Scene-based mine car unmanned driving safety control method and system

Abstract

The invention discloses a scene-based mine car unmanned driving safety control method and system, and belongs to the technical field of mine automatic driving. The method comprises the steps of collecting scene data of mining area work in real time; performing feature extraction according to the scene data and the task data, and dividing the extracted features into a macroscopic scene layer and a microscopic scene layer; using the decision tree model to make a vehicle behavior decision and output a driving path and a target speed; a rule engine is adopted for scene risk assessment, control arbitration is conducted on control instructions of a vehicle driving system, a braking system, gears, an operation system and a warning system according to scene risk levels, safety measures are dynamically adjusted, and a safety control instruction for the vehicle is output; and the vehicle executes the control instruction according to the safety control instruction and the current vehicle state. According to the invention, scenes with different complexity and risk levels in a mining area can be identified through a double-layer classification method, and proper safety measures can be taken by vehicles in different scenes.

Description

Technical Field

[0001] The present invention relates to a scenario-based mine car unmanned driving safety control method and system, belonging to the technical field of mine automatic driving. Background Art

[0002] In modern mining operations, the development of driverless mine car technology has significantly improved production efficiency and reduced labor costs. However, the complex terrain, changing climate conditions, and particularity of mine car operation scenarios in mining areas have put forward higher requirements for the functional safety of driverless technology. Traditional functional safety methods are mostly used for flat roads and standardized driving environments, but in high-risk, highly dynamic scenarios such as mining areas, existing safety methods are not perfect enough. There is an urgent need for a functional safety method and system that can adapt to the diverse operation scenarios of mining areas to ensure the safe operation of mine cars in various environments.

[0003] Most existing unmanned mine car systems rely on fixed preset algorithms and path planning solutions, and cannot flexibly respond to emergencies in different scenarios. For example, when a mine car drives to an area with a large slope, the system may not be able to accurately judge and adjust the driving strategy in time, causing the mine car to lose control or landslide.

[0004] Therefore, there is an urgent need for a functional safety method and system for unmanned mining vehicles with dynamic adjustment based on scene recognition, which can perceive environmental changes in real time, accurately identify the current scene, and adjust the driving strategy according to the different characteristics of the scene to ensure the safety and stability of the mining vehicle in various operating scenarios.

[0005] Patent CN115743169A discloses a functional safety assurance device, method and unmanned vehicle, including a controlled system: a complex controller, used to control the controlled system to operate under normal conditions; a safety controller, used to control the controlled system to operate under a safety-critical state; the safety controller has higher safety than the complex controller; a decision module, used to monitor the operating state of the controlled system in real time, and predict the state of the controlled system during operation based on the monitoring results, and switch between the complex controller and the safety controller based on the prediction results; when the prediction is judged to be no violation of the safety constraints, the complex controller is used to control the controlled system to operate under normal conditions; when the prediction is judged to be a violation of the safety constraints, the safety controller is switched to control the controlled system to operate under a safety-critical state to ensure the safety of the operating state of the controlled system.

[0006] Patent CN111791896A discloses a safety control method and system for an unmanned vehicle, the safety control method comprising: monitoring whether the unmanned vehicle has a safety risk, the safety risk comprising at least one of the following: failure of the vehicle power supply, failure of the unmanned driving domain controller ADCU, failure of the vehicle controller VCU, and failure of the core functional module of the ADCU; in the case of monitoring the safety risk of failure of the vehicle power supply, controlling the start of the backup power supply; in the case of monitoring the safety risk of failure of the ADCU, the VCU automatically controls the unmanned vehicle to brake and stop; in the case of monitoring the safety risk of failure of the VCU, the ADCU unauthorizedly controls the unmanned vehicle to perform parking braking; and in the case of monitoring the failure of the core functional module, the ADCU controls the parking according to the set safe parking trajectory.

[0007] Patent CN110816443B discloses a vehicle monitoring method, device and unmanned vehicle, wherein the vehicle monitoring method includes: monitoring the target information of a functional module by a first monitoring module to obtain a first monitoring result, the functional modules include multiple modules, transmitting the first monitoring result to a second monitoring module, and the second monitoring module monitors the target information of the functional module to obtain a second monitoring result, if the first monitoring module fails, the second monitoring module monitors the target information of the functional module, if the second monitoring module fails, the first monitoring module monitors the target information of the functional module; if both the first monitoring module and the second monitoring module fail, the functional modules form a monitoring system to monitor the target information of the functional modules.

[0008] Wo2006066507A1 discloses a method for implementing a scene security function, by setting limited area information on the network side, and judging whether the mobile terminal belongs to a limited user when it is judged that the mobile terminal enters the limited area. The security of special scenes is realized by a special function that cannot be authorized by mobile.

[0009] CN115743169A discloses a functional safety assurance device, method and unmanned vehicle. The method controls the system in normal and safety-critical states respectively through the coordinated work of a complex controller and a safety controller to ensure the safety of the system. The method introduces multiple controllers, which increases the complexity of the system and makes the design, verification, maintenance and other processes more cumbersome; the safety of the system depends largely on the decision-making module's ability to predict the system's operating status. If the prediction algorithm is inaccurate, wrong switching decisions may occur, causing the system to operate in an inappropriate control mode.

[0010] CN115743169A discloses a safety control method and system for unmanned vehicles. This safety control method and system for unmanned vehicles has a multi-level safety assurance mechanism, and adopts different response measures for different types of faults to ensure the safety of the vehicle. However, this design may also have some potential disadvantages: limitations in fault handling. This method mainly focuses on handling specific hardware failures without fully considering other potential safety risks such as software failures, sensor failures, or communication failures; reliance on a single failure mode. This design assumes that only one failure occurs at a time, and the system can independently respond to these failures. However, in actual applications, multiple failures may occur simultaneously, which may exceed the design capability of the system and lead to unforeseen safety hazards.

[0011] CN110816443B discloses a vehicle monitoring method, device and unmanned vehicle. This vehicle monitoring method and device design has a multi-level redundant monitoring mechanism, that is, the monitoring reliability of the vehicle functional modules is ensured by mutual backup of the first monitoring module and the second monitoring module, and the monitoring system composed of various functional modules when both fail. However, this design may also have the following potential disadvantages: the system complexity increases, and multiple monitoring modules and a backup monitoring system are introduced into the design, which greatly increases the complexity of the system. The increased complexity may lead to an increase in development and maintenance costs, while increasing potential failure points in the system; the fault detection complexity is high, and when the first and second monitoring modules fail, the system relies on the monitoring system composed of various functional modules. In this case, the complexity of fault detection and fault management will increase significantly, which may lead to extended fault recovery time or improper fault handling.

[0012] Wo2006066507A1 discloses a method for implementing scene security functions, which does not support the adjustment of dynamic security requirements and has high requirements for pre-setting. The applied scene cannot be applied to unmanned driving in mines. Summary of the invention

[0013] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a scenario-based mine car unmanned driving safety control method and system, which can identify scenes of different complexity and risk levels in the mining area and ensure that the vehicle takes appropriate safety measures in different scenarios; In order to achieve the above purpose / solve the above technical problems, the present invention is implemented by adopting the following technical solutions: A first aspect: A scenario-based mine car unmanned driving safety control method, the method comprising: In response to task information instructions, real-time collection of mining scene data; Extract features based on scene data and task information, and divide the extracted features into macro scene layer and micro scene layer; Based on the task information and the macroscopic and microscopic scene layers, a decision tree model is used to make vehicle behavior decisions and output the driving path and target speed; According to the driving path and target speed combined with vehicle feedback information, control and output vehicle drive system, brake system, gear position, operating system and warning system control instructions; Based on the scenario risk assessment level and vehicle fault level, the system conducts control arbitration on the vehicle drive system, brake system, gear position, operating system and warning system control commands, and outputs safety control commands for the vehicle; The vehicle executes control instructions based on safety control instructions and the current vehicle status.

[0014] Optionally, the scene data includes position information, posture information, height information, point cloud data and camera images.

[0015] Optionally, the decision tree model adopts a feature matching and optimization method to construct a decision tree based on features, each branch is split according to a feature value, and is constructed by using global optimization and removing redundant branches and nodes.

[0016] Optionally, the step of using a decision tree model to make a vehicle behavior decision based on the task information and the macroscopic scene layer and the microscopic scene layer to output a driving path and a target speed includes: Use the rule engine to evaluate the scenario risk level based on the macro scenario layer and micro scenario layer; Dynamically adjust abnormal detection according to the scenario risk assessment level and output the highest level of all current faults; Based on the highest level of all current faults and the scenario risk assessment level, safety control instructions for the vehicle are output.

[0017] Optionally, the rule engine consists of a rule base, a rule interpreter and the definition of logical rules. The rule base stores predefined logical rules, which are composed of features of different scenarios in the mining area and are used to assess risk levels. The rule interpreter is used to interpret and execute the rules in the rule base, match corresponding rules based on current scenario feature data, and output preliminary risk assessment results.

[0018] Optionally, the definition of the logic rules adopts a priority principle, setting priorities for different rules, and when multiple rules are matched simultaneously, the rule with the highest priority is executed first.

[0019] A second aspect: A scenario-based unmanned driving safety control system for a mine car, the system comprising: A task management module, used to send task information to the scene perception and recognition module; The perception and positioning module is used to collect the scene data of the vehicle in real time according to the task information; The scene perception and recognition module is used to extract features based on scene data and task information, and divide the extracted features into macro scene layer and micro scene layer; The decision-making planning module is used to make vehicle behavior decisions based on the mission information and the macroscopic and microscopic scene layers using a decision tree model to output the driving path and target speed; A control module is used to control and output vehicle drive system, brake system, gear position, operating system and warning system control instructions according to the driving path and target speed combined with vehicle feedback information; Functional safety module, which is used to control and arbitrate the vehicle drive system, brake system, gear, operating system and warning system control instructions based on the scenario risk assessment level and vehicle fault level, and output safety control instructions for the vehicle; The wire control module is used for the vehicle to execute control instructions based on safety control instructions and the current vehicle status.

[0020] Optionally, the perception and positioning module hardware consists of a laser radar, a millimeter wave radar, a camera, an inertial navigation device and an antenna.

[0021] Optionally, the functional safety module includes a scenario risk assessment unit, an anomaly detection unit, a vehicle fault management unit and a safety arbitration unit; The scenario risk assessment unit uses a defined rule engine to perform scenario risk assessment based on the scenario classification results output by the scenario perception and classification module; the anomaly detection unit dynamically adjusts the detection of anomalies based on the scenario risk level output by the scenario risk assessment unit; the vehicle fault management unit is used to summarize all vehicle faults in real time and output the highest level of all current faults; the safety arbitration unit performs control arbitration based on the output results of the scenario risk assessment unit and the vehicle fault management unit, and outputs control instructions for the drive system, braking system, gear, operating system, and warning system.

[0022] Optionally, the task information at least includes the driving reference line and the position, speed, elevation, curvature and area attribute information of each reference point.

[0023] Compared with the prior art, the present invention has the following beneficial effects: The present invention can identify scenes of different complexity and risk levels in the mining area through a two-layer classification method, ensuring that vehicles take appropriate safety measures in different scenarios; The present invention collects and analyzes scene features, combines with a rule engine, dynamically evaluates the current risk level, and timely adjusts the vehicle's safety strategy to prevent potential dangers to the greatest extent; The present invention can dynamically adjust the safety strategy of the mine car according to the specific operating scenario in which the mine car is located, thereby improving the safety and reliability of the mine car in various complex mining operating environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 The figure shows a schematic diagram of an embodiment of a mine car operation scene of the present invention; Figure 2 Shown is a composition diagram of the perception and positioning module of the present invention; Figure 3 Shown is a structural diagram of an embodiment of the system of the present invention; Figure 4 Shown is a flow chart of an embodiment of the method of the present invention; In the figure: 1-base station; 2-mine car body; 3-control center; 4-excavator; 5-grader; 6-inertial navigation device; 7-5G communication equipment; 8-industrial computer; 9-laser radar; 10-millimeter wave radar; 11-camera; 12-wire control controller; DETAILED DESCRIPTION

[0025] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods. Example

[0026] like Figure 4 As shown, the present invention discloses a scenario-based unmanned driving safety control method for a mine car, the method comprising: S101, in response to the task information instruction, real-time collection of scene data of the mining area work; S102, extracting features according to the scene data and task information, and dividing the extracted features into a macro scene layer and a micro scene layer; S103, based on the task information and the macroscopic scene layer and the microscopic scene layer, a decision tree model is used to make a vehicle behavior decision and output a driving path and a target speed; S104, controlling and outputting vehicle drive system, brake system, gear position, operating system and warning system control instructions according to the driving path and target speed combined with vehicle feedback information; S105, based on the scenario risk assessment level and the vehicle fault level, control arbitration is performed on the vehicle drive system, brake system, gear position, operating system and warning system control instructions, and a safety control instruction for the vehicle is output; S106, the vehicle executes the control instruction according to the safety control instruction and the current vehicle state.

[0027] In the specific implementation process, step S101 collects the scene data of the mining area in real time, including position information, posture information, height information, point cloud data, camera images, etc.

[0028] In the specific implementation process, step S102 first extracts features based on the original data of the sensing and positioning module and the task data of the task management module. The features include any area attributes, location, size, shape, speed, relative position, etc. of the object.

[0029] After the scene perception and recognition module extracts features from the raw data, it enters the scene classification, which is based on two-layer scene classification, namely the macro scene layer and the micro scene layer. Scene classification is based on spatial scope, functional dimension, and risk dimension.

[0030] The macro-scenario layer classifies the large-scale environment of the mining area. The main scenes include but are not limited to loading area scenes, transportation area scenes, unloading area scenes, and special scenes.

[0031] The micro-scenario layer is responsible for classifying the details in the specific operating environment, paying more attention to local and immediate risk factors. The micro-scenario of the mining area scene mainly includes: queuing, loading operations, and parking at loading points; the micro-scenario of the transportation area scene mainly includes: flat ground driving, slope driving, curve driving, obstacle avoidance, meeting vehicles, etc.; the micro-scenario of the unloading area scene mainly includes: parking at unloading points, material unloading, cargo box towing, etc.; the micro-scenario of special scenes mainly includes: bad weather, temporary scheduling, forced driving, etc.

[0032] In the specific implementation process, the decision planning module in step S103 uses a decision tree model to make vehicle behavior decisions based on the scene and task information output by the scene perception and recognition module. The decision tree model is constructed using feature matching and optimization methods. A decision tree is constructed based on features, and each branch is split according to the feature value. Global optimization is used to remove redundant branches and nodes and simplify the tree structure.

[0033] First, define the nodes and branches of the decision tree. Each node represents a classification guidance item, and each branch represents a decision item. In the implementation case, the classification results of the macro scenario serve as the nodes of the decision tree model. The results of the macro scenario enter the next branch through the guidance of the decision tree, and each classification result of the micro scenario serves as a branch.

[0034] Secondly, by combining the classification results of macro and micro scenarios, the branches of the decision tree will output a comprehensive operational decision. In the implementation case, in the macro scenario of "transportation zone scenario", if the micro scenario identifies "oncoming vehicles", the system will output the decision of "slowing down to meet the vehicles".

[0035] Among them, the decision-making planning module is only responsible for task-related decisions, while fault-related decisions are the responsibility of the functional safety module.

[0036] In this embodiment, step S104 controls the gear position, drive system, brake system and steering system of the vehicle according to the driving path and target speed output by the decision-making planning module and the feedback information of the vehicle.

[0037] In this embodiment, the functional safety module of step S105 is composed of a scenario risk assessment unit, an anomaly detection unit, a vehicle fault management unit, and a safety arbitration unit. Among them, the scenario risk assessment unit uses a defined rule engine to perform scenario risk assessment based on the scenario classification results output by the scenario perception and classification module. In the implementation case, the scenario risk assessment is divided into 4 levels: high risk, medium risk, low risk, and no risk.

[0038] The rule engine consists of a rule base, a rule interpreter, and the definition of logical rules. The rule base stores a series of predefined logical rules, which are based on the characteristics of different scenarios in the mining area and are used to assess the risk level. The rule interpreter is responsible for interpreting and executing the rules in the rule base, matching the corresponding rules based on the current scenario feature data, and outputting preliminary risk assessment results. The definition of logical rules adopts the priority principle, setting priorities for different rules to ensure that when multiple rules are matched at the same time, the highest priority rule is executed first.

[0039] The anomaly detection unit dynamically adjusts the detection of anomalies according to the scene risk level output by the scene risk assessment unit. In the implementation case, the micro-scene "meeting" belongs to a high-risk level scene. The threshold for determining whether the vehicle deviates from the driving route will be reduced, and if a vehicle deviates from the driving route, it will be controlled to stop and avoid in time.

[0040] The vehicle fault management unit summarizes all vehicle faults in real time and outputs the highest level of all current faults.

[0041] The safety arbitration unit conducts control arbitration based on the output results of the scenario risk assessment unit and the vehicle fault management unit, dynamically adjusts safety measures, and outputs control instructions for the drive system, braking system, gear, operating system, and warning system.

[0042] In this embodiment, in step S106, the wire control module executes the control instruction according to the control instruction output by the functional safety module and the current vehicle status, and sends the real-time status of the vehicle to the functional safety module via the CAN bus.

[0043] The present invention can identify scenes of different complexity and risk levels in a mining area through a double-layer classification method, ensuring that vehicles take appropriate safety measures in different scenes.

[0044] Embodiment 2, as Figure 3As shown, the present invention discloses a scenario-based unmanned driving safety control system for a mine car, the system comprising: A task management module, used to send task information to the scene perception and recognition module; The perception and positioning module is used to collect the scene data of the vehicle in real time according to the task information; The scene perception and recognition module is used to extract features based on scene data and task information, and divide the extracted features into macro scene layer and micro scene layer; The decision-making planning module is used to make vehicle behavior decisions based on the mission information and the macroscopic and microscopic scene layers using a decision tree model to output the driving path and target speed; A control module is used to control and output vehicle drive system, brake system, gear position, operating system and warning system control instructions according to the driving path and target speed combined with vehicle feedback information; Functional safety module, which is used to control and arbitrate the vehicle drive system, brake system, gear, operating system and warning system control instructions based on the scenario risk assessment level and vehicle fault level, and output safety control instructions for the vehicle; The wire control module is used for the vehicle to execute control instructions based on safety control instructions and the current vehicle status.

[0045] like Figure 1 and Figure 2 As shown, in the specific implementation process, this embodiment at least includes: a base station 1, a mining car body 2, a control center 3, an excavator 4, a grader 5, a wire control controller 12, an industrial computer 8, a laser radar 9, a millimeter wave radar 10, a camera 11, an inertial navigation device 6, and a 5G communication device 7; like Figure 2 As shown, the hardware of the perception and positioning module consists of a laser radar 9, a millimeter wave radar 10, a camera 11, and an inertial navigation device 6, and is connected to an industrial computer 8, and the software runs on the industrial computer 8. The task management module is deployed on the server of the control center 3. The task management module and the industrial computer communicate through the 5G communication device 7. The perception and positioning module, the scene perception and recognition module, the decision-making and planning module, the control module, and the functional safety module are all deployed and run on the industrial computer 8. The wire control module is deployed and runs on the wire control controller 12. The wire control controller and the industrial computer are connected through the CAN bus, and the communication signals between the modules are sent and received through the base station 1; When the task management module issues a task to the excavator 4 or the grader 5, the scene perception and recognition module first extracts features based on the original data of the perception and positioning module and the task data of the task management module. Features include any area attributes, location, object size, shape, speed, relative position, etc.

[0046] After the scene perception and recognition module extracts features from the raw data, it enters the scene classification, which is based on two-layer scene classification, namely the macro scene layer and the micro scene layer. Scene classification is based on spatial scope, functional dimension, and risk dimension.

[0047] The macro-scenario layer classifies the large-scale environment of the mining area. The main scenes include but are not limited to loading area scenes, transportation area scenes, unloading area scenes, and special scenes.

[0048] The micro-scenario layer is responsible for classifying the details in the specific operating environment, paying more attention to local and immediate risk factors. The micro-scenario of the mining area scene mainly includes: queuing, loading operations, and parking at loading points; the micro-scenario of the transportation area scene mainly includes: flat ground driving, slope driving, curve driving, obstacle avoidance, meeting vehicles, etc.; the micro-scenario of the unloading area scene mainly includes: parking at unloading points, material unloading, cargo box towing, etc.; the micro-scenario of special scenes mainly includes: bad weather, temporary scheduling, forced driving, etc.

[0049] The decision planning module uses a decision tree model to make vehicle behavior decisions based on the scene and task information output by the scene perception and recognition module.

[0050] The decision tree model is constructed using feature matching and optimization methods. A decision tree is built based on features. Each branch is split according to the feature value. Global optimization is used to remove redundant branches and nodes and simplify the tree structure.

[0051] First, define the nodes and branches of the decision tree. Each node represents a classification guide item, and each branch represents a decision point. In the implementation case, the classification results of the macro scenario serve as the nodes of the decision tree model. The results of the macro scenario enter the next branch through the guidance of the decision tree, and each classification result of the micro scenario serves as a branch.

[0052] Secondly, by combining the classification results of macro and micro scenarios, the branches of the decision tree will output a comprehensive operational decision. In the implementation case, in the macro scenario of "transportation zone scenario", if the micro scenario identifies "oncoming vehicles", the system will output the decision of "slowing down to meet the vehicles". The decision planning module is only responsible for task-related decisions, and the functional safety module is responsible for fault-related decisions.

[0053] The control module controls the vehicle's gear, drive system, brake system and steering system according to the driving path and target speed output by the decision-making and planning module.

[0054] The functional safety module consists of a scenario risk assessment unit, anomaly detection unit, vehicle fault management unit, and safety arbitration unit. The scenario risk assessment unit uses a defined rule engine to perform scenario risk assessment based on the scenario classification results output by the scenario perception and classification module.

[0055] The anomaly detection unit dynamically adjusts the detection of anomalies according to the scene risk level output by the scene risk assessment unit. In the implementation case, the micro-scene "meeting" belongs to a high-risk level scene. The threshold for determining whether the vehicle deviates from the driving route will be reduced, and if a vehicle deviates from the driving route, it will be controlled to stop and avoid in time.

[0056] The vehicle fault management unit summarizes all vehicle faults in real time and outputs the highest level of all current faults.

[0057] The safety arbitration unit conducts control arbitration based on the output results of the scenario risk assessment unit and the vehicle fault management unit, and outputs control instructions for the drive system, braking system, gear, operating system, and warning system.

[0058] The wire control template executes control instructions according to the control instructions output by the functional safety module and the current vehicle status, and sends the real-time status of the vehicle to the functional safety module through the CAN bus.

[0059] The system of this embodiment can dynamically adjust its safety strategy according to the specific operating scenario of the mine car, thereby improving the safety and reliability of the mine car in a variety of complex mining operating environments.

[0060] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A scenario-based mine car unmanned driving safety control method, characterized in that: The method comprises: In response to task information instructions, real-time collection of mining scene data; Extract features based on scene data and task information, and divide the extracted features into macro scene layer and micro scene layer; Based on the task information and the macroscopic and microscopic scene layers, a decision tree model is used to make vehicle behavior decisions and output the driving path and target speed; According to the driving path and target speed combined with vehicle feedback information, control and output vehicle drive system, brake system, gear position, operating system and warning system control instructions; Based on the scenario risk assessment level and vehicle fault level, the system conducts control arbitration on the vehicle drive system, brake system, gear position, operating system and warning system control commands, and outputs safety control commands for the vehicle; The vehicle executes control instructions based on safety control instructions and the current vehicle status.

2. The scenario-based unmanned driving safety control method for a mine car according to claim 1 is characterized in that: The scene data includes position information, posture information, height information, point cloud data and camera images.

3. The scenario-based unmanned driving safety control method for a mine car according to claim 1 is characterized in that: The decision tree model adopts a feature matching and optimization method to construct a decision tree based on features, each branch is split according to the feature value, and is constructed by using global optimization and removing redundant branches and nodes.

4. The scenario-based unmanned driving safety control method for a mine car according to claim 1 is characterized in that: Based on the scenario risk assessment level and the vehicle fault level, the vehicle drive system, brake system, gear, operating system and warning system control instructions are arbitrated and the safety control instructions for the vehicle are output, including: Use the rule engine to evaluate the scenario risk level based on the macro scenario layer and micro scenario layer; Dynamically adjust abnormal detection according to the scenario risk assessment level and output the highest level of all current faults; Based on the highest level of all current faults and the scenario risk assessment level, safety control instructions for the vehicle are output.

5. The scenario-based unmanned driving safety control method for a mine car according to claim 4 is characterized in that: The rule engine consists of a rule base, a rule interpreter and the definition of logical rules. The rule base stores predefined logical rules, which are based on the characteristics of different scenarios in the mining area and are used to assess the risk level. The rule interpreter is used to interpret and execute the rules in the rule base, match the corresponding rules based on the current scenario feature data, and output preliminary risk assessment results.

6. The scenario-based unmanned driving safety control method for a mine car according to claim 5 is characterized in that: The definition of the logic rules adopts the priority principle, setting priorities for different rules, and when multiple rules are matched at the same time, the rule with the highest priority is executed first.

7. A scenario-based mine car unmanned driving safety control system, characterized in that: The system comprises: A task management module, used to send task information to the scene perception and recognition module; The perception and positioning module is used to collect the scene data of the vehicle in real time according to the task information; The scene perception and recognition module is used to extract features based on scene data and task information, and divide the extracted features into macro scene layer and micro scene layer; The decision-making planning module is used to make vehicle behavior decisions based on the mission information and the macroscopic and microscopic scene layers using a decision tree model to output the driving path and target speed; A control module is used to control and output vehicle drive system, brake system, gear position, operating system and warning system control instructions according to the driving path and target speed combined with vehicle feedback information; Functional safety module, which is used to control and arbitrate the vehicle drive system, brake system, gear, operating system and warning system control instructions based on the scenario risk assessment level and vehicle fault level, and output safety control instructions for the vehicle; The wire control module is used for the vehicle to execute control instructions based on safety control instructions and the current vehicle status.

8. The scenario-based unmanned driving safety control system for mining vehicles according to claim 7 is characterized in that: The hardware of the perception and positioning module is composed of a laser radar, a millimeter-wave radar, a camera, an inertial navigation device and an antenna.

9. The scenario-based unmanned driving safety control system for mining vehicles according to claim 7 is characterized in that: The functional safety module includes a scenario risk assessment unit, an anomaly detection unit, a vehicle fault management unit and a safety arbitration unit; The scenario risk assessment unit uses a defined rule engine to perform scenario risk assessment based on the scenario classification result output by the scenario perception and classification module; The anomaly detection unit dynamically adjusts the detection of anomalies according to the scenario risk level output by the scenario risk assessment unit. The vehicle fault management unit is used to summarize all vehicle faults in real time and output the highest level of all current faults. The safety arbitration unit performs control arbitration according to the output results of the scenario risk assessment unit and the vehicle fault management unit, and outputs control instructions for the drive system, braking system, gear, operating system, and warning system.

10. The scenario-based unmanned driving safety control system for mining vehicles according to claim 7 is characterized in that: The task information at least includes the driving reference line and the position, speed, elevation, curvature and regional attribute information of each reference point.

Images