Methods and systems for controlling vehicles performing tasks

Abstract

A method for controlling a vehicle (101) performing a task, the vehicle (101) including a first power source and a second power source (ICE, EM) for driving the vehicle (101) itself, wherein the first power source includes an engine configured to generate power from fuel and an after-treatment system (ATS) coupled to the internal combustion engine, the method comprising the steps of: solving a convex first optimal control problem based on a mathematical model of the vehicle (101), the first optimal control problem relating to a set of state variable constraints and a cost function of the after-treatment system (ATS), the cost function having discrete variables (i gb i ATS b ATS ) and control variables (P, i) of continuous variables (P) gb i ATS b ATS The solution includes discrete variables (i) gb i ATS b ATS The initial determination of (110; 210) and the iterative execution of the following steps: minimizing the cost function relative to the continuous variable (P) after replacing the discrete variable (111; 212); updating (114; 221) the discrete variable (i) gb i ATS b ATS ); and verify that the convergence criteria of (112; 212) are met.

Description

[0001] Cross-references to related applications

[0002] This patent application claims priority to Italian Patent Application No. 102020000011254, filed on May 15, 2020, the entire disclosure of which is incorporated herein by reference. Technical Field

[0003] The present invention relates to methods and systems for controlling vehicles performing tasks, particularly through control by means of optimal control inputs determined to minimize fuel consumption and / or pollutant emissions. Background Technology

[0004] Some vehicles are known to have high flexibility because they can meet their power supply needs by utilizing at least one first power source and at least one second power source in series and / or in parallel.

[0005] Typically, the first power source supplies power from the first energy storage device and cannot replenish energy, while the second power source supplies power from the second energy storage device and can provide energy to the second power source.

[0006] Typically, the primary energy storage device is a fuel tank, and the primary power source uses fuel from the fuel tank to generate energy. Typically, the primary power source is an internal combustion engine, but other sources can also be considered, such as fuel cells that operate on gaseous fuels.

[0007] In other possible cases, such as when the first energy storage device is represented by an overhead line and the first power source is a power converter, the first power source is not dependent on fuel.

[0008] On the other hand, the second power source is usually an electric motor / generator, and the second energy storage device is a battery.

[0009] It is also known that the aforementioned vehicle is equipped with a control system adapted to optimally manage the distribution of required power among available power sources.

[0010] Specifically, when the first energy storage device is a fuel tank and the second energy storage device is a battery, the control system operates to minimize fuel consumption while satisfying constraints on the battery's charge level.

[0011] Sometimes, minimization is also achieved by taking into account constraints on the level of pollutant emissions, such as nitrogen oxide (NOx) emissions.

[0012] Some known control systems rely on prior information about the tasks that the vehicle must perform.

[0013] The term "task" is used here to identify the completion of a vehicle's journey along a predetermined path or route.

[0014] Such prior information typically includes the required speed of the vehicle and the characteristics of the route to be followed, such as road gradient and surface conditions.

[0015] To determine the optimal input for controlling the power flow within a vehicle, known control systems employ optimization methods based on the construction and solution of optimal control problems.

[0016] The latter optimal control problem is posed as the constrained minimization of a specific cost function. In particular, the cost to be minimized typically includes a term representing the total fuel consumption.

[0017] The above terms are accurately modeled using nonlinear and nonconvex functions, which typically depend on a large number of dynamic state variables and optimization variables.

[0018] For example, state variables may include quantities that indicate the operation of the engine and / or the engine-associated aftertreatment system (ATS).

[0019] However, given the complexity of the underlying model, known optimization methods are difficult to implement on computers, especially in real time.

[0020] In fact, the complexity of the cost function and the constraints to be satisfied leads to the need for a large amount of time-consuming computation to solve the relevant optimal control problems.

[0021] In particular, the complexity increases significantly when considering models of aftertreatment systems and pollutant emissions.

[0022] Therefore, it is necessary to provide an optimization method for controlling vehicles performing tasks, which takes into account pollutant emission models, and whose actual implementation on relevant systems requires less computational effort compared to the optimization methods mentioned above.

[0023] The purpose of this invention is to meet the above-mentioned needs. Summary of the Invention

[0024] The aforementioned objective is achieved by the methods and systems for controlling a vehicle as claimed in the appended claims.

[0025] The dependent claims set forth specific embodiments of the invention. Attached Figure Description

[0026] To better understand the present invention, preferred embodiments are described below by way of non-limiting examples with reference to the accompanying drawings, in which:

[0027] · Figure 1 A vehicle including a system for controlling the vehicle according to the present invention is illustrated schematically;

[0028] · Figure 2 This illustrates the control method according to the present invention. Figure 1 A block diagram of the method for the vehicle;

[0029] · Figure 3 This illustrates an embodiment of the invention for execution. Figure 2 A flowchart illustrating the operational plan for one step of the method;

[0030] · Figure 4 This illustrates another embodiment of the invention for execution. Figure 2 A flowchart illustrating the operational plan for one step of the method;

[0031] · Figure 5 It is shown Figure 2 The method involves surface maps of non-convex mappings between quantities; and

[0032] · Figure 6 This demonstrates how adaptive adjustments can be made. Figure 5 The surface map obtained by the region-wise convex mapping shown in the figure is a map of the region-wise convex mapping. Detailed Implementation

[0033] exist Figure 1 In the accompanying drawings, reference numeral 101 indicates a vehicle that includes a first power source or generator and a second power source or generator.

[0034] The vehicle 101 also includes a first energy storage device and a second energy storage device, which are respectively coupled to the first power source and the second power source, and the first power source and the second power source draw energy from the first energy storage device and the second energy storage device to drive the vehicle 101.

[0035] The first power source consumes only energy from the first energy storage device, while the second power source can provide energy to the second energy storage device.

[0036] Vehicle 101 is a hybrid electric vehicle (HEV), wherein, in particular, the second energy storage device and the second power source include a battery BY and an electric motor EM, respectively.

[0037] On the other hand, the first energy storage device and the first power source respectively include a fuel tank (not shown) and an engine ICE, or more precisely, an internal combustion engine.

[0038] The motor EM and the battery BY are coupled to each other, for example, via a DC / DC converter, so that the motor EM is powered by the battery BY.

[0039] Vehicle 101 also includes an aftertreatment system (ATS) coupled to the engine ICE to receive exhaust gases from the engine ICE and remove pollutants from the received exhaust gases. The aftertreatment system (ATS) is of a known type and includes, for example, a particulate filter (PF) and a catalyst for polluting gases (e.g., nitrogen oxides or NOx). Specifically, the catalyst is defined by a selective catalytic reduction (SCR) system, and more particularly, is provided via an aqueous urea solution.

[0040] Vehicle 101 also includes an axle with a pair of traction wheels WL and a vehicle differential D, the traction wheels WL being coupled to the vehicle differential D. The axle also includes a pair of friction brakes BK, each of which is coupled to the traction wheels WL to apply braking torque to the traction wheels WL.

[0041] In addition, vehicle 101 includes a torque distributor TS, which is coupled to both the engine ICE and the electric motor EM to receive engine power and electric motor power from the engine ICE and the electric motor EM respectively, and accordingly provides the sum of the received power to the differential D in an input manner.

[0042] In this way, the wheel WL can be driven by both the engine ICE and the electric motor EM in parallel. In other words, vehicle 101 is a parallel HEV.

[0043] Although the following description will only refer to parallel HEVs without any loss of generality, the concepts disclosed below can be easily converted into series HEVs or any other type of HEV without any effort.

[0044] Vehicle 101 also includes a clutch CL and a transmission GB, which selectively connect the engine ICE to the torque distributor TS. Specifically, the transmission GB has an input shaft (not shown) that can be selectively driven by the engine ICE via the clutch CL, and an output shaft (not shown) that is directly coupled to the torque distributor TS.

[0045] More specifically, the transmission GB introduces multiple gear ratios between the engine ICE and the torque distributor TS. In particular, six additional gear ratios are considered.

[0046] Additionally, vehicle 101 includes a transmission such as a gear reducer GR, which couples the motor EM to the torque distributor TS. Specifically, the gear reducer GR introduces a fixed gear ratio between the motor EM and the torque distributor TS.

[0047] In addition, vehicle 101 includes a control unit ECU, specifically an on-board computer, which is coupled to battery BY, motor EM, transmission GB, clutch CL, internal combustion engine ICE, torque distributor TS, and also specifically coupled to friction brake BK.

[0048] The control unit (ECU) is configured to receive and store multiple parameters about a given task to be performed by the vehicle 101.

[0049] Based on these parameters, the control unit (ECU) is configured to determine the power to be supplied by the engine (ICE) and the electric motor (EM), and is configured to select the gearbox (GB) corresponding to the appropriate gear ratio in order to ensure task completion, minimized fuel consumption, and satisfaction of one or more constraints.

[0050] Advantageously, one of the constraints relates to the total mass of pollutants, such as nitrogen oxides or NOx, emitted by vehicle 101 at the end of the mission.

[0051] The control unit (ECU) stores the above constraints, as well as another constraint regarding the battery charge level at the end of the mission.

[0052] In addition, the control unit ECU preferably stores further constraints regarding the physical limitations of the vehicle 101.

[0053] The control unit (ECU) receives and stores the above parameters before the mission begins and uses them as a function of the variable z, which is associated with the actual position of the vehicle 101 after the mission begins.

[0054] The variable z can directly represent the actual position of vehicle 101 along the route.

[0055] Otherwise, the variable z can be defined, for example, by the time elapsed since the start of the task. In fact, the elapsed time is associated with the actual position of vehicle 101; this association holds given the length / shape of the route and the land speed of vehicle 101, since, given the task, it is assumed that the length / shape of the route and the land speed of vehicle 101 are known or estimated a priori.

[0056] Therefore, the variable z can take the given interval [z0 z] f The values ​​between ], where z0 can be arbitrarily set to null, and z f It depends on the assumed characteristics of the route to be followed given the task.

[0057] Specifically, the control unit (ECU) receives and stores two parameters, Γ and v, which are functions of the variable z.

[0058] The parameters Γ and v indicate the assumed gradient of the route that vehicle 101 is to follow and the assumed land speed that vehicle 101 should use along the same route, respectively.

[0059] Specifically, the parameters Γ and v are defined by the aforementioned gradient and land speed of the route, respectively.

[0060] The control unit (ECU) uses parameters Γ and v to calculate other parameters as a function of variable z, which are useful for the control unit (ECU) to determine the appropriate power supply for wheel WL, as will become clearer below in this disclosure.

[0061] More specifically, the following holds true:

[0062]

[0063]

[0064] Where, γ fd r w ω ts F trac T trac,ts These represent the gear ratio of differential D, the radius of wheel WL, the angular rate at the output of torque distributor TS, the tangential traction force applied to the road by wheel WL, and the total traction torque output from torque distributor TS and input to differential D, respectively.

[0065] γ fd r w It is a constant value stored by the control unit ECU.

[0066] On the other hand, due to the mathematical model stored in the control unit ECU, F trac The calculation is performed by the control unit (ECU) based on parameters Γ and v.

[0067] For example, the control unit ECU stores F trac The following mathematical model:

[0068] F trac =F d (Γ,v)+F m (a)+F br (1)

[0069] Among them, F d F m F br These represent the resistances acting on vehicle 101 (e.g., aerodynamic drag, rolling resistance, and uphill driving force), the inertial forces acting on vehicle 101, and the braking force of brake BK, respectively. Parameter a is the acceleration of vehicle 101 and is calculated by the control unit ECU as the derivative of parameter v.

[0070] The control unit (ECU) is configured to control the amount of power to be directly supplied by the electric motor (EM) and to control the gear selection of the transmission (GB). These parameters, along with other parameters determined by the control unit (ECU), will function in this control process.

[0071] The control unit (ECU) is also configured to operate the engagement and disengagement of the clutch (CL), so that when the clutch (CL) is open, there is no need to select a gear.

[0072] The control unit (ECU) manages the clutch (CL) and transmission (GB) through the following operations:

[0073] • Update at least one first variable, which is discrete and represents the engagement or disengagement state of the clutch CL and the selected gear; and

[0074] • Generate a command signal associated with the at least one first variable to operate the clutch CL and the gearbox GB accordingly.

[0075] Specifically, the control unit ECU updates the discrete variable i gb To control the clutch CL and the gearbox GB.

[0076] In particular, discrete variable i gb Take an integer value, each integer value is associated with a corresponding gear (e.g., 1, 2, 3, 4, 5, 6), or with the disengaged state of the clutch CL (e.g., 0).

[0077] Obviously, when variable i gb When a value associated with the gear position is taken, the engagement state of the clutch CL is necessarily implicit.

[0078] Given the above, the first variable indicates the operation or driving mode of vehicle 101. In fact, for example, variable i... gb The value equal to zero indicates that vehicle 101 operates in fully electric mode. Furthermore, the possibility of selecting a gear indirectly suggests the adaptability of vehicle 101 to different road conditions.

[0079] The control unit (ECU) also manages the power supply to the torque distributor (TS) through the following operations:

[0080] • Update at least one second variable, which is continuous and indicates the power supplied by at least one of the electric motor (EM) and the engine (ICE); and

[0081] • Generate a command signal associated with the at least one second variable to operate at least one of the electric motor EM and the engine ICE accordingly.

[0082] Specifically, the control unit (ECU) updates a single continuous variable u to control the power supply from the electric motor (EM) and the engine (ICE).

[0083] Specifically, the variable u is defined by torque distribution, which is defined here as:

[0084]

[0085] Among them, T m,ts This represents the torque provided by the motor EM at the torque distributor TS, and T trac,ts This represents the total traction torque output from the torque distributor TS and input to the differential D.

[0086] More precisely, the following holds true:

[0087] T m,ts =(1-u)T trac,ts (6)

[0088] Among them, T e,ts This indicates the torque provided by the engine ICE at the torque distributor TS.

[0089] The variable u indicates the instantaneous power P provided by the motor EM, because, on the one hand, P is the product of T. m,ts ·ω ts Proportional, and on the other hand, T trac,ts ω ts They can be derived from the known parameters Γ and v using Equations 2 and 3, respectively.

[0090] In fact, the control unit ECU can calculate T according to Equation 4. m,ts The value of , and then by multiplying the product T m,ts ·ω ts The value of P is obtained by multiplying by a known proportional constant, which is stored in the control unit ECU and is essentially defined by the product of the respective efficiencies of the gear reducer GR and the torque distributor TS.

[0091] The control unit (ECU) updates the first and second variables by solving an optimal control problem. The optimal control problem involves determining multiple control variables based on variable z, which leads to the minimization of a selected cost function.

[0092] The control variables include a first variable and a third variable related to the second variable. Variable z is the independent variable of the optimal control problem solved by the control unit (ECU). Preferably, the third variable is defined by the instantaneous dynamic force P.

[0093] A cost function is determined for at least a portion of the task, and the cost function represents a first amount of energy consumed by the first power source after that portion has been completed.

[0094] More precisely, the first quantity indicates the amount of fuel consumed by vehicle 101.

[0095] The cost function is a convex function among the control variables.

[0096] Constraints can be expressed in mathematical terms as equations or inequalities that are functions of at least one of the control variables.

[0097] The constraints include at least one first-order dynamic constraint, which relates to the derivative of the state variable with respect to the variable z, as a function of at least one of the control variables.

[0098] Such a state variable indicates the operation of the aftertreatment system (ATS). Specifically, the state variable indicates the amount of pollutants released into the environment due to the operation of vehicle 101, more precisely, the amount of NOx. This state variable is denoted by the symbol... This indicates, and corresponds to, the total NOx emitted by the engine ICE minus the amount converted in the aftertreatment system ATS, particularly the selective catalytic reduction (SCR) system. This clearly clarifies why the variable... Instructs the operation of the post-processing system (ATS).

[0099] Specifically, the constraints include additional first-order dynamic constraints with respect to the other corresponding state variables.

[0100] For a given value of a control variable that is an integer or a discrete variable, each constraint in the constraints defines a convex set of permissible values ​​for at least one control variable that is a continuous variable and at least one state variable that is a state variable.

[0101] In this way, the optimal control problem solved by the control unit (ECU) leads to a mixed integer convex optimization problem.

[0102] In other words, for a given value of the control variable as an integer variable, the optimal control problem is convex because the domain of the admissible values ​​of the other control variables is convex, and because the cost function allows only one minimum value in that domain.

[0103] According to the disclosed implementation, the optimal control problem involves three state variables. The first state variable is... The third state variable is defined, and it also indicates the operation of the after-processing system (ATS). Specifically, the third state variable indicates the temperature of the after-processing system (ATS) and is denoted by the symbol θ. ATS Indicator. Variable It is θ ATS The function.

[0104] The second state variable indicates the amount of energy available from the second power source, enabling the second power source to generate power to drive vehicle 101. More specifically, the amount of energy is stored in battery BY and defines the state of charge of battery BY. Specifically, the state variable indicates the state of charge of battery BY and is indicated by the symbol ξ.

[0105] The derivative of ξ with respect to z is preferably modeled as a function of the power generated by the second power source.

[0106] More preferably, using an equivalent circuit model makes the following true:

[0107]

[0108] Among them, P b It draws power from the battery BY, Q max It is the battery's maximum capacity, and V oc It is the open-circuit voltage.

[0109] θ ATS The derivative with respect to z is preferably modeled as the enthalpy flow rate from the exhaust gas to the aftertreatment system (ATS). Heat loss from the aftertreatment system (ATS) to the environment A function of the difference between them.

[0110] More preferably, the following holds true:

[0111]

[0112] Where, m ATS and c ATS This refers to the total mass and specific heat capacity of the after-treatment system (ATS). Heat loss. and θ ATS With ambient temperature θ amb The difference between them is proportional, where the proportionality constant is the heat transfer coefficient α. h1 Multiply by the outer surface S of the subsequent processing system ATS.

[0113] The derivative with respect to z is preferably modeled as θ ATS and the mass flow rate of pollutants (specifically, NOx) through the aftertreatment system ATS. function In other words, the following is true:

[0114]

[0115] Therefore, equations 4a, 6a, and 18 define three dynamic constraints.

[0116] The constraints also include at least one first static equality constraint reflecting the mathematical model of vehicle 101, the mathematical model of vehicle 101 including, for example, F in Equation 1. trac The mathematical model. Such a first equality constraint includes two members, which in turn include convex functions of at least one of the first and third variables.

[0117] Therefore, given the characteristics and constraints of the cost function, and based on the known Pontyragin's minimum principle, the optimal control problem is solvable at least for the second variable. The control unit (ECU) solves the optimal control problem based on this principle.

[0118] Specifically, the cost function includes fuel power P. f From z0 to z f The integral of the fuel power P f Associated with a second quantity indicating the instantaneous power output from the first power source, and consisting of parameters Γ, v, and variable i. gb A function of P.

[0119] Preferably, P f Model in the following way:

[0120]

[0121] Among them, P e,ts This represents the power supplied by the engine ICE at the torque distributor TS, where κ0, κ1, and κ2 represent the output angular rate ω depending on the engine ICE. e The coefficient.

[0122] angular velocity ω e With ω ts Proportional, and the relevant proportionality constant is based on the selected gear of the transmission GB (i.e., determined by the discrete variable i). gb The variable is assumed to have a value. More precisely, ω ts equal to ω e Multiplied by variable i gb The value is associated with the gear ratio, multiplied by the efficiency of the clutch CL and the gearbox GB.

[0123] The control unit (ECU) stores the values ​​of κ0, κ1, and κ2 respectively and ω. e or ω ts The values ​​are associated with three mappings. These mappings depend on the engine ICE, and their derivation is performed experimentally—specifically by fitting experimental data using second-order polynomials.

[0124] Vehicle 101 includes a transducer H1 coupled to the engine ICE and control unit ECU; transducer H1 detects the indicated angular rate ω.e Another quantity and generates the following signal, which is related to such another quantity and is received by the control unit ECU for extracting the angular rate ω. e The actual value.

[0125] In addition, P b Model in the following way:

[0126] P b =αp 2 +βP (12)

[0127] Where α and β represent the output angular rates ω depending on the motor EM. m The coefficient ω. m With ω ts Proportional, and specifically by using ω ts It is obtained by dividing by the gear ratio of the gear reducer GR.

[0128] In addition, β also depends on the sign of P, which may be negative when the motor EM is operated as a generator.

[0129] Therefore, β is represented as a discontinuous function of P, taking the value β1 when P is greater than or equal to 0, and β2 otherwise.

[0130] The control unit (ECU) stores the values ​​of α, β1, and β2 respectively, along with ω. m or ω ts The values ​​are associated with three mappings. These mappings depend on the motor EM, and their derivation is performed experimentally—specifically by fitting experimental data using second-order polynomials.

[0131] Vehicle 101 includes a transducer H2 coupled to a motor EM and a control unit ECU; the transducer H2 detects the indicated angular rate ω. m Another quantity and generates the following signal, which is related to such another quantity and is received by the control unit ECU for extracting the angular rate ω. m The actual value.

[0132] Fuel Power P f It explicitly depends on the third variable P rather than the first variable (variable i) gb This is because the following direct relationship holds:

[0133] P e，ts =P ts -P (5)

[0134] Among them, P ts The power supplied at the torque distributor, along with v and F trac The product is consistent, including the braking force applied by the braking system BS.

[0135] The first and third state variables, which indicate the operation of the aftertreatment system (ATS), are functions of the power generated by the engine (ICE), and therefore are P. e,ts The function of θ. More precisely, the function of θ. ATS enthalpy flow Including P e,ts The function.

[0136] Specifically, the following equation holds true:

[0137]

[0138] Where δ0, δ1, and δ2 represent the output angular rates ω depending on the engine ICE. e The coefficients. The control unit ECU stores the values ​​of δ0, δ1, and δ2 respectively and ω. e or ω ts The values ​​are associated with three mappings. These mappings depend on the engine ICE, and their derivation was performed experimentally.

[0139] about mass flow rate of pollutants Including P e,ts The function. In particular, the following equation holds:

[0140]

[0141] Where v1 and v2 represent the output angular rates ω depending on the engine ICE. e The coefficients. The control unit ECU stores the values ​​of v1 and v2 respectively and ω. e or ω ts The values ​​are associated with two mappings. These mappings depend on the engine ICE, and their derivation was performed experimentally.

[0142] Therefore, for P e,ts and θ ATS This understanding enables us to assess the dynamics of state variables.

[0143] Expressed by Equation 18 With θ ATS and The functional relationships can be experimentally identified by interpolating data about the after-processing system (ATS). For example, such functional relationships can be stored in the control unit (ECU) as a mapping obtained experimentally, such as... Figure 5 The mapping shown in the image. However, Figure 5 The mapping is non-convex and therefore unsuitable for constrained convex problems. Therefore, the control unit (ECU) stores a convex approximation of the experimental mapping.

[0144] Preferably, the experimental map is divided into multiple convex regions, such that the control unit (ECU) stores the regional convex map obtained in this way.

[0145] More preferably, the convex region includes regions for values ​​below a first threshold θ. thr θ ATS The first region defined by the value of the first threshold θ thr Specifically, this corresponds to the temperature of the aftertreatment system (ATS) that is insufficient to cause a reduction in pollutant emissions (i.e., NOx reduction). Therefore, in this first region, Simply equal to

[0146] Other regions are characterized by values ​​greater than the first threshold θ. thr θ ATS The value is used to define these other regions. These other regions correspond to... The values ​​are in different ranges. The ranges do not overlap, but each range is adjacent to another range within the range.

[0147] In particular, such as Figure 6 As shown, there are four other areas. Figure 6 An example of a region-based convex map stored in the control unit (ECU) is shown. Clearly, Figure 6 The total number of regions shown is 5.

[0148] Each region can be represented by a pair of integer variables, such as binary variable b. ATS and discrete variable i ATS To identify.

[0149] Specifically, binary variable b ATS Take the first integer value (e.g., 0) to be independent of i ATS The chosen value is used to identify the first area where no pollutant reduction occurred, and a second integer value (e.g., 1) is taken based on i. ATS Identify other areas.

[0150] Similarly, discrete variable i ATS Take an integer value, and each integer value (e.g., 0, 1, 2, 3) is associated with a corresponding region in other regions.

[0151] Preferably, each region in the other regions is modeled using a set of linear functions, where each linear function has the following shape, more preferably:

[0152]

[0153] Where τ2, τ1, and τ0 represent the coefficients of one of the linear functions.

[0154] Conveniently, the control unit (ECU) stores the corresponding vector for each other region. T 2. T 1. T Triples of 0, the vector T 2. T 1. T The triplet of 0 includes the coefficients of the linear function used to model the other regions.

[0155] In this way, i ATS Each permissible value corresponds to a corresponding vector. T 2. T 1. T A triplet of 0.

[0156] In detail, the following formula holds true:

[0157]

[0158] Where, vector T 2. T 1. T 0 depends on i ATs Equation 17 specifically corresponds to Figure 6 The convex mapping shown.

[0159] Preferably, integer variable b ATS i ATS These are part of the control variables that constitute the optimal control problem solved by the control unit (ECU).

[0160] In addition to the first static equality constraint and dynamic constraints, constraints in optimal control problems can also include constraints on variables P and i. gb Static inequality constraints are set for feasible values ​​to ensure that the physical limitations of vehicle 101 are not violated.

[0161] In particular, the following expresses static inequality constraints:

[0162] ω m ∈[0, ω m，max (11j)

[0163] P∈[P min (ω m ,ξ), P max (ω m ,ξ)] (11k)

[0164]

[0165] P e，ts ∈[O, P e，ts，max (ω e (11m)

[0166] Based on the physical limitations of vehicle 101, subscripts min and max are added to the symbols to indicate the minimum and maximum permissible values ​​of the corresponding parameters, respectively.

[0167] Furthermore, the constraints specifically include another static inequality constraint, which is expressed as follows:

[0168] ξ(z fin )≥ξ0 (11e)

[0169] The latter inequality constraint requires that the charge level of battery BY at the end of the task cannot be lower than the charge level of battery BY at the start of the task, indicated by ξ0.

[0170] In addition, the constraints include another static inequality constraint (11h), which is specifically expressed as follows:

[0171]

[0172] This other constraint requires that the level of pollutants emitted cannot exceed a given threshold stored in the control unit (ECU).

[0173] In addition, the constraints include static equality constraints to set the initial values ​​ξ0, θ of the state variables. ATS，0 :

[0174] ξ(0)=ξ0 (11c)

[0175] θ ATS (0)=θ ATS，0 (11f)

[0176]

[0177] In addition, constraints may include one or more other static inequality constraints to set the allowable values ​​for state variables, such as:

[0178] ξ∈|ξ min ξ max | (11e)

[0179] In the disclosed implementation, the following static model constraints are also considered in the optimal control problem:

[0180] P b ≥αP 2 +β1P (19e)

[0181] P b ≥αP 2 +β2P (19f)

[0182]

[0183]

[0184]

[0185]

[0186]

[0187] in, and This indicates the lower and upper limits of the mass flow rate of emitted pollutants, i.e., the effective mass flow rate. Area (see) Figure 6 The lower and upper limits of ). Furthermore, It is a vector of linear functions used to model other regions of a convex mapping.

[0188] The control unit (ECU) solves the optimal control problem under the above constraints. Solving the optimal control problem involves determining the optimal control variables and the Lagrange multiplier. λ The determination of the optimal costate or vector.

[0189] Here, there are three common states λ. ξ λ θ , λ NOX Specifically with the three state variables ξ, θ ATS , Related. Common state λ ξ , λ NOx It is a constant or a piecewise constant, because ξ and The derivative with respect to z does not depend on ξ and It itself. On the other hand, the common state λ θ It is time-related.

[0190] The optimal control problem is a mixed-integer problem because it involves both discrete control variables (specifically, variable i) and... gb i ATS b ATS This also involves continuous control variables (especially a single variable, i.e., variable P). The discrete variables of the optimal control problem can be collected in a discrete vector. i middle.

[0191] The control unit (ECU) solves the optimal control problem iteratively. Essentially, the ECU first solves the optimal control problem only for continuous control variables and the common-state condition after replacing the discrete variables. Then, the ECU applies the Hamiltonian function associated with the problem. The optimal common state is replaced, and its minimum value is found based on the control variables. Therefore, the discrete variables are updated and replaced to find and update the optimal common state again. This process continues iteratively until the convergence criterion stored in the control unit (ECU) is met.

[0192] The first substitution of discrete variables is based on guesswork or initial determination using any known method.

[0193] When the convergence criterion is met, the control unit (ECU) outputs a Hamiltonian function. The control variables are minimized, and the latest updated common state is obtained. Furthermore, the control unit (ECU) calculates and outputs state variables based on the optimal common state and control variables. These state variables can be collected in a state vector. x middle.

[0194] In other words, solving the optimal control problem involves iterative execution:

[0195] • After replacing the discrete variables, the cost function is minimized relative to the continuous variables and subject to constraints, thus determining the optimal common state;

[0196] • Update discrete variables based on the determined optimal common state;

[0197] • Verify that the convergence criterion is met, and

[0198] • Repeat this process based on the updated discrete variables until the convergence criterion is met.

[0199] More specifically, the update of the discrete variables involves a new minimization of the cost function, which is achieved by minimizing the Hamiltonian function based on Potryagin's minimum principle. Minimize the Hamiltonian function. Specifically, it is preferred to differ from the exact Hamiltonian function H of the optimal control problem by lacking members that serve as static constraints and by neglecting dynamic constraints. In other words, the Hamiltonian function is approximated relative to the exact Hamiltonian function H in the sense of not considering static constraints and state dynamics.

[0200] After replacing the Hamiltonian function After finding the optimal common state, perform a new minimization.

[0201] according to Figure 3 In this implementation, solving the optimal control problem involves an initialization step (box 110), in which discrete vectors... i 0 is either guessed or determined. Typically, discrete vectors... i jThe superscript j indicates the number of iterations; a null value indicates the start of the iteration operation.

[0202] For example, the control unit (ECU) determines the discrete vector by solving another optimal control problem. i 0 This other optimal control problem involves the first and third variables, but not other control variables or state variables and constraints related to the operation of the post-processing system ATS.

[0203] More precisely, the control unit (ECU) can implement the processing disclosed in the SAE technical paper "Fuel-optimal power split and gear selection strategies for a hybrid electric vehicle" published by J. Ritzmann et al. in 2019.

[0204] Having determined the first and third variables, the control unit (ECU) also determines other control variables by positively evaluating the dynamics of the state variables. This results in a fully discrete vector. i 0 The determination of b. In fact, given the dynamic constraints, b... ATS and i ATS Depends on P e，ts Therefore, it depends on P.

[0205] Once the discrete vectors are determined i 0 The control unit (ECU) uses discrete variables to solve the problem. Figure 3 (Box 111) Optimal control problem, where the discrete variables are discrete vectors. i 0 The values ​​of the corresponding variables are then considered. Here, the optimal control problem is no longer a mixed integer problem, but rather one that involves only continuous variables. Therefore, the control unit (ECU) implements one of the known methods for solving continuous problems, such as, in particular, the direct method using multiple targets and interior point solvers.

[0206] Figure 3 Box 111 is preferably solved using vectors according to the solution of continuous problems. λ j and x j Output common state and state variables in the form of [format].

[0207] In addition, box 111 also outputs parameter E. f The parameter E f Indicates the total dissipated fuel energy during the mission, specifically as P throughout the mission. f The integral of a function.

[0208] Conveniently, at box 111, the constraints that are valid at the end of the task, namely the constraints of equations 11d and 11h, are replaced by soft constraints that include slack variables:

[0209] ξ(z fin )≥ξ0-∈ ξ (25)

[0210]

[0211] Where, ∈ ξ ,∈ NOx It is a slack variable.

[0212] Accordingly, in this case, the cost function also includes a terminal cost, which is a weighted sum of slack variables:

[0213]

[0214] Among them, w ξ and w NOx It's the weight.

[0215] Once the optimal common state is determined, the control unit (ECU) verifies that the convergence criterion is met (box 112).

[0216] The convergence criterion can be verified by one or more differences between the quantity computed at iteration j (e.g., common state) and the same quantity computed at the previous iteration j-1 or the next iteration j+1.

[0217] For example, a convergence criterion can be satisfied when the relative increment or decrement of the quantity obtained through iteration is lower than the corresponding tolerance.

[0218] Specifically, the ECU considers the convergence criterion satisfied when at least one of the following inequalities is met:

[0219]

[0220]

[0221]

[0222] More specifically, a convergence criterion is considered met when all of the above inequalities are satisfied.

[0223] The tolerances given in the right-hand members of the inequality are purely exemplary and without any loss of generality.

[0224] If the convergence criterion is not met, the control unit (ECU) can optionally dampen the optimal common state. λ j(Box 113). Damping is based on the damping factor ψ, in particular making the damping increase with j, and more specifically until the maximum damping threshold is reached.

[0225] For example, the control unit (ECU) dampes the optimal common state according to the following rules. λ j :

[0226]

[0227] Here, the damping factor ψ is a multiplicative factor between the minimum damping value (greater than zero) and 1. The damping factor ψ decreases as j decreases until the minimum damping value is reached.

[0228] In particular, the following holds true:

[0229]

[0230] After damping, the control unit (ECU) will achieve the optimal damping common state. Replace with Hamiltonian function And determine that the same Hamiltonian function Minimize P and i gb The value (box 114).

[0231] In particular, the Hamiltonian function Approximately or limited to:

[0232]

[0233] Equation 29 exemplarily illustrates a function lacking constraints representing physical limitations of vehicle 101, such as an indicator function. This is due to the optimal commonality, which is particularly damped here. λ j Based on the assessment, such constraints are considered to have been actually met.

[0234] With variables P and i gb Make Hamiltonian function Minimize, while integer variable b ATS and i ATS By making P and i gb The dynamics of the state variables are minimized during forward simulation to determine θ. ATS The value is provided by box 111.

[0235] Preferably, the Hamiltonian function is found statically. Minimize P and i gb The value, that is, through targeting i gb A grid evaluation Hamiltonian function for all possible values ​​and multiple quantized values ​​of P. Alternatively, make the Hamiltonian function Minimize P and i gb The value of ξ can also be found through dynamic programming or any other known optimization method. For example, this becomes advantageous in guaranteeing convergence when the constraints on the charging state ξ are strict.

[0236] Box 114 outputs the optimal discrete variables, especially the optimal discrete vector i. j+1 The optimal discrete variables collected in the process, the optimal discrete vector i j+1 The next iteration for solving is shown in box 111.

[0237] If the convergence criterion is met in box 112, the solution terminates (box 115), and the optimal control variable P can be output. * , and preferably output Similarly, the corresponding state variables can also be output. x * and costate λ * .

[0238] Figure 4 Another embodiment of the invention for solving optimal control problems using an iterative method is shown. In block 210, the control unit (ECU) determines or guesses the common state. λ 0 Then, the control unit ECU knows λ 0 Determine the initial discrete vector under the following circumstances i 0 or i j (Here j = 0), for example, by solving the optimal control problem through dynamic programming (box 211) or any other known method (e.g., static optimization as in box 114). Then, knowing... i j In this case, the control unit ECU executes block 212, which is equivalent to block 111. Therefore, the output of the latter block 212 is the optimal common state. λ j .

[0239] Therefore, the control unit (ECU) verifies the convergence criterion in block 213, as in block 112. If the convergence criterion is not met, the damped optimal common-state is verified in block 214, as in block 113. λ j Therefore, a damped common state is provided in the input to box 211 for the repetition of the operation.

[0240] If the convergence criteria are met, the solution is stopped in box 215, as in box 115.

[0241] After obtaining the solution to the optimal control problem, the control unit (ECU) updates the second variable based on the optimal value of the third variable. In this way, the optimal value of the second variable is obtained. Specifically, the following equation holds:

[0242] u * =P * / P ts (38)

[0243] Furthermore, the output state variable corresponds to the optimal trajectory of the third quantity indicating the same state variable.

[0244] Control Unit (ECU) Figure 2 The method outlined herein for controlling vehicle 101 during a mission is essentially based on solving an optimal control problem, particularly as already disclosed in detail.

[0245] The control unit (ECU) includes a first logic block RTG or reference trajectory generator, which outputs an optimal trajectory starting from the input parameters Γ and v. x * More specifically, the first logic block RTG solves the optimal control problem for the entire task in the manner detailed above.

[0246] Specifically, the first logic block RTG solves the optimal control problem to evaluate the optimal common state. λ * .

[0247] The operation of the first logic block RTG is based solely on prior information about the task to be performed by vehicle 101. In fact, the only required input is the prior known parameters Γ and v stored by the control unit ECU.

[0248] In principle, the first-frame RTG can also calculate the optimal torque distribution u according to Equation 38. * And correspondingly output the same optimal torque distribution u. * and optimal discrete variables

[0249] According to the embodiment not shown, the output u * and The ECU (Electronic Control Unit) controls the electric motor (EM) and engine (ICE) on one hand, and the clutch (CL) and transmission (GB) on the other. In this case, vehicle 101 will be under open-loop control.

[0250] The vehicle 101 also includes a transducer device T1 coupled to the battery BY and the control unit ECU. The transducer device T1 is configured to detect the actual value of a third quantity indicating a state variable, and is configured to generate a signal related to the detected value.

[0251] The control unit (ECU) receives the signal generated by the transducer device T1, and for each value of variable z, extracts the corresponding actual value associated with the actual state variable from the signal. x act .

[0252] The control unit (ECU) also includes a second logic block, MPC or model predictive controller, which receives actual values. x act Parameters Γ, v and reference trajectory x * As input, to determine the updated reference trajectory x + and the updated optimal common state λ + .

[0253] In the following text, superscript + With superscript * The meaning is similar, and refers to the variable updated by the second logic box MPC. By simply considering the appropriate superscripts, all the equations exposed above also apply to the optimal and updated variables.

[0254] The second logic block, MPC, repeatedly solves the optimal control problem for the shift interval of variable z. The shift interval corresponds to the actual value. x act The actual value of z begins, and has a given size that is larger than the size of the entire interval related to the entire task.

[0255] Specifically, the second logic block MPC repeatedly solves problems that are different from those of the first logic block RTG for the interval or prediction range of z.

[0256] Furthermore, the problem solved by the second logic block MPC differs from the problem solved by the first logic block RTG in that it corresponds to the actual value x act The initial set of state variables.

[0257] The size of the prediction range should be selected to enable effective control of vehicle 101. In fact, the larger the prediction range, the longer the computation time for updating the output of the second logic block MPC.

[0258] With variable z representing the distance the vehicle travels along the route, the minimum possible size of the prediction range should be the maximum distance that vehicle 101 can travel during the update time.

[0259] Furthermore, the second logic box MPC also, based on the aforementioned public information, or more specifically, based on... Figure 3 or Figure 4An iterative method is used to evaluate the updated variable P. + , i + .

[0260] Similar to the first logical block RTG, the second logical block MPC calculates and updates the trajectory. x + and output updated λ + .

[0261] According to an embodiment not shown, the second logic block MPC also calculates the updated torque distribution u according to Equation 38. + and updated discrete variables The updated torque distribution u + and updated discrete variables It may be output and used for optimal torque distribution u * and optimal discrete variables Replacement.

[0262] According to the latter embodiment (not shown), the output u + and The ECU (Electronic Control Unit) controls the electric motor (EM) and engine (ICE) on one hand, and the clutch (CL) and transmission (GB) on the other. In this case, the vehicle 101 will be under closed-loop control.

[0263] Vehicle 101 also includes another transducer device T2, which is coupled to the control unit ECU and configured to detect a fourth quantity indicating the total power required at the torque distributor TS, and is configured to generate a signal associated with this fourth quantity. For example, transducer device T2 is coupled to the accelerator (not shown) of vehicle 101; alternatively, transducer device T2 is coupled to the torque distributor TS.

[0264] The control unit ECU receives the signal generated by the transducer device T2 and extracts the corresponding actual value P from the signal for each value of variable z, which is associated with the actual total power required at the torque distributor TS. req .

[0265] The control unit (ECU) includes a third logic block (OCL) or an optimal controller, which receives updated common-state parameters. l + and actual value P req As input, the updated optimal control is output. u c .

[0266] More precisely, the third logic block OCL uses inputs respectively. l + and P reqAs the common state and the required driving force, to solve the optimal control problem.

[0267] Specifically, the third logic box OCL utilizes λ + After replacing the common state and considering P ts equals P req Instead of functions of Γ and v, make the Hamiltonian function The optimal control problem is solved by minimization.

[0268] The solution is the optimal control. u c The optimal control u c The first and second variables are used by the control unit ECU for optimal updates, in order to control the motor EM and engine ICE on the one hand, and the clutch CL and transmission GB on the other hand.

[0269] The aforementioned control unit ECU, together with transducer devices T1 and T2, constitutes part of a system for controlling vehicle 101.

[0270] The logic blocks RTG, MPC, and OCL define a multi-layered control structure.

[0271] In view of the foregoing, the advantages of the method and system for controlling vehicle 101 according to the present invention are obvious.

[0272] Specifically, the disclosed method is computationally inexpensive, especially compared to known methods, and the disclosed system can implement the disclosed method in real time.

[0273] Although it is significantly faster, the control accuracy is at least comparable to that of known methods.

[0274] In particular, compared to known methods, the disclosed multi-layer control structure enables the mixing of prior known information with real-time information for optimal control of vehicle 101 with improved effectiveness.

[0275] The disclosed methods and systems enable not only the control of power distribution between the electric motor (EM) and the engine (ICE), but also the control of the amount of pollutants emitted by introducing state variables related to the operation of the aftertreatment system (ATS).

[0276] By using convex functions to model the dynamics of these state variables in a simple way, the optimal control problem can be solved in a reasonable time and the identified optimal value is guaranteed to be the global optimum.

[0277] The above approximation of the dynamics leads to the introduction of discrete variables into the cost function of the optimal control problem, which is entirely manageable based on the solution of the same problem disclosed.

[0278] according to Figure 3 The implementation method involves only two basic stages: solving problems involving only continuous variables and static optimization problems. These two stages can be easily completed in a simple manner without requiring a large amount of computation.

[0279] Common-mode damping enables the avoidance of oscillations during the iteration of the method; in particular, the choice of increasing damping with iteration enables rapid convergence in the early iterations and avoids unwanted oscillations around the optimal value.

[0280] Using a general variable z makes it possible to reduce the dynamic impact of disturbances on the control of vehicle 101. This variable z may represent the distance traveled by vehicle 101 along the route, rather than the time.

[0281] Finally, the chosen constraints enable a simple solution to the optimal control problem.

[0282] Obviously, modifications can be made to the described methods and systems without exceeding the scope of protection defined in the claims.

[0283] For example, numerical adjectives are purely conventional; specifically, the third and second variables can agree with each other. In other words, the optimal dynamic P * It can be used directly to control vehicle 101.

[0284] The electric motor EM is an example of a second power source for vehicle 101. Therefore, the electric motor EM can more generally be replaced by different power sources with appropriate direct adaptability.

[0285] For example, the motor EM can be replaced by a converter configured to draw energy from an overhead line. In such a case, the parameters, quantities, and variables related to the motor EM and the battery BY must be replaced with the corresponding parameters, quantities, and variables related to the converter and the overhead line.

[0286] Furthermore, internal combustion engines (ICEs) can be replaced by fuel cells that run on gaseous fuels or other types of fuel power sources.

[0287] The gear reducer GR can also introduce multiple gear ratios; in this case, the variable i can be used. gb The possible values ​​are taken into account to account for the existence of other gear ratios.

[0288] Brake BK may be an emergency brake and remain inactive during normal operation of vehicle 101; in such cases, the mathematical model of vehicle 101 must be modified accordingly.

[0289] Instead of requiring charging to maintain operation, Equation 11d can be rewritten for a predetermined charging state of battery BY that differs from its initial charging state. This is particularly promising for plug-in hybrid electric vehicles where it is expected that battery BY will be depleted during a mission, as it can be recharged using electricity from the grid at the destination.

[0290] Ultimately, the optimal control problem can involve only one constraint, forming a constraint set with only one element. Similarly, the problem can involve only one state variable, forming a state variable set with only one element. More generally, the set includes one or more members.

Claims

1. A method for controlling a vehicle (101) performing a task, the vehicle (101) comprising: A first power source and a second power source (ICE, EM) for driving the vehicle (101) itself, wherein the first power source includes a heat engine (ICE) configured to generate power from fuel and an aftertreatment system (ATS) coupled to the engine (ICE). The method includes the following steps: • Solve the convex first optimal control problem based on the mathematical model of the vehicle (101), the first optimal control problem involving the set of state variables. A constraint set and a cost function, the cost function having at least one discrete variable (i gb i ATS b ATS ) and control variables (P, i) of at least one continuous variable (P) gb i ATS b ATS ), The solution includes at least one discrete variable (i) gb i ATS b ATS The initial determination or guess (110; 210, 211) and the iterative execution of the following steps: a) After replacing the at least one discrete variable, relative to the at least one continuous variable (P) and subject to the set of constraints, based on the set of state variables. The associated optimal set of common-state variables (λ) ξ , λ θ , λ NOx The determination of ) minimizes the cost function (111; 212); b) Based on the optimal set of common-state variables determined in step a), update at least one discrete variable (i) of (114; 211). gb i ATS b ATS ); c) Based on the results of step a), verify that the convergence criteria (112; 213) are satisfied, and d) If the convergence criterion is not met, repeat step a) based on the update in step b). Otherwise, exit from the iterative execution, and • The vehicle (101) is controlled based on the solution of the first optimal control problem. Among them, the set of state variables Includes a first state variable indicating the operation of the post-processing system (ATS). The control variables include a first discrete variable (i) that indicates the operating mode of the vehicle (101). gb ) and a first continuous variable indicating the power (P) supplied by at least one of the first power source and the second power source (ICE, EM); Furthermore, the cost function is determined for at least a portion of the task and represents a first quantity, which indicates the energy (E) consumed by the first power source (ICE) after the completion of at least a portion of the task. f ); The constraint set includes constraints on the first state variable. At least one endpoint constraint that can take a permissible value at the end of at least a portion of the task.

2. The method according to claim 1, wherein, Step b) includes: e) Based on Pontryagin's minimum principle and Hamiltonian function Minimizing the cost function again relative to the control variables, the Hamiltonian function It has a set of common-state variables that are replaced according to the optimal set of common-state variables; Wherein, the Hamiltonian function With regard to the set of state variables This is related to the convex second optimal control problem of the cost function. Wherein, the at least one discrete variable (i gb i ATS b ATS The update is based on the result of step e).

3. The method according to claim 1, wherein, The optimal common-state variable set is damped based on a damping factor (ψ), such that the damping increases with the number of recursions in step a) until the maximum damping threshold is reached.

4. The method according to claim 1, wherein, The at least one endpoint constraint is defined by a soft constraint, or includes at least one slack variable (∈) to be added to the cost function with corresponding weights. ξ ,∈ NOx ).

5. The method according to claim 1, wherein, The first discrete variable takes one of the following: a first value associated with the disengaged state of the clutch (CL) of the vehicle (101); and a plurality of second values ​​associated with the various selectable gears of the transmission (GB) of the vehicle (101).

6. The method according to claim 1, wherein, The set of state variables This includes a second state variable (ξ), which indicates the amount of power (P) available from the second power source for generating the power used to drive the vehicle (101). b The amount of energy; The total derivative of the second state variable (ξ) with respect to the independent variable (z) of the first optimal control problem is the power (P) generated by the second power source. b The function of ).

7. The method according to claim 2, wherein, The constraint set also includes one or more inequality constraints that express the physical limitations of the vehicle (101).

8. The method according to claim 7, wherein, The Hamiltonian function It is not affected by the one or more inequality constraints mentioned above.

9. The method according to claim 1, wherein, The mathematical model includes multiple prior-known first parameters (Γ, v) associated with the task. The mathematical model enables the determination of a second quantity (P) that is a function of the first parameter (Γ, v), the first discrete variable (i), and the first continuous variable (P). f ), the second quantity (P) f This indicates the instantaneous power output by the first power source. Furthermore, the first amount indicating the energy consumed by the first power source (ICE) is the second amount (P). f The function of ).

10. The method according to claim 9, wherein, The first quantity and the second quantity respectively indicate the fuel consumed by the vehicle (101) and the instantaneous fuel power.

11. The method according to claim 10, wherein, The mathematical model includes: • Fuel-power model, for the fuel-power model, the second quantity (P) f ) is modeled as a polynomial with a second parameter, where the coefficients (κ) o κ1, κ2) depend on the output angular rate (ω) of the engine (ICE). e The second parameter indicates the driving power (P) supplied by the engine (ICE). e,ts ); • Engine power model, in which the second parameter is modeled as a function of the first continuous variable (P) and a third parameter, the third parameter indicating the total driving power (P) required to drive the vehicle (101). req );as well as • Driving dynamics model, in which the third parameter is modeled as a function of the driving resistance of the vehicle (101), the driving resistance depending on the first parameter (Γ, v).

12. The method according to claim 9, wherein, The first state variable The set of state variables indicates the cumulative mass of contaminants in the aftertreatment system (ATS), and wherein... It also includes a third state variable (θ) indicating the temperature of the after-treatment system (ATS). ATS ); The first state variable The total derivative of the independent variable (z) with respect to the first optimal control problem is the third state variable (θ). ATS ) and the mass flow rate of the pollutant The function, Wherein, the mass flow rate Modeled as the second quantity (P) f A polynomial function of ), where the coefficients depend on the output angular rate (ω) of the engine (ICE). e ).

13. The method according to claim 12, wherein, The third state variable (θ) ATS ) and the mass flow rate The function defines the third state variable (θ) as follows. ATS ) and the mass flow rate A pair of values ​​with the first state variable The mapping associated with the value, The mapping is divided into multiple sub-regions, each modeled using one or more linear functions, and consisting of a first integer variable and a second integer variable (i). ATS b ATS A pair of variable identifiers defined by ) The first integer variable (b) ATS Take the third state variable (θ) as a constraint. ATS The values ​​of each separate range of ) The second integer variable (i) ATS Take the specified mass flow rate The values ​​of each separate range, The control variables include the first integer variable and the second integer variable as the second discrete variable and the third discrete variable.

14. The method according to claim 13, wherein, The third state variable (θ) ATS The total derivative with respect to the independent variable (z) is the enthalpy flow rate from the exhaust gas of the engine (ICE) to the aftertreatment system (ATS). Heat loss from the aftertreatment system (ATS) to the environment outside the vehicle (101) The function of the difference between them Wherein, the enthalpy flow rate Modeled as the third state variable (θ) ATS ) and the second quantity (P) f A linear function of ω, which has a value depending on the output angular rate (ω). e The coefficient of ) And among them, the heat loss Modeled as the third state variable (θ) ATS A linear function of .

15. The method according to claim 1, further comprising the following step: • Determine the optimal set of state variables based on the solution to the first optimal control problem. x * );as well as • Generate the optimal set of state variables ( x * The reference trajectory of the related third set.

16. The method of claim 15, further comprising the step of: • Obtain the actual current value set of the third quantity set ( x act ); Solve the convex third optimal control problem based on the mathematical model, and the third optimal control problem involves the set of state variables. The constraint set, and the requirement for the state variable set Initially corresponding to the actual current value set ( x act Another constraint and another cost function having the control variables; • Update the optimal common-state variable set based on the solution to the third optimal control problem; and The third optimal control problem includes iteratively executing steps a) to d) by replacing the cost function with the other cost function; The other cost function is determined for a reduction of at least a portion of the task, and represents the first amount.

17. A system for controlling a vehicle (101) performing a task, the vehicle (101) including a first power source and a second power source (ICE, EM) for driving the vehicle (101) itself; The system includes a control unit (ECU) programmed to implement the method of any one of claims 1 to 16.

Images