Automatic driving vehicle real-time planning method based on sequential logic protocol

Abstract

The invention relates to an automatic driving vehicle real-time planning method based on a sequential logic protocol, and the method comprises the steps: carrying out the signal sequential logic language modeling of an expressway scene driving task based on a customized predicate function set, and obtaining a signal sequential logic protocol corresponding to the expressway scene driving task, the information is stored in a vehicle-mounted path planner; the vehicle-mounted path planner reads a signal sequential logic protocol instruction in a current driving state, and plans a track set according to the extracted time constraint, state constraint and vehicle state information of the driving task; and carrying out asynchronous time sequence robustness risk assessment on the trajectory set, and screening out an optimal trajectory. Compared with the prior art, the automatic driving system is more robust and reliable and can participate in traffic safely and effectively.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving path planning, and in particular to a real-time planning method for an autonomous driving vehicle based on temporal logic specifications. Background Art

[0002] The core challenge of robot learning is to enable robots to perform various tasks according to human instructions, which requires robots to have the ability to understand and execute diverse tasks. Large language models (LLMs) are trained on a large amount of Internet text through self-supervised learning and have become an important tool for robots to understand and execute natural language instructions.

[0003] The proposed general framework uses LLMs as "co-pilots" in driving scenarios to automatically perform certain driving tasks through natural language processing. However, specific AI robot systems still face challenges such as the connection between perception and symbolic representation, interactive learning between humans and machines, and the integration of tasks and motion planning. In autonomous driving systems, precise mathematical driving tasks and traffic rules are particularly important to bridge the gap between natural language instructions generated by LLMs and machine executable logic instructions.

[0004] Formal methods are increasingly being used to ensure the safety of autonomous driving systems. At present, the formal modeling of autonomous driving traffic rules and driving tasks mainly includes methods based on ontology and first-order logic. Ontology uses the Semantic Web Rule Language (SWRL) to qualitatively present rules, which is easy to modularize, but lacks quantitative descriptions of specific traffic scenarios and time information, which hinders the effective modeling of time-related tasks.

[0005] Modeling based on temporal logic can make up for the shortcomings of ontology modeling. Metric temporal logic (MTL) allows the specification of attribute time intervals. Related studies use MTL to quantitatively model the traffic rules of autonomous vehicles under the conditions of maintaining safe distances, speed control, and traffic convergence. Signal temporal logic (STL) is used to formalize the modeling of traffic rules and driving tasks. The robustness is quantitatively expressed by the system's satisfaction with the STL specification, and the robustness is calculated using the reinforcement time algorithm. Existing studies have proposed the concept of robustness trajectory and established STL specifications for intersection traffic rules, but the representation and modeling of driving tasks based on temporal logic are still insufficient.

[0006] Formal rules have a wide range of applications in trajectory planning and monitoring. We propose model predictive monitoring of STL specifications, which can be integrated into trajectory planners to generate rule-compliant trajectories and play a role in real-time trajectory evaluation and correction. Rules are crucial in computing the reachability set of autonomous vehicles, enabling planners to consider rule-compliant trajectories. Rulebooks facilitate decision-making by prioritizing traffic rules and resolving conflicts when one rule must be violated to satisfy another. Neural networks are used to predict traffic violations, enhance understanding of human behavior, and support better anticipation and response in autonomous driving systems. Although existing work has considered traffic rules for trajectory evaluation, research on robust evaluation of stronger temporal attribute specifications remains underexplored. Summary of the invention

[0007] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide a real-time planning method for autonomous driving vehicles based on temporal logic specifications. This method makes the autonomous driving system more robust and reliable, and can participate in traffic safely and effectively.

[0008] The purpose of the present invention can be achieved by the following technical solutions:

[0009] A real-time planning method for an autonomous driving vehicle based on temporal logic specification, the method comprising:

[0010] Based on the self-defined predicate function set, the signal timing logic language modeling of the highway scene driving task is carried out to obtain the signal timing logic specification corresponding to the highway scene driving task and store it in the vehicle path planner;

[0011] The vehicle-mounted path planner reads the signal timing logic protocol instructions under the current driving state, and extracts the time constraints, state constraints and vehicle state information of the driving task from the instructions;

[0012] The vehicle trajectory is planned according to the extracted time constraints, state constraints and vehicle state information of the driving task to obtain an initial trajectory set. The asynchronous timing robustness risk assessment is performed on the initial trajectory set to screen out the optimal trajectory.

[0013] Preferably, the signal timing logic language modeling of the highway scenario driving task based on the custom predicate function set is performed to obtain the signal timing logic specification corresponding to the highway scenario driving task, specifically: encoding is performed according to the custom predicate function set, the logical operators of the signal timing logic and the timing operators, the signal timing logic language modeling of the highway scenario driving task is performed to obtain the signal timing logic specification corresponding to the highway scenario driving task.

[0014] Preferably, the user-defined predicate function set includes:

[0015] Basic predicate functions, used to describe the relationship between the ego vehicle and the scene in terms of direction, lane, and rules;

[0016] Risk predicate function, used to quantitatively evaluate the asynchronous time robustness of the current signal timing logic specification;

[0017] Task predicate function is used to describe the execution logic of the driving task under the signal temporal logic specification.

[0018] Preferably, the basic predicate is a binary predicate, which is used to determine whether the corresponding signal timing logic specification is satisfied, including:

[0019] It means that the comfort is improved by reducing the speed variation, where G is the global operator and a is the vehicle acceleration;

[0020] Indicates whether the current vehicle speed v is less than or equal to the lane limit v limit ;

[0021] Indicates whether to cross only the dashed line or the solid line when changing lanes from the current lane to the target lane;

[0022] Indicates the target lane type, including import and export;

[0023] Indicates whether there are relevant traffic participants in the left front, left rear, right front, right rear, front and rear respectively;

[0024] Lon_safety_distance(ego,o front )=G(|s ego -s o |≥s min ): Indicates whether to maintain a safe longitudinal distance with the front and rear vehicles, where ego, o front Refers to the state parameters of the vehicle and the adjacent vehicles in front and behind, s ego 、s o They represent the longitudinal position of the vehicle and the surrounding vehicle respectively, s min is the minimum longitudinal distance;

[0025] Lat_safety_distance(ego,o)=G(|d ego -d o |≥d min ): Indicates whether to maintain a safe lateral distance with adjacent vehicles, where d ego d o are the lateral position of the vehicle and the surrounding vehicle, respectively, dmin is the minimum horizontal distance.

[0026] Preferably, the risk predicate function Risk_monitor(φ), for the signal timing logic specification φ, the asynchronous timing robustness is measured by the time difference between two vehicles arriving at the same location.

[0027] Preferably, the task predicate function comprises:

[0028] The predicate function of the overtaking task Suipass(ego,o front ):=s ego -s o ≥δ sur , among which ego, o front Refers to the state parameters of the vehicle and the preceding vehicle, s ego 、s o They represent the longitudinal position of the vehicle and the front vehicle respectively, δ sur is the overtaking distance threshold. When the longitudinal position of the vehicle exceeds that of the front vehicle at a certain moment, it means that the overtaking task is completed.

[0029] Lane change task predicate function Change_lane left / right :=|d ego,t -d target |≤δ d , where the subscripts left and right correspond to changing lanes to the left and right, respectively. ego ,d o are the lateral position of the vehicle and the surrounding vehicle, respectively, d min is the minimum lateral distance;

[0030] Emergency avoidance task predicate function Evasion;

[0031] Safety parking predicate function Pull_over:=v ego ==0.

[0032] The predicate function MaxBraking() that triggers the maximum braking deceleration.

[0033] Preferably, the vehicle trajectory is planned according to the extracted time constraint, state constraint and vehicle state information of the driving task to obtain an initial trajectory set, specifically:

[0034] The spatial position x and the task completion time t are sampled, and for a signal timing logic specification, a two-dimensional search space of horizontal time and vertical position is obtained. A series of discrete polynomial trajectories are generated in this two-dimensional search space, and the initial trajectory set is obtained after preliminary screening through the kinematic boundary.

[0035] Preferably, the spatial position x and the task completion time t are sampled, and for a signal timing logic specification, a two-dimensional search space of horizontal time and vertical position is obtained, and a series of discrete polynomial trajectories are generated in this two-dimensional search space. After preliminary screening through the kinematic boundary, an initial trajectory set is obtained, including:

[0036] For different signal timing logic specifications, select different state quantities [s1, s1′, s1″, T] ij =[s i ,s t ' arget ,s t ″ arget ,T j ], where i and j represent the index of the longitudinal position and the time to complete the task, respectively, and s t ' arget 、s t ″ arget represents the given target speed and acceleration values, s1, s1′, s1″, T represent the end point position, speed, acceleration and task completion time;

[0037] The cost function in the process of generating polynomial trajectories is the smoothing cost J t , the weighted sum of time cost g(T) and endpoint constraint h(X1).

[0038] Preferably, performing asynchronous timing robustness risk assessment on the initial trajectory set to screen out the optimal trajectory includes:

[0039] According to the vertical cost C lon and horizontal cost C lat and asynchronous timing robustness θ φ The weighted sum of constructs the joint cost function C tot , select the optimal trajectory from the initial trajectory set.

[0040] Preferably, after the optimal trajectory is screened out, the process further includes performing trajectory rolling planning, specifically:

[0041] At the beginning of each planning cycle, the onboard controller gives instructions to calculate the long-term trajectory and short-term risk assessment. If there is no collision in the short term and there is a safe parking trajectory in the long term, the feasibility of the current trajectory is verified and confirmed;

[0042] If a collision is expected to occur in a short time, the vehicle will use the backup trajectory that can be used for safe parking at the previous moment to decelerate, and then accelerate to restore to the original expected speed after passing verification at the next moment.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] (1) The real-time planning method for autonomous driving vehicles based on highway scenario temporal logic specifications proposed in this paper uses signal temporal logic language to formulate driving tasks and traffic rules in a precise and mathematical way, which can solve the key problem of the gap between natural language and machine-executable logical instructions. It defines asynchronous temporal robustness to quantify the risk of trajectory not meeting safety specifications, and integrates this metric into trajectory planning to improve the safety of planning.

[0045] (2) The temporal logic specifications corresponding to all driving tasks in highway scenarios constructed by the present invention cover the action instructions required for the autonomous driving system to handle various highway scenarios. The task coverage is complete and provides a sound framework for meeting the diverse needs of highway driving.

[0046] (3) Through rolling trajectory planning, if there is no collision in the short term and there is a safe parking trajectory in the long term, the verification is passed and the feasibility of the current trajectory is determined; if a collision occurs in the short term, the safe parking backup trajectory of the previous moment is used for deceleration, and then accelerated to the original expected speed after passing the verification at the next moment. This long-term and short-term risk calculation architecture effectively avoids overly conservative behavior while ensuring safety and collision-free. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is a framework diagram of the real-time planning method for an autonomous driving vehicle based on the highway scenario temporal logic specification of the present invention;

[0048] Figure 2 It is a definition diagram of asynchronous timing robustness risk in an embodiment of the present invention;

[0049] Figure 3 The drivable area is calculated based on the reachable set analysis in the embodiment of the present invention;

[0050] Figure 4 It is a rolling planning framework for the method of the present invention;

[0051] Figure 5 4 is the trajectory simulation result in the embodiment of the present invention. DETAILED DESCRIPTION

[0052] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0053] Example

[0054] like Figure 1As shown, this embodiment proposes a real-time planning method for an autonomous driving vehicle based on a highway scene temporal logic specification, and the method includes the following steps:

[0055] Based on the self-defined predicate function set, the signal timing logic language modeling of the highway scene driving task is carried out to obtain the signal timing logic specification corresponding to the highway scene driving task and store it in the vehicle path planner;

[0056] The vehicle-mounted path planner reads the signal timing logic protocol instructions under the current driving state, and extracts the time constraints, state constraints and vehicle state information of the driving task from the instructions;

[0057] The vehicle trajectory is planned according to the extracted time constraints, state constraints and vehicle state information of the driving task to obtain an initial trajectory set. The asynchronous timing robustness risk assessment is performed on the initial trajectory set to screen out the optimal trajectory.

[0058] Next, the method of this embodiment is introduced in detail.

[0059] Step 1: Establish a common protocol

[0060] General specifications refer to the regulations that autonomous vehicles must generally meet outside of the driving task, including:

[0061] ① Maintain a safe longitudinal distance (hard requirement): The vehicle must maintain a safe distance from the vehicle ahead in the same lane to prevent collision, even in emergency braking situations. This rule applies to every vehicle within the sensor range and must be evaluated accordingly.

[0062]

[0063] Among them: G represents the global operator, that is, the following conditions must be met at all times; It indicates logical implication; is the property of the preceding vehicle, which refers to its existence and its state of motion; ego, o front They respectively represent the state parameters of the vehicle and the vehicle in front; Lon_safe_distance() is the longitudinal safety distance predicate, which represents a mapping function from state parameters to Boolean variables.

[0064] ② Maintain lateral safety distance (hard requirement): The vehicle must maintain a safe distance from vehicles in adjacent lanes to minimize the risk of collision, especially when changing lanes or when vehicles in adjacent lanes make sudden moves. This rule applies to all vehicles within the sensor range and must be evaluated and followed at all times to ensure that the appropriate lateral safety distance is maintained.

[0065]

[0066] Among them: ∨ represents logical "or"; is the property of the adjacent vehicle, which refers to the existence and motion state of the adjacent vehicle; ego, o adiacent Respectively refer to the state parameters of the ego vehicle and the adjacent vehicle; Lat_safe_distance() is the lateral safety distance predicate. ③ Maximum speed limit (hard requirement): The ego vehicle must strictly abide by the specified maximum speed limit to ensure safe and legal operation. This rule applies to all driving conditions and road types. The speed of the vehicle should not exceed the specified limit and must be continuously monitored and adjusted to ensure compliance.

[0067]

[0068] in: is the speed property of the vehicle.

[0069] ④ Ensure comfort (soft requirements): The vehicle should minimize the discomfort caused by sudden acceleration, deceleration or violent operation to provide a comfortable riding experience.

[0070]

[0071] in: It is the comfort attribute of the vehicle.

[0072] Step 2: Extract possible driving tasks from the highway scenario. Driving tasks usually specify a series of actions, so temporal logic is a natural choice for the formal expression of driving tasks. Our formal specification is based on high-level decision instructions given by the upper-level decision system. Based on the general specification, the invention defines a comprehensive set of basic driving tasks for highway scenarios. These driving tasks are divided into two categories: tasks with reference targets and tasks without reference target objects. As shown in Table 1, these tasks cover key operations such as overtaking, avoiding, following, changing lanes, cruising, safe parking, entrance ramps and exit ramps. By including these tasks, the present invention ensures that the autonomous driving system has the action instructions required to handle various highway scenarios. This shows the completeness of the architecture in terms of task coverage, providing a sound framework for addressing the diverse needs of highway driving.

[0073] Step 3: From natural language description of driving tasks to signal temporal logic language.

[0074] As shown in Table 1 below, the signal timing logic STL language modeling is performed for the highway scene driving task.

[0075] Table 1

[0076]

[0077] The first category, tasks with reference targets and their formal specifications are as follows:

[0078] ① Overtaking task: The action of the ego vehicle moving from the current lane to the adjacent lane to overtake the slower vehicle ahead. This task is subject to specific rules and considerations to ensure the safety and efficiency of the lane change operation.

[0079] The STL specification is as follows:

[0080]

[0081] in:

[0082] Indicates that there is a target vehicle in front of the vehicle.

[0083] Stay_in_lane U [0,t] Safe_to_change_lane: The vehicle should stay in the current lane until time t when the Safe_to_change_lane condition should be met. U is the Until operator, which means to maintain the Stay_in_lane state until time t.

[0084] This is a logical implication that means that if the condition on the left (there is a car ahead, stay in the lane until it is safe to change lanes) is true, then the condition on the right should also be true.

[0085] This shows that within the time interval [0, t1], Change_lane left An event must occur which means the vehicle should change to the left lane.

[0086] This shows that in the time interval [0,Tt], the event Surpass(ego,o front ) must happen, Change_lane left The event is completed in the time interval [0, t1].

[0087] In general, the overtaking rule can be interpreted as follows: if there is a vehicle in front, the ego vehicle must stay in its lane until a safe lane change can be made, then within the time interval [0,Tt], the ego vehicle should overtake the vehicle in front and complete the left lane change within the range [0,t1].

[0088] ② Emergency avoidance tasks: In the field of autonomous vehicles, emergency avoidance is crucial, especially when the vehicle in front suddenly slows down, the vehicle behind approaches quickly, or the vehicle accidentally enters the lane. These examples emphasize that the autonomous driving system must demonstrate agile response capabilities when faced with unexpected events and skillfully implement emergency avoidance maneuvers.

[0089] The STL specification is as follows:

[0090] 1) Vehicle cut-in: This is defined as the lateral distance to the vehicle in front not meeting the safe distance, and the speed of the own vehicle is faster than that of the vehicle in front, exceeding the predetermined speed difference threshold.

[0091]

[0092] Where: s o represents the longitudinal position of the pericycle, s ego represents the longitudinal position of the vehicle; v o represents the velocity property of the circumvehicle, v ego represents the speed attribute of the vehicle; δ s , δ v Represents the thresholds of distance and speed respectively. When a vehicle suddenly cuts into the lane of the vehicle, the vehicle will brake at the maximum deceleration until the distance to the vehicle in front is restored to a safe level.

[0093]

[0094] Among them: MaxBraking() represents the predicate function that triggers the maximum braking deceleration.

[0095] 2) Emergency braking of the front vehicle or approaching of the rear vehicle: The scenario where the front vehicle suddenly decelerates and the rear vehicle approaches quickly can be defined by the STL formula:

[0096]

[0097] Among them: a o represents the acceleration property of the pericycle, a ego Represents the acceleration property of the vehicle; δ a They respectively represent the acceleration thresholds.

[0098] If the risk monitoring module is triggered based on the conditions of the status of the ego vehicle and surrounding traffic participants, the system must ensure that evasive action is taken at some point in the corresponding time interval.

[0099]

[0100] Among them: Risk_monitor() represents the predicate function of risk detection.

[0101] ③Following task: The vehicle automatically maintains a safe distance from the vehicle in front on the highway while maintaining a stable cruising speed.

[0102] The STL specification is as follows:

[0103]

[0104] The second category, tasks without reference targets and their formal specifications are as follows:

[0105] ① Lane change task: The vehicle changes from the current lane to the adjacent lane.

[0106] The STL specification is as follows:

[0107]

[0108] ②Cruise mission: The action of maintaining the vehicle’s position within a specific lane while driving on the road.

[0109] The STL specification is as follows:

[0110]

[0111] Where: v suggestion Indicates recommended speed.

[0112] ③Safe pull-over: The vehicle pulls out of the main traffic flow and stops at the roadside or designated parking area. It is usually used to respond to various situations, such as emergencies, traffic violations, or to allow other vehicles to pass. When performing this task, the vehicle follows a series of steps to safely and efficiently transition from active driving to a stationary state.

[0113] The STL specification is as follows:

[0114]

[0115] ④Task 4Entrance (Exit) Ramp: The entry and exit of the ramp can be regarded as a lane change process, where the vehicle transitions from the main road to the ramp and vice versa.

[0116]

[0117] Step 4: Extract predicates and functions from the signal temporal logic specification. Signal temporal logic is a formal logic language defined on predicates. It is necessary to define predicates and functions that formalize the driving task. The basic elements are systematically organized and defined through a task-level hierarchy.

[0118] The present invention defines predicate functions from three levels. According to the self-defined predicate function set, the signal timing logic specification corresponding to each driving task can be constructed.

[0119] First category: basic predicates

[0120] The satisfaction of a binary predicate is determined by a feature function that outputs a 0-1 variable to indicate whether the STL formula is satisfied under quantitative semantics. The basic predicates describe the relationship between the ego-vehicle and the scene in three aspects: direction, lane, and rules. Due to space limitations, we cannot list all predicates here, however, a scene can be fully defined in this way by a finite number of predicates.

[0121] ① It means that comfort is improved by reducing speed changes;

[0122] ② Indicates whether the speed limit is met;

[0123] ③ Indicates whether to cross only the dotted line (solid line) when changing lanes from the current lane to the target lane;

[0124] ④ Indicates the type of the target lane;

[0125] ⑤ These six predicates indicate whether the relevant traffic participants exist in the left front, left rear, right front, right rear, front, and rear, respectively;

[0126] ⑥Lon_safety_distance(ego,o front )=G(|s ego -s o |≥s min ): Indicates whether to maintain a safe distance from the vehicles in front and behind;

[0127] ⑦Lat_safety_distance(ego,o)=G(|d ego -d o |≥d min ): Indicates whether to maintain a safe distance from adjacent vehicles.

[0128] Where: v limit is the speed limit of the corresponding lane; s min ,d min They are the minimum longitudinal distance and the minimum lateral distance, respectively, and their calculation refers to the responsibility-sensitive safety model.

[0129] The second category, risk predicates

[0130] The Risk_monitor(φ) function is used to quantitatively evaluate the asynchronous time robustness of the specification φ. By comparing with a predetermined threshold, the spatiotemporal satisfaction of the specification φ can be determined. Figure 3,In the given ramp example, the convention ensures that the red and blue cars do not collide.,Asynchronous timing robustness is measured by the time difference between the two cars arriving at the same,location.

[0131] The third category, task predicate function

[0132] The task function clarifies the specific implementation and execution logic of the driving action under the STL specification. It inputs the time task and state constraints specified in the STL into the controller and planner. Although real-time trajectory solving based on STL constraints has not yet been achieved, it is feasible to evaluate the task satisfaction based on the STL specification and generate a trajectory that meets the constraints.

[0133] ① The predicate function of the overtaking task Surpass(ego,o front ):=s ego,t -s o,t ≥δ sur , among which ego, o front Refers to the state parameters of the vehicle and the preceding vehicle, s ego 、s o They represent the longitudinal position of the vehicle and the front vehicle respectively, δ sur is the overtaking distance threshold.

[0134] Determine the termination condition of the autonomous driving task and input the condition into the onboard trajectory planner. When the longitudinal position of the self-vehicle exceeds the front vehicle at a certain moment, it means that the overtaking task is completed.

[0135] ② Lane change task predicate function Change_lane left / right :=|d ego,t -d target |≤δ d , where the subscripts left and right correspond to changing lanes to the left and right, respectively. ego ,d o are the lateral position of the vehicle and the surrounding vehicle, respectively, d min is the minimum horizontal distance.

[0136] When a left lane change is required, the target lateral position constraint is input into the planner and the system then executes the left lane change task.

[0137] ③Safety parking predicate function Pull_over:=v ego,t = = 0

[0138] Enables the driver to quickly and safely pull over to the side of the road in an emergency. The system detects the driver's intention to pull over and assesses the surroundings to identify a safe parking location for the vehicle.

[0139] ④Emergency avoidance task predicate function Evasion:

[0140] Use ensemble analysis to predict the movement trends of other traffic participants, such as Figure 3 As shown, the predicted occupied areas of the surrounding traffic participants are constructed, from which the drivable area is extracted.

[0141] Step 5: Path planning.

[0142] This embodiment adopts a path planner based on asynchronous timing robustness risk assessment, which is provided with a signal timing logic specification set encoded according to a user-defined predicate function set, a signal timing logic logical operator, and a timing operator. By reading the signal timing logic language instructions under the current driving state, according to the time constraints, state constraints and vehicle state information of the driving task extracted therefrom, the vehicle trajectory is planned according to the time constraints, state constraints and vehicle state information of the driving task extracted therefrom, an initial trajectory set is obtained, and the asynchronous timing robustness risk assessment is performed on the initial trajectory set to screen out the optimal trajectory and complete the path planning task.

[0143] The trajectory generation function is used for autonomous driving path planning. Given the initial position and the time constraints of the driving task, an optimal path that meets safety (no collision), efficiency, and STL specifications is generated. A common method for trajectory generation is to use a polynomial function to represent the trajectory.

[0144] Sampling the spatial position x and the task completion time t reduces the search amount compared to the method of sampling all state spaces. In general, for an STL specification, the planner will eventually get a two-dimensional search space of time and longitudinal position, generate a series of discrete polynomial trajectories in this space, and then perform preliminary screening through the kinematic boundaries. Of course, different state quantities can be selected for different STL specifications.

[0145] [s1,s1′,s1″,T] ij =[s i ,s t ' arget ,s t ″ arget ,T j ] (16)

[0146] Where: i, j represent the index of the vertical position and the time to complete the task respectively; s t ' arget 、s t ″ arget represents the given target speed and acceleration value; s1, s1′, s1″, T represent the end point position, speed, acceleration and task completion time.

[0147] Considering the smoothing cost J t, time cost g(T) and endpoint constraint h(X1), construct the cost function.

[0148] C=k j J t +k t g(T)+k f h(X1) (17)

[0149] Where: k j , k t , k f is the weight coefficient.

[0150] Considering the smoothness cost, time cost and endpoint constraint, a cost function is constructed.

[0151] C=k j J t +k t g(T)+k f h(X1) (17)

[0152] Where: k j , k t , k f is the weight coefficient.

[0153] The joint cost function is the weighted sum of the longitudinal cost, the transverse cost and the asynchronous temporal robustness, while the calculation of the asynchronous temporal robustness requires the conversion of coordinates from the Frenet coordinate system to the Cartesian coordinate system. By combining the objective function, the trajectory set is evaluated and the optimal trajectory is selected.

[0154] C tot =k lat C lat +k lon C lon +k risk θ φ (18)

[0155] Figure 5 The trajectory generation process for the overtaking specification is outlined. At the initial time step, the overtaking command, state vector, and prediction horizon are received. First, a function is called to obtain the target region and constraints. Subsequently, a set of trajectories are generated by sampling. The last step is to calculate the cost by checking the comfort score, time score, and asynchronous time robustness according to the STL specification.

[0156] Step 6: Design a rolling planning framework. This embodiment designs a Figure 4The planning cycle shown in Figure 1 is designed to follow the STL instructions provided by the decision system and perform rolling planning. The goal of the planning process is to generate a trajectory that meets the specified STL requirements while ensuring safety and efficiency. At the beginning of each planning cycle, three decision instructions (such as Figure 4 , c = 1, Task1, Task2, Task3 three decision instructions are received), long-term (such as 5 seconds) trajectory and short-term (such as 1.5 seconds) risk assessment are calculated. If there is no collision in the short term and there is a safe parking trajectory in the long term, the feasibility of the AI ​​trajectory is verified and confirmed; if a collision occurs in the short term, the safe parking backup trajectory of the previous moment is used for deceleration, and then accelerated to the original expected speed after passing the verification at the next moment. This long-term and short-term risk calculation architecture effectively avoids overly conservative behavior while ensuring safety and collision-free.

[0157] Step 7: Design a simulation platform for the planning framework proposed by the present invention. The present invention develops a simulation framework to implement path planning for autonomous vehicles operating on dynamic highways. Figure 5 It is shown how to convert a driving task specified in Signal Timing Logic (STL) into an accurate and executable trajectory. The effectiveness of the proposed method is verified by effectively applying the STL specification and leveraging predefined predicates.

[0158] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A real-time planning method for an autonomous driving vehicle based on temporal logic specification, characterized in that: The method includes: Based on the self-defined predicate function set, the signal timing logic language modeling of the highway scene driving task is carried out to obtain the signal timing logic specification corresponding to the highway scene driving task and store it in the vehicle path planner; The vehicle-mounted path planner reads the signal timing logic protocol instructions under the current driving state, and extracts the time constraints, state constraints and vehicle state information of the driving task from the instructions; The vehicle trajectory is planned according to the extracted time constraints, state constraints and vehicle state information of the driving task to obtain an initial trajectory set. The asynchronous timing robustness risk assessment is performed on the initial trajectory set to screen out the optimal trajectory.

2. The method for real-time planning of an autonomous driving vehicle based on temporal logic specification according to claim 1, characterized in that: The signal timing logic language modeling of the highway scenario driving task based on the custom predicate function set is performed to obtain the signal timing logic specification corresponding to the highway scenario driving task, specifically: encoding is performed according to the custom predicate function set, the logical operators of the signal timing logic and the timing operators, the signal timing logic language modeling of the highway scenario driving task is performed to obtain the signal timing logic specification corresponding to the highway scenario driving task.

3. The method for real-time planning of an autonomous driving vehicle based on temporal logic specification according to claim 1, characterized in that: The custom predicate function set includes: Basic predicate functions, used to describe the relationship between the ego vehicle and the scene in terms of direction, lane, and rules; Risk predicate function, used to quantitatively evaluate the asynchronous time robustness of the current signal timing logic specification; Task predicate function is used to describe the execution logic of the driving task under the signal temporal logic specification.

4. The method for real-time planning of an autonomous driving vehicle based on temporal logic specification according to claim 3, characterized in that: The basic predicate is a binary predicate used to determine whether the corresponding signal timing logic specification is satisfied, including: It means that the comfort is improved by reducing the speed variation, where G is the global operator and a is the vehicle acceleration; Indicates whether the current vehicle speed v is less than or equal to the lane limit v limit ; Indicates whether to cross only the dashed line or the solid line when changing lanes from the current lane to the target lane; Indicates the target lane type, including import and export; Indicates whether there are relevant traffic participants in the left front, left rear, right front, right rear, front and rear respectively; Lon_safety_distance(ego,o front )=G(|s ego -s o |≥s min ): Indicates whether to maintain a safe longitudinal distance with the front and rear vehicles, where ego, o front Refers to the state parameters of the vehicle and the adjacent vehicles in front and behind, s ego 、s o They represent the longitudinal position of the vehicle and the surrounding vehicle respectively, s min is the minimum longitudinal distance; Lat_safety_distance(ego,o)=G(|d ego -d o |≥d min ): Indicates whether to maintain a safe lateral distance with adjacent vehicles, where d ego d o are the lateral positions of the vehicle and the surrounding vehicle, respectively, and d min is the minimum horizontal distance.

5. The method for real-time planning of an autonomous driving vehicle based on temporal logic specification according to claim 3, characterized in that: The risk predicate function Risk_monitor(φ), for the signal timing logic specification φ, the asynchronous timing robustness is measured by the time difference between two vehicles arriving at the same location.

6. The method for real-time planning of an autonomous driving vehicle based on temporal logic specification according to claim 3, characterized in that: The task predicate function includes: The overtaking task predicate function Suipass(ego,o front ):=s ego -s o ≥δ sur , among which ego, o front Refers to the state parameters of the vehicle and the preceding vehicle, s ego 、s o They represent the longitudinal position of the vehicle and the front vehicle respectively, δ sur is the overtaking distance threshold. When the longitudinal position of the vehicle exceeds that of the front vehicle at a certain moment, it means that the overtaking task is completed. Lane change task predicate function Change_lane left / right :=|d ego,t -d target |≤δ d , where the subscripts left and right correspond to changing lanes to the left and right, respectively. ego d o are the lateral positions of the vehicle and the surrounding vehicle, respectively, and d min is the minimum lateral distance; Emergency avoidance task predicate function Evasion; Safety parking predicate function Pull_over:=v ego ==0. The predicate function MaxBraking() that triggers the maximum braking deceleration.

7. The method for real-time planning of an autonomous driving vehicle based on temporal logic specification according to claim 1, characterized in that: The vehicle trajectory is planned according to the extracted time constraint, state constraint and vehicle state information of the driving task to obtain an initial trajectory set, which is specifically: The spatial position x and the task completion time t are sampled, and for a signal timing logic specification, a two-dimensional search space of horizontal time and vertical position is obtained. A series of discrete polynomial trajectories are generated in this two-dimensional search space, and the initial trajectory set is obtained after preliminary screening through the kinematic boundary.

8. The method for real-time planning of an autonomous driving vehicle based on temporal logic specification according to claim 7, characterized in that: The spatial position x and the task completion time t are sampled, and for a signal timing logic specification, a two-dimensional search space of horizontal time and vertical position is obtained. A series of discrete polynomial trajectories are generated in this two-dimensional search space, and an initial trajectory set is obtained after preliminary screening through the kinematic boundary, including: For different signal timing logic specifications, select different state quantities [s1, s1′, s1″, T] ij =[s i ,s t ' arget ,s t ″ arget ,T j ], where i and j represent the index of the longitudinal position and the time to complete the task, respectively, and s t ' arget 、s t ″ arget represents the given target speed and acceleration values, s1, s1′, s1″, T represent the end point position, speed, acceleration and task completion time; The cost function in the process of generating polynomial trajectories is the smoothing cost J t , the weighted sum of time cost g(T) and endpoint constraint h(X1).

9. The method for real-time planning of an autonomous driving vehicle based on temporal logic specification according to claim 7, characterized in that: The asynchronous timing robustness risk assessment is performed on the initial trajectory set to screen out the optimal trajectory, including: According to the vertical cost C lon and horizontal cost C lat and asynchronous timing robustness θ φ The weighted sum of constructs the joint cost function C tot , select the optimal trajectory from the initial trajectory set.

10. The method for real-time planning of an autonomous driving vehicle based on temporal logic specification according to claim 1, characterized in that: After the optimal trajectory is screened out, the trajectory rolling planning is also performed, specifically: At the beginning of each planning cycle, the onboard controller gives instructions to calculate the long-term trajectory and short-term risk assessment. If there is no collision in the short term and there is a safe parking trajectory in the long term, the feasibility of the current trajectory is verified and confirmed; If a collision is expected to occur in a short time, the vehicle will use the backup trajectory that can be used for safe parking at the previous moment to decelerate, and then accelerate to restore to the original expected speed after passing verification at the next moment.

Images