the current moment;

[0346] 63) Using a corner module-level coordination allocation method based on control efficiency matrix, consistency constraints and weighted quadratic optimization, the target load of each execution domain determined in step 62) at the current time is solved at the corner module level, and the steering control quantity, drive control quantity and braking control quantity of each corner module are calculated to form corner module control commands.

[0347] Step 61) is as follows:

[0348] Based on the corner module capability parameter set formed in step 3) The execution domain reconstruction result formed in step 4) and the control objectives of each execution domain obtained in step 5). A hierarchical coordinated control allocation model consisting of vehicle targets, execution domain targets, and corner module control variables is established.

[0349] Suppose that at the unified discrete time... Next The domain-level generalized force bearing vector of each execution domain at the current moment. for:

[0350] ;

[0351] In the formula, To achieve a unified discrete time Next Domain-level generalized force-bearing vector of each execution domain; , and At the unified discrete time respectively The longitudinal force, lateral force, and yaw moment borne by the lower execution domain;

[0352] The vehicle's executable control objectives then satisfy the following:

[0353] ;

[0354] In the formula, To achieve a unified discrete time The following can execute the vehicle control objectives;

[0355] The execution domain control targets formed in step 5) As domain-level reference bearing vectors, they respectively correspond to the basic bearing targets of the steering execution domain, drive execution domain, and braking execution domain in the generalized force space of the whole vehicle;

[0356] For those belonging to the first For each module in the execution domain, construct its module-level control efficiency vector for the generalized forces of the entire vehicle. And thus form at a unified discrete time Next Control efficiency matrix of each execution domain as follows:

[0357] ;

[0358] In the formula, , For the first Number of corner modules within each execution domain From the unified discrete time Next The effective capability boundary, installation location, current configuration, and control channel type of each corner module are jointly determined;

[0359] Therefore, a foundation was established. For the goals of the higher authorities, As a domain-level reference target, with A hierarchical coordinated control allocation model for the control variables of the corner module.

[0360] Step 62) is as follows:

[0361] For each execution domain control objective obtained in step 5), considering the differences in functional roles, control permissions, effectiveness levels, and takeover states of the internal modules of different execution domains, the following is constructed: The corner module in the first The role-permission factors in each execution domain are as follows:

[0362] ;

[0363] In the formula, To achieve a unified discrete time Next The corner module in the first Role-permission factors in an execution domain To achieve a unified discrete time The role weight coefficients corresponding to the functional roles are as follows: To achieve a unified discrete time Lower control permission parameters, To achieve a unified discrete time The trigger variable for the sub-module is: when a certain module triggers the exit or downgrade condition, its corresponding role-permission factor automatically decays.

[0364] Combined with the execution domain target acceptance coefficient in step 5) and the module-level control efficiency vector in step 61). Construct the first The responsibility representation matrix of each execution domain is as follows:

[0365] ;

[0366] In the formula, To achieve a unified discrete time Next The responsibility representation matrix of each execution domain at the current moment is used to characterize the comprehensive ability of the execution domain to bear the various components of the generalized force of the whole vehicle under the combined effect of role structure, authority structure and actual control capability.

[0367] Based on domain-level reference bearing vector The residuals are allocated in a coordinated manner between construction domains as follows:

[0368] ;

[0369] In the formula, Indicates at the unified discrete time The vehicle can execute control objectives Compared with the control objectives of each execution domain in step 5) The remaining coordination amount between the sums;

[0370] The remaining coordination quantity is allocated according to the responsibility representation matrix of each execution domain to obtain the first... The final target workload of each execution domain at the current moment is as follows:

[0371] ;

[0372] In the formula, To achieve a unified discrete time Next The ultimate target workload of each execution domain. To prevent the matrix from having singular tiny positive numbers, It is the identity matrix;

[0373] The control objectives of each execution domain are retained as the basic workload; the remaining coordination workload caused by changes in mode, capability differences and permission changes is then distributed secondary based on the responsibility and capability of each execution domain under the current role-permission structure; thereby realizing the inter-domain coordination and allocation that adaptively adjusts with role level, permission ratio, module availability and real-time capability status.

[0374] Definition of the first The coordination weights of each execution domain, to characterize the degree of coordination dominance of each execution domain at the current moment, are as follows:

[0375] ;

[0376] In the formula, Indicates the first Execution domain responsibility representation matrix The trace is used to compress the comprehensive responsibility capacity of the execution domain into a scalar; To achieve a unified discrete time Next The coordination weight of each execution domain; the coordination weight reflects the degree of dominance of the execution domain in the overall vehicle coordination task under the current role-permission structure.

[0377] Step 63) is as follows:

[0378] For the Let there be an execution domain, and its internal corner module control vector be defined. for:

[0379] ;

[0380] In the formula, To achieve a unified discrete time Next Control vectors of internal corner modules within each execution domain; To achieve a unified discrete time Next The execution domain is numbered as follows The angle module control quantity; For the first The corner module number that participates in control allocation within each execution domain. ; For the first The number of corner modules participating in control allocation within each execution domain; ; Indicates at the unified discrete time The angular module control vector in the downstairs steering execution domain. Indicates at the unified discrete time The control vector of the corner module in the lower drive execution domain. Indicates at the unified discrete time The angular module control vector in the lower braking execution domain;

[0381] No. The corner module-level coordination allocation of each execution domain is solved by the following weighted quadratic optimization problem:

[0382] ;

[0383] In the formula, To achieve a unified discrete time Next At the current moment, each execution domain The optimal angle module control vector; For the first Each execution domain corner module control quantity optimization variable; To achieve a unified discrete time Next Control efficiency matrix of each execution domain; To achieve a unified discrete time Next The ultimate target workload of each execution domain; The target tracking weight matrix; This is the consistency constraint matrix; This is the consistency and coordination coefficient; To control the allocation balance coefficient; To achieve a unified discrete time Bottom corner module weight matrix; To achieve a unified discrete time The square root of the weight matrix of the lower corner module;

[0384] The corner module weight matrix is ​​defined as follows:

[0385] ;

[0386] This indicates that corner modules with high functional roles and high authority levels, and which have not triggered takeover conditions, have smaller weight penalties and are given priority to undertake more control tasks in the optimization solution; while corner modules with low authority or in a degraded state have their workload automatically limited.

[0387] The optimization problem satisfies the following constraints:

[0388] ;

[0389] In the formula, and At the unified discrete time respectively Next The lower and upper bounds of the control variables of each corner module within the execution domain. This is the constraint vector for the rate of change of the control quantity;

[0390] Based on the control channel type corresponding to each execution domain, the optimization results are mapped to corner module control instructions as follows:

[0391] ;

[0392] In the formula, To achieve a unified discrete time Next Steering control commands from each corner module To achieve a unified discrete time Send drive control commands, To achieve a unified discrete time Issue a braking control command; , and At the unified discrete time respectively Next The optimal control quantity of each corner module in the steering execution domain, drive execution domain, and braking execution domain; Indicates at the unified discrete time Next The corner module belongs to the first Find the optimal execution domains; synthesize the optimization results of each execution domain to construct a unified discrete time step. Lower corner module control instruction set as follows:

[0393] .

[0394] 7) Based on the corner module control commands in step 6), combined with the mode switching timing in step 2) and the execution domain allocation results and control authority parameters in step 4), a transition takeover process is performed on the execution domain affiliation and control authority change process before and after reconstruction. The processed control commands are then sent to the steering actuator, drive actuator, and braking actuator corresponding to each corner module. Simultaneously, based on the actual execution results of each corner module and the vehicle motion response, the state dataset in step 1), the target operating mode and mode switching timing in step 2), the corner module capability parameter set in step 3), and the execution domain allocation results in step 4) are updated to achieve closed-loop coordinated control under multi-mode execution domain reconstruction. The specific steps are as follows:

[0395] 71) Based on the corner module control instructions in step 6), the execution domain allocation results in step 4), and the mode switching timing in step 2), construct the transition takeover relationship corresponding to the control channel participation relationship and control authority change relationship before and after reconstruction;

[0396] 72) Based on the permission gradual change and instruction continuous processing method, the transition takeover relationship constructed in step 71) is smoothed to form continuous control instructions, which are then sent to the steering actuator, drive actuator and braking actuator corresponding to each corner module.

[0397] 73) Based on the actual execution results of each corner module and the vehicle motion response, update the state dataset in step 1), the target operating mode and mode switching timing in step 2), the corner module capability parameter set in step 3), and the execution domain allocation results in step 4), thereby realizing closed-loop coordinated control under multi-mode execution domain reconstruction.

[0398] Specifically, step 71) is as follows:

[0399] Based on the corner module control instruction set in step 6) The execution domain reconstruction results in step 4) and the timing of mode switching in step 2). Extract the execution domain participation relationships and control permission change relationships before and after the reconstruction; let the first... The corner module in the first The pre-reconstruction participation flag and permission parameters in each control channel are as follows: and After reconstruction, the participating tags and permission parameters are as follows: and Then define the first The corner module in the first Transition takeover coefficient in each control channel for:

[0400] ;

[0401] In the formula, To achieve a unified discrete time Lower transition gradient factor, The transition processing time is determined accordingly; this establishes the transition takeover relationship corresponding to the changes in control channel participation and control authority before and after the reconstruction.

[0402] Specifically, step 72) is as follows:

[0403] Suppose that at the unified discrete time... Next The corner module in the first The pre-reconfiguration control commands in each control channel are: The corner module control command obtained in step 6) is Then its transition instruction is expressed as:

[0404] ;

[0405] Applying a rate-of-change constraint to the transition command yields the following continuous control command:

[0406] ;

[0407] In the formula, To achieve a unified discrete time Next Each corner module in the channel Continuous control commands in; For the previous discrete time... The corresponding continuous control commands; To achieve a unified discrete time The transition instruction obtained after reconstruction and transition processing; Indicates at the unified discrete time The previous sampling time; For the first Each corner module in the channel The maximum allowable single-step instruction change; Indicates The saturation function is the amplitude limiting boundary;

[0408] Based on the control channel type, continuous control commands are issued to the corresponding actuators of each corner module, i.e.:

[0409] ;

[0410] In the formula, To switch control commands to the actuator, To drive the actuator control commands, For control commands of the braking actuator.

[0411] Specifically, step 73) is as follows:

[0412] Based on the actual execution results of each module and the vehicle's motion response, a unified discrete time step is constructed. Next The corner module in the first Execution deviation in each control channel as follows:

[0413] ;

[0414] In the formula, To achieve a unified discrete time Next The corner module in the first The actual execution volume in each control channel;

[0415] Based on the aforementioned execution deviation, a closed-loop correction is performed on the channel validity parameter in step 3), expressed as follows:

[0416] ;

[0417] In the formula, At the unified discrete moment The following is the closed-loop correction of the first... Each corner module in the channel The channel validity parameters below; To achieve a unified discrete time Channel validity parameters before correction; To achieve a unified discrete time Next Each corner module in the channel Execution deviations below; To achieve a unified discrete time Next Each corner module in the channel Continuous control commands issued by the user; This is a saturation function used to limit the corrected channel validity parameters within an allowable range; This is the validity correction factor; To prevent the division of tiny positive numbers with a denominator of zero;

[0418] Based on the corrected channel validity parameters, the control authority parameters in step 4) are normalized and updated as follows:

[0419] ;

[0420] In the formula, To achieve a unified discrete time The next updated version The corner module in the first Control permission parameters in each execution domain; To achieve a unified discrete time Control permission parameters before the update; To achieve a unified discrete time Channel validity parameters after lower closed-loop correction; To achieve a unified discrete time Next The optimal execution domain set corresponding to each control channel; This is the index of the corner module within the execution domain, used for traversal. All corner modules in;

[0421] This enables closed-loop coordinated control under multi-mode execution domain reconstruction.

[0422] This invention has many specific applications. The above description is only a preferred embodiment of this invention. It should be noted that for those skilled in the art, several improvements can be made without departing from the principle of this invention, and these improvements should also be considered within the scope of protection of this invention.

Claims

1. A multi-mode execution domain reconfiguration system for a drive-by-wire chassis of a transport vehicle, characterized in that, include: The system comprises a state-aware unit, an execution domain reconstruction unit, a coordination and control unit, and an execution unit. The state perception unit is used to collect environmental information, operating condition information, task requirement information, vehicle motion state information and state information of each module of the transport equipment, and output a state dataset. The execution domain reconstruction unit is used to determine the target operating mode of the carrier equipment based on the state dataset output by the state perception unit, and to reconstruct the functional division and control authority of each module according to the target operating mode, and output mode information, execution domain allocation results and control authority parameters. The coordination and control unit is used to generate a vehicle control target based on the mode information output by the execution domain reconstruction unit, the execution domain allocation result, and the control authority parameters, and to decompose the vehicle control target into control instructions for each module. The execution unit is used to process the corner module control commands output by the coordination control unit and send them to each corner module execution mechanism. At the same time, it collects execution feedback information and returns it to the state perception unit, the execution domain reconstruction unit, and the coordination control unit.

2. The multi-mode execution domain reconfiguration system for a drive-by-wire chassis of a transport vehicle according to claim 1, characterized in that, The execution domain reconstruction unit includes: a mode determination module, a capability assessment module, and a reconstruction management module; The mode determination module is used to perform long-term domain prediction and intelligent decision-making on the evolution of operating conditions within a preset prediction interval based on the state dataset output by the data processing module, determine the target operating mode, and send the target operating mode to the target generation module in the reconfiguration management module and the coordination control unit; the target operating mode is determined based on environmental characteristics, operating condition characteristics, vehicle attitude characteristics, and task requirement characteristics. The capability assessment module is used to quantitatively assess the available control capabilities of each corner module under the current operating conditions and target operating mode constraints based on the status dataset output by the data processing module, forming a corner module capability parameter set, and sending the corner module capability parameter set to the control allocation module in the reconfiguration management module and the coordination control unit; the corner module capability parameter set includes steering adjustment capability parameters, drive output capability parameters, braking adjustment capability parameters and allocable weight parameters for each corner module; The reconstruction management module is used to reconstruct the execution domain affiliation, functional role, and control authority of each corner module in the current mode according to the target operating mode output by the mode determination module and the corner module capability parameter set output by the capability assessment module, forming execution domain allocation results and control authority parameters, and sending the execution domain allocation results and control authority parameters to the coordination and control unit; the execution domain allocation results represent the execution domain to which each corner module belongs and its functional role, and the control authority parameters represent the control authority and constraint boundaries of each corner module in the target operating mode.

3. The multi-mode execution domain reconfiguration system for a drive-by-wire chassis of a transport vehicle according to claim 1, characterized in that, The coordination and control unit includes: a target generation module, a target decomposition module, and a control allocation module; The target generation module is used to generate a vehicle control target that matches the current operating requirements based on the target operating mode output by the mode determination module, the status dataset output by the data processing module, and the task requirement information, and then send the vehicle control target to the target decomposition module; the vehicle control target includes one or more of the following: trajectory tracking target, attitude stabilization target, passability target, and motion response target; The target decomposition module is used to decompose the vehicle control target into execution domain control targets based on the vehicle control target output by the target generation module and the execution domain allocation result output by the reconfiguration management module, and send the execution domain control targets to the control allocation module; the execution domain control targets represent the steering adjustment task, drive allocation task and braking allocation task undertaken by the corresponding execution domain. The control allocation module is used to calculate and allocate the control tasks of each corner module according to the control targets of each execution domain, the corner module capability parameter set, and the control permission parameters, to form corner module control instructions, and send the corner module control instructions to the transition processing module in the execution unit; the corner module control instructions include the target steering angle, target driving torque, and target braking force corresponding to each corner module.

4. The multi-mode execution domain reconfiguration system for a drive-by-wire chassis of a transport vehicle according to claim 1, characterized in that, The execution unit includes: a transition processing module, an instruction issuing module, and a feedback module; The transition processing module is used to continuously process the corner module control commands output by the control allocation module during the operation mode switching, execution domain adjustment, or corner module function migration, and send the processed control commands to the command issuing module; the continuous processing is used to make the target steering angle, target driving torque, and target braking force change continuously before and after reconstruction according to a preset transition law; The instruction issuing module is used to receive control instructions output by the transition processing module, and send the control instructions to the steering actuator, drive actuator and braking actuator corresponding to each corner module respectively, so as to drive each corner module to perform corresponding control actions; the control instructions include target steering angle instruction, target drive torque instruction and target braking force instruction; The feedback module is used to collect the actual execution results of each corner module and the vehicle motion response, and return the feedback results to the data processing module, the mode determination module and the reconstruction management module to update the status dataset, the current operating mode and the execution domain allocation results; the feedback results include the actual steering response, actual driving response, actual braking response and vehicle motion response of each corner module.

5. A coordinated control method for a multi-mode execution domain reconfiguration system of a drive-by-wire chassis for transport equipment, based on the system described in any one of claims 1-4, characterized in that, The steps are as follows: 1) Collect environmental information, operating condition information, mission requirement information, vehicle motion status information, and status information of each module of the transport equipment, and form a status dataset after unifying the time base, denoising correction and feature extraction; 2) Based on the state dataset, perform long-term time-domain prediction and intelligent decision-making on the evolution of working conditions within the future preset prediction interval, and determine the target operating mode and the timing of mode switching. 3) Based on the state dataset, the available controllability of each corner module under the constraints of the current working condition and the target operating mode is quantitatively evaluated, and an actuator effectiveness model is established to form a corner module capability parameter set; 4) Based on the target operating mode and mode switching timing in step 2) and the corner module capability parameter set in step 3), the execution domain affiliation, functional role and control authority of each corner module in the target operating mode are reconstructed to form the execution domain allocation result and control authority parameters; 5) Based on the target operation mode in step 2), the state dataset in step 1), and the task requirement information, generate a vehicle control target that matches the requirements of the target operation mode, and perform executability verification and constraint correction on the vehicle control target in conjunction with the corner module capability parameter set in step 3); and decompose the vehicle control target into control targets for each execution domain according to the corrected vehicle control target and the execution domain allocation results in step 4). 6) Based on the corner module capability parameter set in step 3), the control authority parameters in step 4), and the control objectives of each execution domain in step 5), construct a hierarchical coordination control allocation model, perform inter-domain coordination allocation and corner module-level optimization allocation of the control tasks of each corner module, and form corner module control instructions; 7) Based on the corner module control instructions in step 6), combined with the mode switching timing in step 2) and the execution domain allocation results and control authority parameters in step 4), a transition takeover process is performed on the execution domain affiliation and control authority change process before and after reconstruction. The processed control instructions are sent to the steering actuator, drive actuator and braking actuator corresponding to each corner module. At the same time, according to the actual execution results of each corner module and the vehicle motion response, the state dataset in step 1), the target operating mode and mode switching timing in step 2), the corner module capability parameter set in step 3), and the execution domain allocation results in step 4) are updated to achieve closed-loop coordinated control under multi-mode execution domain reconstruction.

6. The coordinated control method according to claim 5, characterized in that, The specific steps of step 2) are as follows: 21) Based on the state dataset in step 1), establish a knowledge base for the operation mode of the launch vehicle, including the applicable working conditions, task requirements, performance target weights and execution constraint boundaries corresponding to each operation mode; 22) A multimodal temporal prediction network based on spatiotemporal attention mechanism is adopted to jointly encode environmental features, working condition features, task requirement features, vehicle motion features and corner module state features and extrapolate long-term trends to obtain the working condition evolution sequence within the future preset prediction interval. 23) Using a pattern matching method based on contrastive representation learning, the working condition evolution sequence obtained in step 22) is matched with the feature prototypes of each pattern in the operation mode knowledge base to generate a set of candidate operation modes; 24) Using a comprehensive evaluation method based on the mode benefit prediction network and the switching cost assessment model, the task completion benefit, stability benefit, energy consumption benefit and mode switching cost of the candidate operating mode set obtained in step 23) are jointly predicted within the prediction interval to form a comprehensive benefit ranking result for each operating mode. 25) Using a rolling time-domain optimization-based mode decision-making method, the comprehensive benefit ranking results obtained in step 24) are used to make time-series decisions to determine the target operating mode and the timing of mode switching.

7. The coordinated control method according to claim 6, characterized in that, The specific steps of step 3) are as follows: 31) Based on the state dataset formed in step 1), extract the road surface adhesion conditions, wheel end normal load, tire slip state, and driving, braking, and steering execution boundary information to construct the wheel end constraint state set of each corner module; 32) Based on the tire force constraint analysis and friction circle correction method, the wheel end constraint state set obtained in step 31) is processed to solve the available boundaries of the longitudinal and lateral forces at the wheel end of each corner module under the current working condition and to perform mode constraint correction, so as to obtain the steering execution capability, driving execution capability and braking execution capability of each corner module under the target operating mode. 33) Based on the generalized force feasible domain construction and control efficiency matrix modeling method, the execution capabilities of each corner module obtained in step 32) are mapped to the actuator effectiveness model, generating the corner module capability parameter set and channel weight parameters.

8. The coordinated control method according to claim 7, characterized in that, The specific steps of step 4) are as follows: 41) Based on the target operating mode and mode switching timing determined in step 2) and the corner module capability parameter set formed in step 3), construct the mode demand-capability supply coupling relationship to form the set of functions that each corner module can undertake in different execution domains and the candidate set of execution domains; 42) An execution domain reconstruction method based on pattern-capability coupling graph matching and priority constraint optimization is adopted to make reconstruction decisions on the set of achievable functions and the candidate set of execution domains formed in step 41), and to determine the execution domain affiliation of each module in the target operating mode, as well as the main functional role, auxiliary functional role and support functional role. 43) Based on the execution domain affiliation and functional role determined in step 42), classify the participation level, control sequence and takeover conditions of each module in the steering, driving and braking execution domains to form control authority parameters; 44) Output the execution domain allocation results and control permission parameters.

9. The coordinated control method according to claim 8, characterized in that, The specific steps of step 5) are as follows: 51) Based on the state dataset formed in step 1), the target operation mode determined in step 2), and the task requirement information, generate a vehicle control target that matches the operation requirements of the future preset prediction interval. 52) Combining the vehicle generalized force feasible boundary corresponding to the corner module capability parameter set formed in step 3), perform executability verification and constraint correction on the vehicle control target generated in step 51) to obtain the executable vehicle control target under the target operation mode; 53) Based on the execution domain allocation results and control authority parameters formed in step 4), establish the mapping relationship between the vehicle control target and the control targets of each execution domain, and decompose the executable vehicle control target obtained in step 52) into the control targets of each execution domain, and output the control targets of each execution domain and their priority information.

10. The coordinated control method according to claim 9, characterized in that, The specific steps for control allocation in step 6) are as follows: 61) Based on the corner module capability parameter set in step 3), the control authority parameters in step 4), and the control objectives of each execution domain obtained in step 5), construct the control efficiency relationship of each execution domain and the control efficiency relationship of each corner module to form a hierarchical coordination control allocation model; 62) Adopt a role-permission constraint-based inter-domain coordination and allocation method to allocate responsibility and contribution ratios for the control objectives of each execution domain, and determine the target responsibilities and coordination weights of each execution domain at the current moment; 63) Using a corner module-level coordination allocation method based on control efficiency matrix, consistency constraints and weighted quadratic optimization, the target load of each execution domain determined in step 62) at the current time is solved at the corner module level, and the steering control quantity, drive control quantity and braking control quantity of each corner module are calculated to form corner module control commands.

Images