Speed planning method, speed planning device, vehicle and readable storage medium

Abstract

A speed planning method, a speed planning device, a vehicle and a readable storage medium. The method comprises: on the basis of a state transition equation for a vehicle, constructing a speed planning model from a start point to an end point; on the basis of the speed planning model, planning a first speed curve from the current state to an end-point state, and determining whether the first speed curve meets a speed upper limit constraint; if the first speed curve does not meet the speed upper limit constraint, expanding the current state, selecting a target candidate state from candidate states obtained from expansion, using same as the current state, and returning to execute the step of: on the basis of the speed planning model, planning the first speed curve from the current state to the end-point state; and if the first speed curve meets the speed upper limit constraint, obtaining a speed planning result on the basis of the first speed curve.

Description

Speed ​​planning methods, speed planning equipment, vehicles, and readable storage media

[0001] Cross-reference to related applications

[0002] This disclosure claims priority to Chinese Patent Application No. 202410799384.1, filed on June 19, 2024, entitled "Speed ​​Planning Method, Speed ​​Planning Device, Vehicle and Readable Storage Medium", the entire contents of which are incorporated herein by reference. Technical Field

[0003] This disclosure relates to the field of intelligent driving technology, and in particular to a speed planning method, speed planning device, vehicle, and readable storage medium. Background Technology

[0004] Speed ​​planning is one of the key technologies in the field of intelligent driving, which determines the comfort and safety of a vehicle in intelligent driving mode.

[0005] Improving the adaptability of speed planning and saving computing power and time are technical problems that urgently need to be solved by those skilled in the art. Summary of the Invention

[0006] In view of this, the purpose of this disclosure is to provide a speed planning method, speed planning device, vehicle, and readable storage medium to improve the adaptability of speed planning and save computing power and time.

[0007] To achieve the above objectives, this disclosure provides the following technical solution:

[0008] A speed planning method includes: constructing a speed planning model from a starting point to an ending point based on the vehicle's state transition equation; planning a first speed curve from the current state to the ending state based on the speed planning model; determining whether the first speed curve satisfies the speed upper limit constraint; if not, expanding the current state, selecting a target candidate state from the expanded candidate states as the current state, and returning to the step of planning the first speed curve from the current state to the ending state based on the speed planning model; if yes, obtaining the speed planning result based on the first speed curve.

[0009] Optionally, the speed limit constraint is a function of the travel distance.

[0010] Optionally, the velocity planning model includes a range of jerk values;

[0011] The speed planning method further includes: determining candidate behavior vectors based on the range of jerk values, wherein the candidate behavior vectors include jerk and the duration of the jerk;

[0012] Expanding the current state and selecting a target candidate state as the current state from the expanded candidate states includes: expanding the current state according to the candidate behavior vector, planning a second velocity curve according to the obtained candidate states, determining the target candidate state according to the second velocity curve and the velocity upper limit constraint, and determining the target candidate state as the current state.

[0013] Optionally, determining the target candidate state based on the second velocity curve and the velocity upper limit constraint includes: determining the candidate state corresponding to the candidate behavior vector that satisfies the velocity upper limit constraint and has the largest acceleration as the target candidate state.

[0014] Optionally, before determining the target candidate state as the current state, the method further includes: determining whether the planned driving distance of the vehicle exceeds the destination driving distance based on the target candidate state; if yes, obtaining a speed planning result based on the target candidate state; if no, performing the step of determining the target candidate state as the current state.

[0015] Optionally, it also includes: arranging the candidate behavior vectors in descending order of accelerometer speed;

[0016] The process of expanding the current state based on candidate behavior vectors, planning a second velocity curve based on the obtained candidate states, and determining a target candidate state based on the second velocity curve and the velocity upper limit constraint includes: taking the first candidate behavior vector as the current candidate behavior vector, expanding the current state based on the current candidate behavior vector to obtain a candidate state; planning a second velocity curve based on the candidate state and the velocity planning model, and determining whether the second velocity curve satisfies the velocity upper limit constraint; if yes, then determining the candidate state corresponding to the current candidate behavior vector as the target candidate state; if no, then taking the next candidate behavior vector as the current candidate behavior vector, and returning to the step of expanding the current state based on the current candidate behavior vector.

[0017] Optionally, after expanding the current state based on the current candidate behavior vector, the method further includes: determining whether the obtained candidate state satisfies the speed upper limit constraint; if yes, then performing the step of planning a second speed curve based on the candidate state and the speed planning model; if no, then performing the step of taking the next candidate behavior vector as the current candidate behavior vector.

[0018] Optionally, when the current candidate behavior vector is the last candidate behavior vector in the sequence, if it is determined that the second speed curve does not meet the speed upper limit constraint, the method further includes: determining that the planning has failed and outputting a planning failure message.

[0019] Optionally, after constructing a speed planning model from the starting point to the ending point based on the vehicle's state transition equation, the method further includes: constructing a steady-constant constraint planning model for the distance boundary based on the speed planning model; the steady-constant constraint planning model for the distance boundary includes an objective function with the shortest time, and the constraints of the steady-constant constraint planning model for the distance boundary include a starting boundary, an ending boundary, a speed limit, an acceleration limit, and the range of acceleration values, wherein the upper speed limit in the speed limit is a fixed value;

[0020] Planning a first velocity curve from the current state to the endpoint state based on the velocity planning model includes: planning a first velocity curve with the shortest time from the current state to the endpoint state based on the steady constraint planning model of the distance boundary.

[0021] Optionally, after constructing a speed planning model from the starting point to the ending point based on the vehicle's state transition equation, the method further includes: constructing a steady-constraint planning model for the speed boundary based on the speed planning model; the steady-constraint planning model for the speed boundary includes an objective function with the shortest time, and the constraints of the steady-constraint planning model for the speed boundary include the starting speed, the starting acceleration, the ending speed, the ending acceleration, the acceleration limit, and the range of the acceleration value;

[0022] The second velocity curve is planned based on the candidate state, including: using the steady constraint planning model of the velocity boundary to plan the second velocity curve with the shortest time from the candidate state to the stopping state.

[0023] Optionally, before planning the second velocity curve with the shortest time from the candidate state to the stopping state using the steady-constraint programming model of the velocity boundary, the method further includes: constructing an acceleration boundary velocity programming model, solving the acceleration boundary velocity programming model, and obtaining a first solution result; the acceleration boundary velocity programming model includes an objective function with the shortest time, and the constraints of the acceleration boundary velocity programming model include the starting acceleration, the ending acceleration, and the range of the jerk value, wherein the starting acceleration, the ending acceleration, and the upper and lower limits of the jerk are all constants;

[0024] Planning a second speed curve with the shortest time from the candidate state to the parking state using a steady-constraint programming model of the speed boundary includes: solving the steady-constraint programming model of the speed boundary based on the first solution result to obtain a second solution result; and obtaining the second speed curve with the shortest time from the candidate state to the parking state based on the candidate state and the second solution result.

[0025] Optionally, planning a first velocity curve that minimizes the time required to reach the endpoint state from the current state, based on the steady-constrained programming model of the distance boundary, includes: solving the steady-constrained programming model of the distance boundary based on the second solution result to obtain a third solution result; and obtaining the first velocity curve that minimizes the time required to reach the endpoint state from the current state based on the current state and the third solution result.

[0026] Optionally, the velocity planning model is:

[0027] Among them, A i Let Δt be the vector of the i-th row. i Let j be the duration of the i-th action vector. i Let be the acceleration of the vehicle in the i-th action vector. Let N be the set of real numbers, N-1 be the number of behavior vectors, N be the number of vehicle states, and j be the number of vehicle states. max X is the upper limit of jerk in velocity planning. i =[t i ,s i ,v i ,a i ] T Let s be the i-th state vector of the vehicle. i For the vehicle at t i Distance traveled at any given time, v i For the vehicle at t i The speed of time, a i For the vehicle at t i acceleration at time X i+1 =f(X) i A i Let be the state transition equation for the vehicle, where s1, v1, and a1 are variables representing the distance traveled, speed, and acceleration at the starting point, respectively. S v S a S Given the distance traveled, speed, and acceleration determined at the starting point, s N v N a N Let be the variables representing the distance traveled, speed, and acceleration at the destination, s E v E a E Let v be the distance traveled, speed, and acceleration determined at the destination, where 0 ≤ v i ≤v max (s i () represents the speed limit, v max (s i ) represents the distance traveled in s i The maximum speed at time amin ≤a i ≤a max For acceleration limitation, a min For minimum acceleration, a max This is the maximum acceleration.

[0028] Optionally, the time-constant programming model for the distance boundary is:

[0029] Among them, t N The time to reach the final state.

[0030] Optionally, the steady-constant programming model for the velocity boundary is:

[0031] Optionally, the acceleration boundary velocity planning model is as follows:

[0032] A speed planning device includes: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of the speed planning method as described in any of the preceding claims.

[0033] A vehicle comprising the speed planning device described above.

[0034] A readable storage medium storing a computer program that, when executed by a processor, implements the steps of the speed planning method as described in any of the preceding claims.

[0035] This disclosure provides a speed planning method, speed planning device, vehicle, and readable storage medium. The method includes: constructing a speed planning model from a starting point to an ending point based on the vehicle's state transition equation; planning a first speed curve from the current state to the ending point based on the speed planning model, and determining whether the first speed curve satisfies the speed upper limit constraint; if not, expanding the current state, selecting a target candidate state from the expanded candidate states as the current state, and returning to the step of planning the first speed curve from the current state to the ending point based on the speed planning model; if yes, obtaining the speed planning result based on the first speed curve.

[0036] The technical solution disclosed herein constructs a speed planning model from the starting point to the ending point based on the vehicle's state transition equation. Based on this model, a first speed curve from the current state to the ending state is planned, and it is determined whether the first speed curve satisfies the speed upper limit constraint. If the first speed curve does not satisfy the speed upper limit constraint, the current state is expanded, and a target candidate state is selected as the current state from the expanded candidate states. The process then returns to the step of planning the first speed curve. This process ensures that each expansion layer corresponds to only one target candidate state as the current state, meaning that the number of calculations for each expansion layer is at most equal to the number of expansions for the current state. This makes the computational complexity a linear function of the path length, thereby reducing computational complexity, computation time, and computational power consumption, and facilitating accurate speed planning results. If the first speed curve satisfies the speed upper limit constraint, the planning is considered successful, and the speed planning result is obtained based on the first speed curve, greatly reducing computation time and computational power consumption, and obtaining accurate speed planning results. This process achieves a smooth speed curve that satisfies the smoothness constraint by recursively traversing the path once from the starting point, thus reducing the computational power and time consumption of speed planning. Furthermore, the speed planning process described above does not require a predefined speed curve and can perform speed planning for any scenario, thereby improving the adaptability of speed planning.

[0037] Additional aspects and advantages of this disclosure will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this disclosure. Attached Figure Description

[0038] Figure 1 is a flowchart of a speed planning method provided in an embodiment of this disclosure;

[0039] Figure 2 is a flowchart of another speed planning method provided in an embodiment of this disclosure;

[0040] Figure 3 is a flowchart of another speed planning method provided in an embodiment of this disclosure;

[0041] Figures 4(a)-(f) are schematic diagrams of the periodic decomposition of speed planning provided in the embodiments of this disclosure;

[0042] Figures 5(a)-(e) are schematic diagrams of five solutions to the steady constraint programming model of the velocity boundary provided in the embodiments of this disclosure;

[0043] Figures 6(a)-(c) are schematic diagrams of three solutions to the steady-constant constraint programming model for distance boundary provided in the embodiments of this disclosure;

[0044] Figures 7(a)-(b) are schematic diagrams of speed planning results under the step speed limit condition provided in the embodiments of this disclosure;

[0045] Figures 8(a)-(b) are schematic diagrams of the velocity planning structure under the upper limit of sinusoidal velocity provided in the embodiments of this disclosure;

[0046] Figures 9(a)-(b) are schematic diagrams of speed planning results under the random speed limit condition provided in the embodiments of this disclosure. Detailed Implementation

[0047] Currently, speed planning is mainly conducted using two methods. One method is based on predefined parametric speed curves. The parameters are modified according to online data (path length, speed limits, etc.) to obtain the actual speed curve used for control. In this method, the predefined parametric speed curve can generally only be optimized, yielding feasible solutions but not optimal solutions. This results in poor planning performance, making it unsuitable for demanding application scenarios. Furthermore, limited by the flexibility of the predefined speed curve, this planning technique is only applicable to a limited range of scenarios and may yield unusable results in some unconsidered situations. The other method is optimization-based speed planning. This requires establishing a longitudinal motion model of the vehicle, defining optimization objectives, and setting constraints. The speed planning problem is transformed into a mathematical model for solving an optimization problem. This method requires iterative calculations for each node and may involve backtracking. The computational complexity is an exponential function of the path length, resulting in high time and computational costs, making it difficult to apply on platforms with limited computing power.

[0048] To this end, this disclosure provides a speed planning method, speed planning device, vehicle, and readable storage medium to improve the adaptability of speed planning and save computing power and time.

[0049] Embodiments of this disclosure are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this disclosure, and should not be construed as limiting this disclosure.

[0050] Referring to Figure 1, which shows a flowchart of a speed planning method provided in an embodiment of this disclosure, the speed planning method provided in an embodiment of this disclosure may include:

[0051] S11: Construct a speed planning model from the starting point to the destination based on the vehicle's state transition equation.

[0052] It should be noted that the speed planning method provided in this disclosure specifically plans the longitudinal motion process of a vehicle traveling along a one-dimensional path. In this disclosure embodiment, X = [t, s, v, a] TDefined as a state vector, it represents the distance s traveled, the speed v, and the acceleration a of the vehicle along a one-dimensional path at time t. Let A = [Δt, j]. T Defined as a behavior vector, it represents the vehicle's jerk j and the duration Δt of this jerk.

[0053] According to the above definition, starting from a certain state X and taking action A, we can obtain the subsequent state X'. This process can be expressed in equation form, that is, the state transition equation of the vehicle can be expressed as shown in equation (1): X′=f(X,A) (1)

[0054] The travel distance s, vehicle speed v, and acceleration a can be calculated using the integral formula shown in equation (2), where τ is the integral variable with respect to time:

[0055] Assuming that the jerk j remains unchanged during the time interval Δt, then equation (2) can be written as equation (3) as follows:

[0056] According to equation (3), the state transition equation in equation (1) can be written as equation (4) as follows:

[0057] From equation (4), it can be seen that starting from a certain initial state X1, if there exists a series of behavior vectors A i (i = 1, 2, ..., N-1), then a series of subsequent states, namely X, can be recursively derived. i =f(X) i-1 A i-1 (i = 1, 2, ..., N), which completely represents the motion state of a vehicle on a one-dimensional path. Here, N is the number of vehicle states, and N-1 is the number of behavior vectors.

[0058] Based on the vehicle state transition equation shown in equation (4), the speed planning problem from the starting point to the destination can be abstracted and constructed to build a speed planning model from the starting point to the destination. This speed planning model includes a behavior vector A. i (i = 1, 2, ..., N-1, N = 2, 3, ...) and constraints, etc.

[0059] The constraints in the constructed velocity planning model need to consider the following aspects:

[0060] (1) Longitudinal motion model

[0061] The longitudinal motion of the vehicle needs to meet certain rules, and the states of any two cycles need to satisfy the relationship in equation (4).

[0062] (2) Boundary conditions

[0063] The starting boundary and the ending boundary, where the starting boundary specifically includes the travel distance s at the starting point. S Speed ​​v S acceleration a S The endpoint boundary specifically includes the travel distance s at the endpoint. E Speed ​​v E acceleration a E .

[0064] (3) Speed ​​limit

[0065] The lower limit of vehicle speed is 0, and the upper limit of vehicle speed is v. max (s), which means the upper limit of speed is a function of the travel distance s.

[0066] (4) Acceleration Limitation

[0067] Specifically, it includes an upper limit and a lower limit for acceleration.

[0068] (5) Time monotonicity constraint

[0069] In the velocity planning process, the time variable t must be monotonically increasing, i.e., Δt>0.

[0070] In other words, the constraints in the constructed speed planning model include the vehicle's state transition equation, the starting boundary, the ending boundary, and the speed limit (the upper speed limit is v). max (s), which means the speed limit is a function of the travel distance s), acceleration limit, and time monotonic constraint.

[0071] Including jerk as a parameter in the constructed velocity planning model can achieve smoothing of the planned velocity, that is, to achieve smooth velocity planning.

[0072] S12: Based on the velocity planning model, plan the first velocity curve from the current state to the destination state, and determine whether the first velocity curve meets the upper speed limit constraint. If not, proceed to step S13; if yes, proceed to step S14.

[0073] Based on step S11, a first speed curve from the current state to the final state can be planned based on the constructed speed planning model. Specifically, the first speed curve can be planned according to a fixed speed upper limit. The planned first speed curve is a curve of speed versus distance traveled, i.e., the horizontal axis represents the distance traveled, and the vertical axis represents the speed.

[0074] After obtaining the first speed curve through planning, it can be determined whether the planned first speed curve satisfies the upper speed limit constraint. Specifically, determining whether the first speed curve satisfies the upper speed limit constraint can be done by checking whether the first speed curve intersects with the upper speed limit constraint. If they intersect, the first speed curve does not satisfy the upper speed limit constraint; if they do not intersect, the first speed curve satisfies the upper speed limit constraint.

[0075] It should be noted that before speed planning, or after constructing a speed planning model from the starting point to the ending point based on the vehicle's state transition equations, and before planning the first speed curve from the current state to the ending state based on the speed planning model for the first time, the starting point boundary and the ending point boundary can be obtained and initialized to facilitate subsequent solution to obtain the speed planning results. Specifically, the initialization mentioned here can involve setting the index p of the expansion layer to 1, the candidate state index q to 1 (q represents the index of the current expansion layer's candidate state), and setting X... 11 =[0,s S ,v S ,a S ] T s S v S a s Given the distance traveled, speed, and acceleration determined at the starting point, X 11 The state when p=1 (i.e., the state in the first extended layer) (i.e., the starting state).

[0076] S13: Expand the current state, select the target candidate state from the expanded candidate states as the current state, and return to step S12.

[0077] If it is determined that the planned first velocity curve does not meet the upper speed limit constraint, the current state (which is contained in the p-th extension layer; when p=1, the current state is the starting state) can be expanded to obtain candidate states (candidate states can be considered to exist in the (p+1)-th extension layer, used to determine whether to use them as the current state based on subsequent processing). The number of candidate states obtained by expanding the current state is the same as the number of times the current state in the p-th extension layer is expanded; that is, each expansion of the current state in the p-th extension layer yields a candidate state in the (p+1)-th extension layer.

[0078] Then, a target candidate state can be selected from multiple candidate states in the (p+1)th extended layer as the current state in the (p+1)th extended layer, and the process returns to the step of planning the first speed curve from the current state to the endpoint state based on the speed planning model, that is, returning to step S12. That is, using the target candidate state as the current state indicates that the target candidate state is a planned state. This target candidate state can be placed at the end of the state sequence, or the target candidate state (specifically, the travel distance and speed in the target candidate state) can be incorporated into the end of the planned speed curve. In other words, the travel distance and speed in the starting state are used as the starting point of the planned speed curve. Each time a target candidate state is determined, the travel distance and speed in that target candidate state can be incorporated into the planned speed curve. During speed planning, a state buffer can be pre-allocated to store the selected target candidate states (i.e., the target candidate states can be stored as a state sequence or as a planned speed curve), so that the speed planning result can be obtained subsequently based on the stored target candidate states.

[0079] The above process ensures that only one candidate state is used as the current state in subsequent velocity planning at each expansion layer (i.e., only one candidate state is incorporated into the planned velocity curve at each expansion layer), and that each expansion layer calculates at most m (m equal to the number of expansions of the current state) candidate states, making the computational complexity a linear function of the path length. Furthermore, the process allows for recursion from the starting point without backtracking. Therefore, compared to existing optimization-based velocity planning methods that use a tree structure and calculate every node in each layer of the tree structure (where the number of nodes in each layer is exponentially related to the layer number; for example, if the first layer contains x nodes, and each node derives k nodes in the next layer, then the second layer contains x*k nodes, the third layer contains x*k nodes, and so on), this method achieves significantly higher computational complexity. 2 There are n nodes, and the nth layer contains x*k nodes. (n-1) The computational complexity of speed planning (which involves multiple nodes) is an exponential function of the path length, and existing optimization-based methods often backtrack to previously calculated layers for recalculation if the results do not meet requirements. This embodiment of the present disclosure reduces the computational complexity of speed planning, thereby reducing computational and time consumption, lowering the demand on platform computing power, improving speed planning efficiency, and obtaining accurate speed planning results. Furthermore, this embodiment does not predefine speed curves but directly performs speed planning; therefore, it can obtain optimal speed planning results for any scenario, improving speed planning performance and adaptability to different scenarios.

[0080] S14: Obtain the speed planning result based on the first speed curve.

[0081] If the planned first velocity curve from the current state to the final state satisfies the upper speed limit constraint, the planning is considered successful, and the velocity planning result can be obtained from the first velocity curve. Specifically, if the first velocity curve from the starting state to the final state is planned using the starting state as the current state and the planned first velocity curve satisfies the upper speed limit constraint (i.e., without going through step S12, the first velocity curve planned using the starting state as the current state satisfies the upper speed limit constraint), then the velocity planning result can be obtained directly from the first velocity curve. Specifically, the first velocity curve can be used as the final velocity planning curve (i.e., the velocity planning result can be in the form of a velocity planning curve), or the states corresponding to the first velocity curve (i.e., state vectors) can be obtained, and the states can be arranged in order to form a state sequence to obtain the velocity planning result (i.e., the velocity planning result can be in the form of a state sequence). If step S12 is required to plan the first speed curve that satisfies the speed upper limit constraint, the speed planning result can be obtained based on the target candidate state and the first speed curve. Specifically, the target candidate state can be concatenated with the first speed curve in sequence to obtain the speed planning result (i.e., the expression is a speed planning curve). In other words, the planned speed curve can be concatenated with the first speed curve to obtain the final speed planning curve. Of course, the various states corresponding to the first speed curve can also be obtained. The obtained target candidate state and the various states corresponding to the first speed curve can be arranged in sequence to obtain the speed planning result (i.e., the expression is a state sequence).

[0082] After obtaining the speed planning result, a success message and the speed planning result (i.e., the speed planning result can be in the form of a state sequence or a speed planning curve) can be output. This allows relevant personnel to confirm the success of the speed planning and obtain the corresponding speed planning result, facilitating intelligent driving of the vehicle based on the planned speed planning result. The process of planning the first speed curve from the current state to the destination state based on the speed planning model, determining the planning success when the planned first speed curve meets the upper speed limit constraint, and directly obtaining the speed planning result eliminates the need for a recursive approach to speed planning. This significantly reduces the computational complexity and power consumption of speed planning, improves its efficiency, and provides accurate speed planning results.

[0083] As can be seen from the above speed planning process, this embodiment first constructs a speed planning model, then plans a first speed curve from the current state to the destination state. If the first speed curve does not meet the speed upper limit constraint, the current state is expanded. A target candidate state is selected from the candidate states included in the next expansion layer as the current state, and the planning of the first speed curve from the current state to the destination state and related processes are continued. If the first speed curve meets the speed upper limit constraint, the speed planning result is obtained based on the first speed curve. This achieves the goal of obtaining a smooth speed curve that meets the smoothness constraint by recursively traversing the path once, starting from the starting point. It does not use a predefined speed curve or an optimization-based method. Compared with the speed planning method based on a predefined speed curve, the speed planning method provided by this embodiment can obtain the optimal speed planning result for any scenario. Compared with the speed planning method based on optimization, the computational complexity of this embodiment is only a linear function of the path length, not an exponential function, greatly reducing computation time and computing power consumption.

[0084] The technical solution disclosed in this embodiment constructs a speed planning model from the starting point to the ending point based on the vehicle's state transition equation. Based on this model, a first speed curve from the current state to the ending state is planned, and it is determined whether the first speed curve satisfies the speed upper limit constraint. If the first speed curve does not satisfy the speed upper limit constraint, the current state is expanded, and a target candidate state is selected as the current state from the expanded candidate states. The process then returns to the step of planning the first speed curve. This process ensures that each expansion layer corresponds to only one target candidate state as the current state, meaning that the number of calculations for each expansion layer is at most equal to the number of expansions for the current state. This makes the computational complexity a linear function of the path length, thereby reducing computational complexity, computation time, and computational power consumption, and facilitating accurate speed planning results. If the first speed curve satisfies the speed upper limit constraint, the planning is considered successful, and the speed planning result is directly obtained based on the target candidate state and the first speed curve, greatly reducing computation time and computational power consumption, and obtaining accurate speed planning results. This process achieves a smooth speed curve that satisfies the smoothness constraint by recursively traversing the path once from the starting point, thus reducing the computational power and time consumption of speed planning. Furthermore, the speed planning process described above does not require a predefined speed curve and can perform speed planning for any scenario, thereby improving the adaptability of speed planning.

[0085] This disclosure provides a speed planning method in which the upper limit of speed is constrained as a function of the travel distance.

[0086] In this embodiment of the disclosure, the speed limit constraint can be a function of the travel distance s, that is, the speed limit constraint can be expressed as v max(s) means that the speed limit constraint is a dynamic speed limit constraint related to the travel distance and does not require approximation or linearization. Compared to current optimization-based methods, the embodiments of this disclosure can accurately handle the dynamic speed limit constraint related to the travel distance to prevent speed planning results from violating the constraint, thereby facilitating the obtaining of accurate speed planning results. Moreover, setting the speed limit constraint as a function of the travel distance also enables the embodiments of this disclosure to achieve speed planning oriented towards the dynamic speed limit, further improving the accuracy of speed planning.

[0087] Correspondingly, when determining whether the first speed curve satisfies the speed limit constraint, it can also be determined whether the speed at each travel distance in the first speed curve exceeds the speed limit at the corresponding travel distance in the speed limit constraint. If the speed at any travel distance in the first speed curve exceeds the speed limit at the corresponding travel distance in the speed limit constraint, then the first speed curve is determined not to satisfy the speed limit constraint. If the speed at each travel distance in the first speed curve does not exceed the speed limit at the corresponding travel distance in the speed limit constraint, then the first speed curve is determined to satisfy the speed limit constraint.

[0088] As can be seen from the above process, the embodiments of this disclosure incorporate dynamic speed upper limit constraints related to driving mileage into speed planning without requiring approximation or linearization. Therefore, dynamic speed upper limit constraints can be accurately processed to prevent speed planning results from violating constraints, thereby obtaining accurate speed planning results.

[0089] This disclosure provides a speed planning method, wherein the speed planning model includes a range of jerk values;

[0090] Velocity planning methods may also include:

[0091] Candidate behavior vectors are determined based on the range of jerk values. Each candidate behavior vector includes the jerk and the duration of the jerk.

[0092] Expanding the current state and selecting a target candidate state as the current state from the expanded candidate states can include:

[0093] The current state is extended based on the candidate behavior vector. A second velocity curve is planned based on the obtained candidate state. The target candidate state is determined based on the second velocity curve and the upper limit constraint of velocity. The target candidate state is then determined as the current state.

[0094] In this embodiment of the disclosure, the constructed velocity planning model may include a range of jerk values, specifically including jerk j. i The objective function is to find the maximum value within the range of jerk values. Specifically, the range of jerk values ​​can be {-j}. max ,0,jmax}, j max This represents the upper limit of jerk in velocity planning. In other words, the constructed velocity planning model can specifically be defined as allowing the jerk j to be increased under certain constraints. i When the maximum is obtained, the corresponding behavior vector A is obtained. i In other words, in the constructed velocity planning model, the design variable to be solved is the sequence A of behavior vectors. i (i = 1, 2, ..., N-1, N = 2, 3, ...), the solution rule is to select the larger value from the set jerk range while satisfying the constraints. By including the jerk parameter in the speed planning model, the final speed planning result can be a smooth speed curve, thereby improving vehicle driving performance.

[0095] Based on the above, before extending the current state, candidate behavior vectors A can be determined according to the range of jerk values. Each determined candidate behavior vector includes the jerk and the duration of that jerk. For example, when the jerk value range is {-j max ,0,j max When}, the candidate behavior vector A can specifically include: A1=[Δt,j max ] T A2 = [Δt, 0] T A3=[Δt,-j max ] T And it can be used with X. pq Let i represent the q-th candidate state of the p-th extended layer. p Used to record the state of the selected layer p.

[0096] Accordingly, the current state can be expanded based on the determined candidate behavior vectors. Specifically, equation (4) can be used to expand the current state to obtain the candidate states (i.e., candidate state vectors) contained in the next expansion layer. That is, if it is determined that the first speed curve does not meet the speed upper limit constraint, the current state (which is contained in the p-th expansion layer at this time) can be expanded based on the candidate behavior vectors to obtain candidate states (the candidate states can be regarded as existing in the p+1-th expansion layer, used to determine whether to use them as the current state based on subsequent processing). At this time, the number of candidate states contained in the p+1-th expansion layer is at most equal to the number of candidate behavior vectors. Compared with the exponential growth of the number of nodes to be calculated in each layer in the existing optimization-based methods, the embodiments of this disclosure can reduce the number of candidate states in each expansion layer, thereby reducing the time consumption and computing power consumption of speed planning.

[0097] After obtaining the candidate states contained in the (p+1)th extended layer, a second speed curve can be planned based on the candidate states in the (p+1)th extended layer and the constructed speed planning model (for example, the parking speed curve with the shortest time from the candidate state to the parking state can be planned). This second speed curve, along with the speed upper limit constraint, determines the target behavior vector to be used when transitioning from the current state in the pth layer to the next state (i.e., as the new current state). Specifically, after planning the second speed curve, it can be determined whether the second speed curve satisfies the speed upper limit constraint. If the second speed curve does not satisfy the speed upper limit constraint, it indicates that the vehicle's speed will exceed the speed upper limit when transitioning to the state using the corresponding candidate behavior vector in the future, and therefore the corresponding candidate behavior vector can be discarded. If the second speed curve satisfies the speed upper limit constraint, it indicates that the vehicle's speed will not exceed the speed upper limit when transitioning to the state using the corresponding candidate behavior vector in the future, and therefore it can be retained. A target behavior vector can be determined from the retained candidate behavior vectors, and the candidate state corresponding to this target behavior vector is determined as the target candidate state. Furthermore, after determining the target behavior vector, the target behavior vector A can be recorded and stored. pq (q=i p i p Used to record the selected state of the p-th layer and the target candidate state X corresponding to the target behavior vector. pq Among them, the target behavior vector A is recorded and stored. pq and its corresponding target candidate state X pq At that time, records can be made in a defined order to form a sequence of behaviors and a sequence of states, that is, X can be recorded starting from p=1. pq and A pq (q=i p ).

[0098] The above methods can reduce the time and computing power required for speed planning and obtain accurate speed planning results, thereby improving vehicle driving performance.

[0099] This disclosure provides a speed planning method that determines a target candidate state based on a second speed curve and a speed upper limit constraint, and may include:

[0100] The candidate state corresponding to the candidate behavior vector that satisfies the upper speed limit constraint and has the largest jerk in the second velocity curve is determined as the target candidate state.

[0101] In this embodiment of the disclosure, when determining the target candidate state based on the second speed curve and the speed upper limit constraint, specifically, the candidate behavior vector with the largest acceleration can be selected from the candidate behavior vectors of the second speed curve that satisfy the speed upper limit constraint, and the candidate state obtained by expanding the current state using the candidate behavior vector is determined as the target candidate state. That is, the candidate state corresponding to the candidate behavior vector of the second speed curve that satisfies the speed upper limit constraint and has the largest acceleration is determined as the target candidate state, so as to obtain accurate and optimal speed planning results, shorten driving time, and improve driving performance.

[0102] Based on the above, after confirming successful planning, the recorded target behavior vector and the behavior vector corresponding to the first velocity curve can be arranged in order to obtain the behavior sequence corresponding to the velocity planning result. Furthermore, the behavior sequence can be output; that is, after successful planning, the recorded X values ​​can be output starting from p=1. pq and A pq (q=i p This means outputting the state sequence and behavior sequence obtained from the planning, so that relevant personnel can obtain more speed planning information.

[0103] Referring to Figure 2, which shows a flowchart of another speed planning method provided by an embodiment of this disclosure, the speed planning method provided by an embodiment of this disclosure may further include, before determining the target candidate state as the current state:

[0104] Determine whether the planned driving distance of the vehicle exceeds the destination driving distance based on the target candidate status;

[0105] If so, the speed planning result is obtained based on the target candidate state;

[0106] If not, then proceed with the step of determining the target candidate state as the current state.

[0107] In this embodiment of the disclosure, before determining the target candidate state as the current state, and after determining the target candidate state based on the second speed curve and the speed upper limit constraint, it can be determined whether the planned driving distance of the vehicle exceeds the destination driving distance based on the target candidate state. Specifically, the driving distance in the target candidate state can be determined as the planned driving distance of the vehicle, and then it is determined whether the planned driving distance of the vehicle exceeds the destination distance.

[0108] If the planned travel distance of the vehicle exceeds the destination distance, it indicates that although the destination state has not been reached, the planned travel distance has exceeded the destination distance. In this case, planning can also be considered successful, and the speed planning result can be obtained based on the target candidate states. Specifically, a speed planning curve can be plotted based on the target candidate states and their corresponding target behavior vectors to obtain the speed planning result. Alternatively, the target candidate states can be arranged in the order determined by the target candidate states to form a state sequence to obtain the speed planning result. Afterwards, the success of planning and the speed planning result can be output (represented as a candidate state sequence or a speed planning curve).

[0109] If the planned driving distance of the vehicle does not exceed the destination distance, it is determined that the planning has not been successful. At this time, subsequent related steps can be executed, such as returning to execute steps to determine the target candidate state as the current state, so as to continue speed planning.

[0110] The above methods can accurately determine whether speed planning is successful, thereby reducing additional computing power and time consumption and improving the accuracy of speed planning.

[0111] Referring to Figures 3 and 4(a)-(f), where Figure 3 shows a flowchart of another speed planning method provided by an embodiment of the present disclosure, and Figures 4(a)-(f) show schematic diagrams of the periodic decomposition of speed planning provided by an embodiment of the present disclosure. A speed planning method provided by an embodiment of the present disclosure may further include:

[0112] Arrange the candidate behavior vectors in descending order of jerk;

[0113] Extending the current state based on candidate behavior vectors, planning a second velocity curve based on the obtained candidate states, and determining the target candidate state based on the second velocity curve and velocity upper limit constraints can include:

[0114] The candidate behavior vector ranked first is taken as the current candidate behavior vector. The current state is then expanded based on the current candidate behavior vector to obtain the candidate state.

[0115] Plan the second velocity curve based on the candidate states and velocity planning model, and determine whether the second velocity curve meets the upper speed limit constraint.

[0116] If so, then the candidate state corresponding to the current candidate behavior vector is determined as the target candidate state;

[0117] If not, the next candidate behavior vector in the order is taken as the current candidate behavior vector, and the process returns to perform the step of expanding the current state based on the current candidate behavior vector.

[0118] In this embodiment of the disclosure, when setting candidate behavior vector A according to the range of jerk values, the candidate behavior vectors can be arranged in descending order of jerk value according to the magnitude of the jerk contained in each candidate behavior vector, forming a candidate behavior sequence. For example, when the jerk value range is {-j... max ,0,j max The candidate behavior vector A specifically includes: A1 = [Δt,j max ] T A2 = [Δt, 0] T A3=[Δt,-j max ] T In this case, the order of these three candidate behavior vectors can be A1, A2, and A3.

[0119] Based on the above, for each extension layer, the specific process of extending the current state according to the candidate behavior vector, planning the second velocity curve according to the candidate state, and determining the target candidate state according to the second velocity curve and the velocity upper limit constraint can be as follows:

[0120] Step 1: Take the first candidate behavior vector in the candidate behavior sequence as the current candidate behavior vector.

[0121] Step 2: Based on the current candidate behavior vector, expand the current state using equation (4) to obtain the candidate state.

[0122] Step 3: Based on the candidate state planning and velocity planning model obtained in Step 2, plan the second velocity curve and determine whether the planned second velocity curve meets the upper speed limit constraint. If the planned second velocity curve meets the upper speed limit constraint, proceed to Step 4. If the planned second velocity curve does not meet the upper speed limit constraint, proceed to Step 5.

[0123] In the second speed curve, the horizontal axis represents the distance traveled, and the vertical axis represents the speed. Furthermore, when planning the second speed curve, the speed v at the destination can be considered. E acceleration a E All values ​​are set to 0, meaning the second speed curve can specifically be a parking speed curve (e.g., the parking speed curve with the shortest time). In other words, deceleration and parking can be performed based on a candidate state obtained in the next extended layer. It is then determined whether the lower speed limit meets the upper speed limit constraint during the deceleration and parking process. If it does not meet the constraint, the speed of the vehicle under other driving conditions is even less likely to meet the upper speed limit constraint. At this point, step 5 can be executed. If it does meet the constraint, the speed of the vehicle under other driving conditions may meet the upper speed limit constraint. Based on this, the second speed curve can be further extended and planned, i.e., step 4 can be executed.

[0124] Step 4: If it is determined that the planned second velocity curve satisfies the upper speed limit constraint, then the candidate state obtained in step 2 can be determined as the target candidate state.

[0125] Step 5: If it is determined that the planned second speed curve does not meet the speed upper limit constraint, the next candidate behavior vector in the candidate behavior sequence can be used as the current candidate behavior vector, and then return to execute step 2.

[0126] For example, the above process is illustrated in Figures 4(a)-(f). In Figures 4(a)-(f), the horizontal axis represents the travel distance, the vertical axis represents the speed, the dotted line represents the planned speed curve, the circled line represents the speed upper limit constraint, the dashed line represents the first speed curve from the current state to the destination state, the solid line represents the speed curve corresponding to the candidate state obtained after expanding the current state using the candidate behavior vector, and the dotted-dash line represents the second speed curve expanded based on the expanded candidate state (specifically, the speed curve that decelerates from the speed curve corresponding to the candidate state to 0). As shown in Figures 4(a)-(f), each time, starting from the current state, a first speed curve from the current state to the destination state is planned, and it is determined whether the planned first speed curve meets the speed upper limit constraint. If it does not meet the constraint, the current state is expanded according to the candidate behavior vector, and a second speed curve (i.e., the parking speed curve) is planned starting from the expanded state. The decision on whether to accept the expanded point is based on whether the planned second speed curve meets the speed upper limit constraint. If an expanded state is accepted, the expanded state is used as the current state and the above process is repeated; otherwise, the process switches to the next expanded point.

[0127] As shown in Figure 4(a), a first speed curve is first planned from the end of the planned speed curve (i.e., the current state) to the end state. It is determined that the planned first speed curve intersects with the speed upper limit constraint, and it is determined that the planned first speed curve does not meet the speed upper limit constraint. At this time, the end of the planned speed curve is extended using the candidate behavior vector A1 to obtain the speed curve corresponding to the extended candidate state. A second speed curve is planned based on the speed curve corresponding to the extended candidate state. It is determined that the planned second speed curve does not intersect with the speed upper limit constraint, and it is determined that the planned second speed curve meets the speed upper limit constraint. At this time, the speed curve corresponding to the candidate behavior vector A1 can be merged into the planned speed curve to obtain the planned speed curve in Figure 4(b).

[0128] As shown in Figure 4(b), a first velocity curve reaching the endpoint state is planned at the end of the planned velocity curve. It is determined that the planned first velocity curve intersects with the upper speed limit constraint, thus the planned first velocity curve does not meet the upper speed limit constraint. At this point, the end of the planned velocity curve is extended using candidate behavior vector A1 to obtain the velocity curve corresponding to the extended candidate state. A second velocity curve is planned based on the extended candidate state velocity curve. It is determined that the planned second velocity curve intersects with the upper speed limit constraint, thus the second velocity curve does not meet the upper speed limit constraint. Next, the end of the planned velocity curve is extended using candidate behavior vector A2 to obtain the velocity curve corresponding to the extended candidate state. A second velocity curve is planned based on the extended candidate state velocity curve. It is determined that the planned second velocity curve intersects with the upper speed limit constraint, thus the second velocity curve does not meet the upper speed limit constraint. Next, the candidate behavior vector A3 is used to extend the end of the planned velocity curve to obtain the velocity curve corresponding to the extended candidate state. A second velocity curve is planned based on the velocity curve corresponding to the extended candidate state. It is determined that the planned second velocity curve does not intersect with the upper limit constraint, and it is considered that the second velocity curve satisfies the upper limit constraint. At this time, the velocity curve corresponding to the candidate behavior vector A3 can be merged into the planned velocity curve.

[0129] It should be noted that Figures 4(a)-4(f) are performed in sequence, but the order may not be consecutive. For example, Figure 4(c) follows Figure 4(b), and Figure 4(d) follows Figure 4(c), but Figure 4(c) does not immediately follow Figure 4(b), and Figure 4(d) does not immediately follow Figure 4(c). Some similar steps are omitted in between... Finally, as shown in Figure 4(f), a first speed curve that reaches the endpoint is planned at the end of the planned speed curve. It is determined that the planned first speed curve meets the upper speed limit constraint. At this point, the planning can be considered successful, and the first speed curve is merged into the planned speed curve to obtain the final speed planning curve.

[0130] By selecting candidate behavior vectors in descending order of accelerometer speed for expansion, the most effective candidate behavior vector can be selected for each expansion layer. This allows the current state of the expansion layer to require less than m expansion calculations (e.g., only 1 or 2 expansion calculations are needed, and the current state of the expansion layer requires at most m expansion calculations), which greatly reduces computational complexity and power consumption, and improves speed planning efficiency.

[0131] Of course, when expanding the current state in the p-th expansion layer based on the candidate behavior vector, we can also use m candidate behavior vectors to expand the current state in the p-th expansion layer simultaneously to obtain m candidate states in the p+1-th layer. Based on the m candidate states in the p+1-th layer, we plan m second velocity curves. We select the second velocity curve that satisfies the upper limit of velocity constraint and has the maximum jerk among the candidate behavior vectors as the target second velocity curve. We determine the candidate behavior vector corresponding to the target second velocity curve as the target behavior vector. We use the target candidate state corresponding to the target behavior vector as the current state in the p+1-th expansion layer to continue velocity planning. In this way, each expansion layer requires m calculations, and the computational complexity is still a linear function of the path length. Therefore, we can still reduce computational complexity and computing power consumption and improve the efficiency of velocity planning.

[0132] The velocity planning method provided in this disclosure, after expanding the current state based on the current candidate behavior vector, may further include:

[0133] Determine whether the candidate state satisfies the speed upper limit constraint;

[0134] If so, proceed with the step of planning the second velocity curve based on the candidate state and the velocity planning model;

[0135] If not, then proceed with the step of taking the next candidate behavior vector as the current candidate behavior vector.

[0136] In this embodiment of the disclosure, after expanding the current state according to the current candidate behavior vector to obtain candidate states, it can be determined whether the obtained candidate states satisfy the speed upper limit constraint, that is, whether the speed in the candidate states does not exceed the corresponding speed upper limit in the speed upper limit constraint. If none of the speeds in the candidate states exceed the corresponding speed upper limit, it is determined that the candidate states in the next expansion layer satisfy the speed upper limit constraint. If there is a speed in the candidate states that exceeds the corresponding speed upper limit, it is determined that the candidate states in the next expansion layer do not satisfy the speed upper limit constraint.

[0137] If the obtained candidate states satisfy the upper limit of velocity constraints, then the step of planning the second velocity curve based on the candidate states and the velocity planning model can be performed. If the obtained candidate states do not satisfy the upper limit of velocity constraints, then the step of using the next candidate behavior vector as the current candidate behavior vector can be performed.

[0138] The above judgment ensures that the target candidate state corresponding to the target behavior vector (that is, the current state in each extended layer) will not exceed the speed upper limit constraint, thereby improving the accuracy of speed planning.

[0139] This disclosure provides a speed planning method that, when the current candidate behavior vector is the last candidate behavior vector in the sequence, further includes the following if it is determined that the second speed curve does not meet the speed upper limit constraint:

[0140] The plan has failed, and a message indicating the failure is displayed.

[0141] In this embodiment of the disclosure, during the process of expanding the current state according to the candidate behavior sequence, when the last candidate behavior vector is taken as the current candidate behavior vector, the current state is expanded according to the current candidate behavior vector, and a second speed curve is planned based on the expanded candidate state. When it is determined whether the planned second speed curve meets the speed upper limit constraint, if it is determined that the planned second speed curve does not meet the speed upper limit constraint, it indicates that the subsequent planned speed will exceed the speed upper limit constraint. Therefore, it can be determined that the constraint condition has no solution. At this time, it is determined that the planning has failed, and a planning failure prompt (which may include content such as the constraint condition having no solution) can be output so that relevant personnel can perform corresponding processing.

[0142] The speed planning method provided in this disclosure, after constructing a speed planning model from the starting point to the ending point based on the vehicle's state transition equation, may further include:

[0143] Based on the velocity programming model, a steady-constrained programming model for distance boundary is constructed. The steady-constrained programming model for distance boundary includes the objective function of the shortest time. The constraints of the steady-constrained programming model for distance boundary include the starting boundary, the ending boundary, the velocity limit, the acceleration limit, and the range of jerk values. The upper limit of the velocity in the velocity limit is a fixed value.

[0144] Planning the first velocity curve from the current state to the final state based on a velocity planning model can include:

[0145] Based on the steady-constraint programming model of distance boundary, plan the first velocity curve that takes the shortest time to reach the final state from the current state.

[0146] In this embodiment of the disclosure, after constructing a speed planning model from the starting point to the ending point based on the vehicle's state transition equation, a steady-constant constraint planning model for the distance boundary can be constructed based on the constructed speed planning model. This steady-constant constraint planning model for the distance boundary contains the objective function of minimizing time, and the constraints of the steady-constant constraint planning model for the distance boundary include the state transition equation and the starting point boundary (specifically including the travel distance s at the starting point). S Speed ​​v S acceleration a S ), finish line boundary (specifically including the driving distance s at the finish line) E Speed ​​v E acceleration aE ), speed limit (the upper limit of speed in the speed limit is a fixed value), acceleration limit (specifically including upper limit of acceleration and lower limit of acceleration), time monotonicity constraint and jerk value range.

[0147] Based on the established steady-constraint programming model for distance to the boundary, a first velocity curve with the shortest time to reach the final state can be planned according to the established steady-constraint programming model for distance to the boundary, with a fixed upper speed limit. Specifically, the first velocity curve considers planning the shortest path from the current state to the final state with a fixed upper speed limit. The objective function for minimizing time in the steady-constraint programming model for distance to the boundary is specifically the objective function for minimizing the time from the current state to the final state.

[0148] By using the above method to plan the first speed curve that takes the shortest time to reach the destination state from the current state, speed planning performance can be improved and the time for intelligent driving can be shortened.

[0149] The speed planning method provided in this disclosure, after constructing a speed planning model from the starting point to the ending point based on the vehicle's state transition equation, may further include:

[0150] Construct a steady-constraint programming model for the velocity boundary based on the velocity programming model; the steady-constraint programming model for the velocity boundary includes the objective function of the shortest time, and the constraints of the steady-constraint programming model for the velocity boundary may include the starting velocity, the starting acceleration, the ending velocity, the ending acceleration, the acceleration limit, and the range of jerk values;

[0151] The second velocity curve can be planned based on the candidate states, which may include:

[0152] The second velocity curve with the shortest time from the candidate state to the stopping state is planned using a steady-constraint programming model based on the velocity boundary.

[0153] In this embodiment, after constructing a speed planning model from the starting point to the ending point based on the vehicle's state transition equation, a steady-constraint planning model for the speed boundary can be constructed based on the constructed speed planning model to plan the second speed curve. This steady-constraint planning model for the speed boundary contains the objective function of minimizing time, and its constraints include the state transition equation, starting speed, starting acceleration, ending speed, ending acceleration, acceleration limits, time monotonicity constraints, and the range of jerk values. Furthermore, the starting speed, starting acceleration, ending speed, and ending acceleration are all fixed values ​​(i.e., all given). In other words, the second speed curve is a speed planning problem that, given the starting speed, starting acceleration, ending speed, and ending acceleration, needs to satisfy the upper and lower limits of acceleration and jerk, and requires the shortest possible travel time.

[0154] Based on the above, when planning the second speed curve, the steady constraint planning model of the speed boundary can be used to plan the second speed curve with the shortest time from the candidate state to the parking state. In order to use the lower speed limit to predict whether the upper speed limit constraint will definitely not be met in the future, the candidate behavior vectors that may exceed the upper speed limit constraint in the future can be screened out, and the backtracking of the calculation process can be avoided. It is possible to obtain the result by recursively traversing the path once starting from the starting point, thereby reducing the computational complexity, computation time and computing power consumption.

[0155] The speed planning method provided in this disclosure, before planning the second speed curve with the shortest time from the candidate state to the stopping state using a steady-constant constraint planning model of the speed boundary, may further include:

[0156] An acceleration boundary velocity programming model is constructed and solved to obtain the first solution result. The acceleration boundary velocity programming model includes the objective function of the shortest time. The constraints of the acceleration boundary velocity programming model include the starting acceleration, the ending acceleration, and the range of jerk values. The starting acceleration, the ending acceleration, and the upper and lower limits of the jerk are all constants.

[0157] Planning the second velocity curve with the shortest time from the candidate state to the stopping state using a steady-constraint programming model based on velocity boundaries can include:

[0158] Based on the first solution result, the steady constraint programming model of the velocity boundary is solved to obtain the second solution result;

[0159] Based on the candidate states and the second solution results, the second velocity curve with the shortest time from the candidate state to the stopping state is obtained.

[0160] In this embodiment, to facilitate solving the steady-constrained programming model for the velocity boundary and the steady-constrained programming model for the distance boundary, an acceleration boundary velocity programming model can be constructed based on the constructed velocity programming model. This acceleration boundary velocity programming model contains the objective function of minimizing time, and its constraints include a state transition equation, initial acceleration, final acceleration, time monotonicity constraints, and a range of jerk values. The initial acceleration, final acceleration, and upper and lower limits of the jerk are all constants. Since the acceleration boundary velocity programming model includes jerk, it is essentially a steady-constrained S-shaped velocity programming problem with acceleration boundaries, specifically: 1) given initial and final accelerations, 2) jerk upper and lower limits are constants, and 3) a velocity programming problem with the objective of minimizing the total time.

[0161] After constructing the acceleration boundary velocity planning model, the model can be solved to obtain the first solution result. This first solution result may include the value of N and the corresponding target behavior vector.

[0162] Based on the above, when planning the second velocity curve with the shortest time from the candidate state to the stopping state using the steady-constraint programming model of the velocity boundary, we can first solve the steady-constraint programming model of the velocity boundary based on the first solution obtained from solving the acceleration boundary velocity programming model, to obtain the second solution. The second solution can include the value of N and the corresponding target behavior vector. Then, based on the candidate state and the second solution, we can obtain the second velocity curve with the shortest time from the candidate state to the stopping state. Specifically, based on the candidate state and the second solution, we can use the state transition equation to obtain the second velocity curve with the shortest time from the candidate state to the stopping state.

[0163] This disclosure provides a velocity planning method that, based on a steady-constant constraint planning model for distance boundaries, plans a first velocity curve that minimizes the time required to reach the final state from the current state. This method may include:

[0164] Based on the second solution result, the steady-constrained programming model of distance boundary is solved to obtain the third solution result;

[0165] Based on the current state and the third solution result, the first velocity curve with the shortest time to reach the final state from the current state is obtained.

[0166] To plan the first velocity curve that takes the shortest time to reach the final state from the current state using a steady-constraint programming model based on distance boundaries, we can first solve the acceleration boundary velocity programming model to obtain a first solution. Then, based on the first solution, we solve the steady-constraint programming model for the velocity boundary to obtain a second solution. Next, based on the second solution, we solve the steady-constraint programming model for the distance boundary to obtain a third solution, which includes the value of N and the corresponding target behavior vector. After obtaining the third solution, we can use the state transition equation to obtain the first velocity curve that takes the shortest time to reach the final state from the current state, based on the current state and the third solution.

[0167] The above methods enable the optimal solution for static speed upper limit constraints to be obtained in a single calculation, without the need for iterative calculations or the calculation of constraints for multiple sets of curves. This reduces computational power consumption, yields accurate results, and improves the efficiency and accuracy of speed planning.

[0168] This disclosure provides a speed planning method, the speed planning model of which is:

[0169] Among them, A i Let Δt be the vector of the i-th row. i Let j be the duration of the i-th action vector. i Let be the vehicle's jerk in the i-th action vector. Let N be the set of real numbers, N-1 be the number of behavior vectors, N be the number of vehicle states, and j be the number of vehicle states. max X is the upper limit of jerk in velocity planning. i =[t i ,s i ,v i ,a i ] T Let s be the i-th state vector of the vehicle. i For vehicles at t i Distance traveled at any given time, v i For vehicles at t i The speed of time, a i For vehicles at t i acceleration at time X i+1 =f(X) i A i Let be the state transition equation for the vehicle, where s1, v1, and a1 are variables representing the distance traveled, speed, and acceleration at the starting point, respectively. S v S a S Given the distance traveled, speed, and acceleration determined at the starting point, s N vN a N Let be the variables representing the distance traveled, speed, and acceleration at the destination, s E v E a E Let v be the distance traveled, speed, and acceleration determined at the destination, where 0 ≤ v i ≤v max (s i () represents the speed limit, v max (s i ) represents the distance traveled in s i The maximum speed at time a min ≤a i ≤a max For acceleration limitation, a min For minimum acceleration, a max This is the maximum acceleration.

[0170] In this embodiment of the disclosure, the speed planning model from the starting point to the ending point constructed based on the vehicle's state transition equation (4) can be specifically shown in equation (5):

[0171] The design variable to be solved is the sequence A of behavior vectors. i The solution rule is to solve the problem within -j while satisfying the constraints. max , 0, j max Choose the larger of the three values, j max It is the upper limit of jerk in speed planning.

[0172] The constraints of the velocity programming model consider five aspects: longitudinal motion model, boundary conditions, velocity limits, acceleration limits, and time monotonicity constraints. Specifically, in the velocity limits, the lower velocity limit is 0, and the upper velocity limit is v. max (s), since the longitudinal motion model in equation (4) is a recursive equation with time as the independent variable, and v max (s) is a function of the travel distance s, causing this constraint to exhibit significant nonlinearity. Compared to existing methods that cannot directly handle this constraint, or use approximation methods to convert the dynamic constraint into a fixed constraint, or use local linearization methods to process it into a linear function, or require updating the constraint in each iteration cycle, this embodiment of the present disclosure, through the aforementioned planning of a first speed curve and a second speed curve, connects the first speed curve and the second speed curve with v. max(s) compares and determines whether the first speed curve and the second speed curve meet the dynamic speed upper limit constraint. Based on the determination result, corresponding processing measures are adopted to accurately process the dynamic speed upper limit constraint related to the travel distance. No approximation or linearization processing is required, which can prevent the speed planning result from violating the constraint, obtain accurate results, and make the computational complexity a linear function of the path length, thereby reducing computational complexity and saving computing power.

[0173] The velocity planning method provided in this embodiment has a steady-constraint planning model for distance boundaries as follows:

[0174] In equation (6), t N min.t is the time to reach the final state. N Let be the objective function of the constant-constraint programming model for distance from the boundary.

[0175] When considering a fixed speed limit for planning the shortest path from the current state to the final state on the first speed curve, the steady-constrained planning model for the distance boundary can be specifically shown in equation (6) above.

[0176] This disclosure provides a velocity planning method, wherein the steady constraint planning model for the velocity boundary is as follows:

[0177] When the second velocity curve is a speed planning problem that is given a starting velocity, starting acceleration, ending velocity, and ending acceleration, and needs to satisfy the upper and lower limits of acceleration and jerk constraints, and requires the shortest travel time, the steady constraint planning model of the distance boundary can be specifically shown in the above equation (7).

[0178] This disclosure provides a velocity planning method, wherein the acceleration boundary velocity planning model is as follows:

[0179] In this embodiment of the disclosure, in order to facilitate the solution of equations (6) and (7), the acceleration boundary velocity planning model shown in equation (8) can be planned, and equation (8) can be solved.

[0180] Regarding the solution of equation (8), it can be seen from equation (8) that the acceleration change between two adjacent states is jΔt. If only the acceleration at the starting point and the ending point is constrained, then j takes the limit of the acceleration (jΔt). max or -j max This will minimize the required Δt, requiring only one action to reach the final acceleration, i.e., the solution result of equation (8) is shown in equation (9):

[0181] Based on the behavior described in the above formula, the resulting change in velocity is:

[0182] The velocity planning problem for the second velocity curve has been defined in equation (7). Compared with the acceleration boundary problem defined in equation (8), acceleration constraints at the starting and ending points, as well as upper and lower limit constraints on acceleration, have been added. Here, the velocity increment from the starting point to the ending point is divided into 5 cases. There are a total of 5 cases in the solution results of equation (7), as shown in Figures 5(a)-(e). Figures 5(a)-(e) show schematic diagrams of the five solutions of the steady constraint planning model of the velocity boundary provided in the embodiments of this disclosure. The specific results of the five solutions in Figure 5 are given below:

[0183] (one)

[0184] As shown in Figure 5(a), its corresponding The distribution of velocity, acceleration, and jerk over time, where the horizontal axis represents time and the vertical axis corresponds to velocity, acceleration, and jerk.

[0185] When the velocity increment from the starting point to the ending point is greater than the acceleration is increased to a... max When the critical value is reached, an acceleration of a will be generated. max The uniformly accelerated segment, namely:

[0186] (two)

[0187] As shown in Figure 5(b), its corresponding The distribution of velocity, acceleration, and jerk over time, where the horizontal axis represents time and the vertical axis corresponds to velocity, acceleration, and jerk.

[0188] When the velocity increment from the starting point to the ending point is increased by adjusting the acceleration to a max The critical value and directly adjusting the acceleration to a E Between particular solutions, there is no uniform acceleration segment; there is only a segment of acceleration increasing and a segment of acceleration decreasing, i.e.:

[0189] (three)

[0190] As shown in Figure 5(c), its corresponding The distribution of velocity, acceleration, and jerk over time, where the horizontal axis represents time and the vertical axis corresponds to velocity, acceleration, and jerk.

[0191] When the velocity increment from the starting point to the ending point is less than a, the acceleration is reduced to a. min When the critical value is reached, an acceleration of a will be generated. min The uniformly decelerated segment, namely:

[0192] (Four)

[0193] As shown in Figure 5(d), its corresponding The distribution of velocity, acceleration, and jerk over time, where the horizontal axis represents time and the vertical axis corresponds to velocity, acceleration, and jerk.

[0194] When the velocity increment from the starting point to the ending point is reduced to a min The critical value and directly adjusting the acceleration to a E Between particular solutions, there is no uniform acceleration segment; there is only a segment of decreasing acceleration and a segment of increasing acceleration, i.e.:

[0195] (five)

[0196] As shown in Figure 5(e), its corresponding The distribution of velocity, acceleration, and jerk over time, where the horizontal axis represents time and the vertical axis corresponds to velocity, acceleration, and jerk.

[0197] When the velocity increment from the starting point to the ending point is exactly equal to directly changing the acceleration from a... S Adjust to a E When the special solution is obtained, there is only one behavior, namely the case of acceleration boundary velocity planning corresponding to equation (9).

[0198] The above five cases are the results of solving equation (7).

[0199] Next, we solve equation (6). The steady constraint programming model for distance boundary is shown in equation (6) (i.e., the S-shaped velocity programming problem for distance boundary is defined in equation (6)). Here, we use the velocity boundary programming problem (i.e., the model shown in equation (7)) as the basis and use the same approach as above for equation (7) to solve equation (6).

[0200] For the purposes of the following representation, the action sequence obtained by solving equation (7) using the above algorithm will be written as A(v). S ,a S ,v E ,a E ), N is represented as N(v S ,a S ,v E ,a E The resulting distance increment is represented as Δs(v) S ,a S ,v E ,a E ).

[0201] Specifically, the solution results of equation (6) can be divided into three cases, as shown in Figures 6(a)-(c). Figures 6(a)-(c) show schematic diagrams of three solutions of the steady constraint programming model for distance boundary provided in the embodiments of this disclosure. The decomposition conditions and calculation methods of these three solutions are given below.

[0202] (I) Case 1: v exists max constant speed segment

[0203] As shown in Figure 6(a), there exists v max The distribution of velocity, acceleration, and jerk during the uniform motion segment, where the horizontal axis represents time and the vertical axis corresponds to velocity, acceleration, and jerk.

[0204] The critical state in this case is v max The duration of the uniform velocity segment is 0, meaning that this situation requires the following condition to be met:

[0205] s E -s S >Δs(v S ,a S ,v max ,0)+Δs(v max ,0,v E ,a E (15)

[0206] When the condition of equation (15) is satisfied, the behavior sequence is:

[0207] (ii) Case 2: There exists a point where the acceleration v' is 0.

[0208] As shown in Figure 6(b), it shows the distribution of velocity, acceleration, and jerk when there is a point where v' acceleration is 0. The horizontal axis represents time, and the vertical axis corresponds to velocity, acceleration, and jerk.

[0209] When the condition in equation (15) is not met, the speed cannot be increased to v. max The speed can only increase to a certain intermediate value v' before deceleration needs to begin (as shown in Figure 6(b)). With the distance increment s... E -s S As it gradually decreases, v' will also continuously decrease, and its lower limit v0 is:

[0210] Considering the lower limit of v', case two should satisfy equation (18) without satisfying equation (15):

[0211] s E -s S ≥Δs(vS ,a S ,v0,0)+Δs(v0,0,v E ,a E (18)

[0212] The resulting sequence of behaviors is:

[0213] The v' required to solve the behavior sequence according to equation (19) can be determined by the following equation:

[0214] s E -s S =Δs(v S ,a S ,v′,0)+Δs(v′,0,v E ,a E v′∈[v0,v max (20)

[0215] (III) Case 3: There is no point where the acceleration is 0.

[0216] As shown in Figure 6(c), it shows the distribution of velocity, acceleration, and jerk when there is no point where the acceleration is 0. The horizontal axis represents time, and the vertical axis represents velocity, acceleration, and jerk.

[0217] If, in case two, v' reaches its lower limit, the distance increment s E -s S If we further reduce the value, and it does not satisfy equation (18) but still satisfies equation (21), then we will enter case three, that is, there is no intermediate state where the acceleration is 0.

[0218] s E -s S ≥Δs(v S ,a S ,v E ,a E ) (twenty one)

[0219] At this point, the sequence of actions is:

[0220] Where a' is not 0.

[0221] The v' and a' required to calculate equation (22) are determined by solving the equation in the following formula:

[0222] The above three cases are the solutions to equation (6).

[0223] Speed ​​planning using the algorithm mentioned above in this disclosure can yield a smooth speed curve with the shortest time that satisfies dynamic speed constraints. Specifically, the algorithm can be verified using three dynamic speed limits, as shown in Figures 7-9. Figures 7(a)-(b) are schematic diagrams of speed planning results under the stepped speed limit condition provided in the embodiments of this disclosure. Figure 7(a) shows the final speed planning curve and the distribution of the stepped speed limit, with the horizontal axis representing the travel distance and the vertical axis representing speed. Figure 7(b) shows the distribution of planned speed, acceleration, and jerk, with the horizontal axis representing time and the vertical axis corresponding to speed, acceleration, and jerk. Figures 8(a)-(b) are schematic diagrams of the speed planning structure under the sinusoidal speed limit condition provided in the embodiments of this disclosure. Figure 8(a) shows the final speed planning curve and the distribution of the sinusoidal speed limit, with the horizontal axis representing the travel distance and the vertical axis representing speed. Figure 8(b) shows the distribution of planned speed, acceleration, and jerk, with the horizontal axis representing time and the vertical axis corresponding to speed, acceleration, and jerk. Figures 9(a)-(b) are schematic diagrams of the speed planning results under the random speed limit provided in the embodiments of this disclosure. Figure 9(a) shows the final speed planning curve and the distribution of the random speed limit, while Figure 9(b) shows the distribution of the planned speed, acceleration, and jerk. The horizontal axis represents time, and the vertical axis corresponds to speed, acceleration, and jerk. As can be seen from Figures 7-9, the algorithm provided in this disclosure can plan a smooth speed curve under all conditions, and the speed curve is tangent to the speed limit in each case. The algorithm provided in this disclosure successfully plans the fastest speed curve.

[0224] This disclosure also provides a speed planning device, which may include:

[0225] Memory, used to store computer programs;

[0226] A processor is used to execute computer programs and implement the steps of any of the above-mentioned speed planning methods.

[0227] Specifically, the speed planning device can be a vehicle computer, controller, etc.

[0228] This disclosure also provides a vehicle including the speed planning device described above.

[0229] This disclosure also provides a readable storage medium storing a computer program, which, when executed by a processor, can implement the steps of any of the above-described speed planning methods.

[0230] For a description of the relevant parts of the speed planning device, vehicle and readable storage medium provided in the embodiments of this disclosure, please refer to the detailed description of the corresponding parts in the speed planning method provided in the embodiments of this disclosure, and will not be repeated here.

[0231] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be specifically implemented in any readable storage medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "readable storage medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of readable storage media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the readable storage medium could even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0232] It should be understood that various parts of this disclosure can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0233] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0234] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this disclosure, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0235] In this disclosure, unless otherwise expressly specified and limited, the terms "installation," "connection," "linking," "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise expressly limited. Those skilled in the art can understand the specific meaning of the above terms in this disclosure according to the specific circumstances.

[0236] Although embodiments of the present disclosure have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present disclosure. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present disclosure.

Claims

1. A speed planning method characterized by, The method comprises the following steps: constructing a speed planning model from a starting point to an ending point according to a state transition equation of a vehicle; planning a first speed curve from a current state to the ending state based on the speed planning model, and determining whether the first speed curve satisfies a speed upper limit constraint; if not, expanding the current state, selecting a target candidate state from candidate states obtained by the expansion as the current state, and returning to the step of planning the first speed curve from the current state to the ending state based on the speed planning model; if yes, obtaining a speed planning result according to the first speed curve.

2. The speed planning method of claim 1, wherein, The speed upper limit constraint is a function of a driving distance.

3. The speed planning method of claim 2, wherein, The speed planning model comprises a jerk value range. The method further comprises the following steps: determining a candidate behavior vector according to the jerk value range, wherein the candidate behavior vector comprises a jerk and a duration of the jerk; expanding the current state, selecting a target candidate state from candidate states obtained by the expansion as the current state, comprises the following steps: expanding the current state according to the candidate behavior vector, planning a second speed curve according to the candidate state, determining a target candidate state according to the second speed curve and the speed upper limit constraint, and determining the target candidate state as the current state.

4. The speed schedule method of claim 3, wherein, Determining a target candidate state according to the second speed curve and the speed upper limit constraint comprises the following steps: determining the candidate state corresponding to the candidate behavior vector with the largest jerk and satisfying the speed upper limit constraint as the target candidate state.

5. The speed planning method of claim 4, wherein, Before determining the target candidate state as the current state, the method further comprises the following steps: determining whether a planned driving distance of the vehicle exceeds an ending driving distance according to the target candidate state; if yes, obtaining a speed planning result according to the target candidate state; if not, performing the step of determining the target candidate state as the current state.

6. The speed schedule method of claim 4, wherein, The method further comprises the following steps: arranging candidate behavior vectors in descending order of jerk; expanding the current state according to the candidate behavior vector, planning a second speed curve according to the candidate state, determining a target candidate state according to the second speed curve and the speed upper limit constraint, comprises the following steps: arranging the candidate behavior vector at the first position as a current candidate behavior vector, expanding the current state according to the current candidate behavior vector to obtain a candidate state; determining whether the second speed curve satisfies the speed upper limit constraint; if yes, determining the candidate state corresponding to the current candidate behavior vector as the target candidate state; if not, arranging the candidate behavior vector at the next position as the current candidate behavior vector, and returning to the step of expanding the current state according to the current candidate behavior vector.

7. The speed scheduling method of claim 6, wherein, After expanding the current state according to the current candidate behavior vector, the method further comprises the following steps: determining whether the obtained candidate state satisfies the speed upper limit constraint; if yes, performing the step of planning the second speed curve according to the candidate state and the speed planning model; if not, performing the step of arranging the candidate behavior vector at the next position as the current candidate behavior vector. If not, the step of arranging the next candidate behavior vector as the current candidate behavior vector is performed.

8. The speed scheduling method of claim 6, wherein, When the current candidate behavior vector is the last candidate behavior vector, if it is determined that the second speed curve does not satisfy the speed upper limit constraint, further comprising: determining that the planning fails, and outputting a prompt of planning failure.

9. The speed planning method according to any one of claims 3 to 8, characterized in that, After constructing a speed planning model from the start point to the end point according to a state transition equation of the vehicle, further comprising: constructing a distance boundary constant constraint planning model according to the speed planning model; the distance boundary constant constraint planning model comprises a time shortest objective function, and constraint conditions of the distance boundary constant constraint planning model comprise a start point boundary, an end point boundary, a speed limit, an acceleration limit and a jerk value range, wherein the speed upper limit in the speed limit is a fixed value; planning a first speed curve from a current state to a terminal state based on the speed planning model, comprising: planning a first speed curve from the current state to the terminal state based on the distance boundary constant constraint planning model.

10. The speed planning method of claim 9, wherein, After constructing a speed planning model from the start point to the end point according to a state transition equation of the vehicle, further comprising: constructing a distance boundary constant constraint planning model according to the speed planning model; the distance boundary constant constraint planning model comprises a time shortest objective function, and constraint conditions of the distance boundary constant constraint planning model comprise a start point boundary, an end point boundary, a speed limit, an acceleration limit and a jerk value range, wherein the speed upper limit in the speed limit is a fixed value; planning a first speed curve from a current state to a terminal state based on the speed planning model, comprising: planning a first speed curve from the current state to the terminal state based on the distance boundary constant constraint planning model.

11. The speed scheduling method of claim 10, wherein, Before planning a second speed curve from a candidate state to a parking state based on the speed boundary constant constraint planning model, further comprising: constructing an acceleration boundary speed planning model, and solving the acceleration boundary speed planning model to obtain a first solving result; the acceleration boundary speed planning model comprises a time shortest objective function, and constraint conditions of the acceleration boundary speed planning model comprise a start point acceleration, an end point acceleration and a jerk value range, wherein the start point acceleration, the end point acceleration and the upper and lower limits of the jerk are constants; planning a second speed curve from the candidate state to the parking state based on the speed boundary constant constraint planning model, comprising: solving the speed boundary constant constraint planning model according to the first solving result to obtain a second solving result; obtaining a second speed curve from the candidate state to the parking state based on the candidate state and the second solving result.

12. The speed scheduling method of claim 11, wherein, planning a first speed curve from a current state to a terminal state based on the distance boundary constant constraint planning model, comprising: solving the distance boundary constant constraint planning model according to the second solving result to obtain a third solving result; According to the current state and the third solving result, a first speed curve with the shortest time from the current state to the terminal state is obtained.

13. The speed scheduling method of claim 12, wherein, The speed planning model is: wherein A i is the ith behavior vector, Δt i is the duration of the ith behavior vector, j i is the jerk of the vehicle in the ith behavior vector, is the set of real numbers, N-1 is the number of state vectors, N is the number of states of the vehicle, j max is the upper limit of jerk in speed planning, X i = [t i , s i , v i , a i ] T is the i-th state vector of the vehicle, s i is the travel distance of the vehicle at time t i , v i is the speed of the vehicle at time t i , a i is the acceleration of the vehicle at time t i , X i+1 = f(X i , A i ) is the state transition equation of the vehicle, s1, v1, a1 are variables representing the travel distance, speed and acceleration at the starting point, s S , v S , a S are the determined travel distance, speed and acceleration at the starting point, s N , v N , a N are variables representing the travel distance, speed and acceleration at the end point, s E , v E , a E are the determined travel distance, speed and acceleration at the end point, 0≤v i ≤v max (s i ) is the speed limit, v max (s i ) is the maximum speed at travel distance s i , a min ≤a i ≤a max is the acceleration limit, a min is the minimum acceleration, a max is the maximum acceleration.

14. The speed scheduling method of claim 13, wherein, The steady constraint programming model of the distance boundary is: where t N is the time to reach the end state.

15. The speed scheduling method of claim 13, wherein, The constant constraint programming model of the speed boundary is:

16. The speed scheduling method of claim 13, wherein, The acceleration boundary speed planning model is:

17. A speed planning device characterized by comprising: The method comprises the steps of: a memory for storing a computer program; a processor for implementing the steps of the speed planning method according to any one of claims 1 to 16 when the computer program is executed.

18. A vehicle characterized by comprising: The speed planning device according to claim 17 is included.

19. A readable storage medium, characterized by, The readable storage medium stores a computer program, and the computer program is executed by the processor to implement the steps of the speed planning method according to any one of claims 1 to 16.

Images