CN108437991A - A kind of intelligent electric automobile adaptive cruise control system and its method - Google Patents

Abstract

An intelligent electric vehicle adaptive cruise control system and a method thereof relate to the auxiliary control of vehicle safety driving. The system includes an information acquisition module, a working mode selection module, a control function switching module, an expected torque calculation module, a converter module and an actuator module. A safety distance control strategy and a drive/brake switching strategy are proposed, and the neurofuzzy-based inversion sliding mode adaptive cruise tracking mode control method can solve the nonlinear problem of the speed control of the adaptive cruise system of electric vehicles and the strong coupling of the system state To ensure the ability of the vehicle to track the vehicle ahead during adaptive cruise driving, improve the utilization rate of traffic roads and the safety and comfort of vehicle driving.

Description

A smart electric vehicle adaptive cruise control system and method thereof

technical field

The invention relates to automobile safety driving auxiliary control, in particular to an intelligent electric vehicle self-adaptive cruise control system and a method thereof.

Background technique

Adaptive cruise control (Adaptive Cruise Control, ACC) is a car safety assisted driving system, which adds the function of automatically tracking the vehicle in front on the basis of the traditional cruise control function, so that the vehicle and the vehicle in front can maintain a suitable safe distance, which can improve the safety of the vehicle. Driving safety, ease traffic congestion, and reduce the burden on drivers. At present, a large number of researches on adaptive cruise control systems and control methods focus on traditional vehicles, but less on smart electric vehicles. Intelligent electric vehicle is a current research hotspot, which has the advantages of energy saving and environmental protection. The adaptive cruise function of intelligent electric vehicle can realize the safe, comfortable and energy-saving driving of the vehicle, and at the same time can improve the utilization rate of traffic roads. Therefore, the adaptive cruise of intelligent electric vehicle The research on the system and control method is of great significance.

The adaptive cruise system of intelligent electric vehicles has the characteristics of nonlinearity and uncertainty. The vehicle has acceleration and deceleration processes in the process of adaptive cruise tracking mode control. When establishing a longitudinal dynamic model, the slip rate becomes an important factor. The adaptive cruise dynamics model with slip rate has the nonlinear problem of speed control, and there is a strong coupling relationship between state variables. Therefore, simple nonlinear system linearization methods and simple control methods are difficult to meet the control requirements of the system. Document 1 (Kayacan E.Multi-objective H _∞ Control for String Stability of Cooperative Adaptive Cruise Control Systems[J].IEEE Transactions on Intelligent Vehicles,2017,2(1):52-61.) Using multi-objective robust H _∞ control can A good adaptive cruise tracking effect can be achieved, but it is assumed that there are no nonlinear factors when establishing the longitudinal dynamic model. Document 2 (Vajedi M, Azad NL. Ecological Adaptive Cruise Controller for Plug-In Hybrid Electric Vehicles Using Nonlinear Model Predictive Control [J]. IEEE Transactions on Intelligent Transportation Systems, 2015, 17(1): 113-122.) using a nonlinear model Predictive control can make adaptive cruise control vehicles safe and energy-saving, but predictive control requires high computing performance of the system. Therefore, when studying adaptive cruise control, it is necessary to establish a more accurate model and select an appropriate control method.

Contents of the invention

The purpose of the present invention is to solve the above-mentioned existing technical problems, and provide an intelligent electric vehicle adaptive cruise control system and its method that can ensure the ability of intelligent electric vehicle adaptive cruise to track the vehicle ahead, and realize vehicle safety, comfort, energy-saving and intelligent driving.

The intelligent electric vehicle adaptive cruise control system includes an information acquisition module, a working mode selection module, a control function switching module, an expected torque calculation module, a converter module and an actuator module; the information acquisition module includes a vehicle-vehicle communication module, Vehicle-road communication module and vehicle-mounted information sensor, the input end of described working mode selection module is connected with the output end of vehicle-vehicle communication module, vehicle-road communication module and vehicle-mounted information sensor, the output terminal of working mode selection module is connected with the control function The input terminal of the switching module is connected, and the switching module of control action includes a driving control module and a braking control module; the expected torque calculation module includes a driving torque calculation module and a braking torque calculation module, and the output terminal of the driving control module is connected to the driving torque calculation module, and the braking control module The output terminal of the module is connected to the braking torque calculation module; the output terminal of the expected torque calculation module is connected to the input terminal of the converter module, and the output terminal of the converter module is connected to the input terminal of the actuator module.

The intelligent electric vehicle adaptive cruise control method comprises the following steps:

1) After the driver sets the cruise speed and activates the adaptive cruise switch, the cruise starts;

2) The information acquisition module of the adaptive cruise system collects the driving state information of the vehicle and the surrounding environment information in real time. The vehicle speed sensor measures the driving speed of the vehicle, the radar measures the distance between the vehicle and the vehicle in front, and the vehicle-vehicle/vehicle-road communication system Obtain the driving speed of the vehicle in front, etc.;

3) According to the information obtained in real time in step 2), the working mode selection module of the control system makes a choice between the cruise control mode or the tracking mode. If the radar detects that there is no vehicle in front, it will enter the cruise control mode, otherwise it will automatically track the vehicle in front mode; if the driver intervenes during the cruise, the cruise will end;

The desired vehicle distance control strategy in the tracking mode adopts a fixed time distance control strategy, and the desired distance varies linearly with the speed, and its expression is s _d =τv _x +d ₀ , where s _d is the desired vehicle distance, τ is the time distance between vehicles, v _x is the speed of the vehicle, d ₀ is the set minimum safe distance between vehicles;

4) If the vehicle enters the mode of automatically tracking the vehicle in front, the control function switching module makes a selection of the driving or braking mode according to the driving/braking switching strategy;

Let s be the actual distance between two vehicles, then the distance error is expressed as Δs=ss _d , and the relative speed of the two vehicles (that is, the difference between the speed of the preceding vehicle and the speed of the own vehicle) is represented by v _r , according to the relationship between the distance error and the relative speed Propose a driving/braking switching strategy to ensure the comfort of the vehicle when driving;

5) The expected torque calculation module calculates the expected control torque for tracking the vehicle in front, and the expected control torque includes driving torque or braking torque;

6) The converter module converts the desired control torque signal into a driving pedal signal or a brake pedal signal, and outputs it to the actuator module to control the corresponding actuator to complete the control of adaptive cruise;

The control target in tracking mode is Δs→0, v _r →0;

For the control target, the electric vehicle tracking mode control method includes the following steps:

(1) Establish a longitudinal dynamics model of intelligent electric vehicle adaptive cruise;

(2) Using the method of combining inversion design and sliding mode control, an inversion sliding mode control law is proposed to track the expected wheel control torque of the preceding vehicle;

(3) Using fuzzy logic to realize the self-adjustment of important parameters c ₁ , c ₂ , c ₃ in the inverse sliding mode control law, the specific method is as follows:

(3.1) Select the difference Δs between the actual distance between the two vehicles and the expected safe distance and the relative speed v _r of the two vehicles as the input variables of the fuzzy model, and the control law parameters c ₁ , c ₂ , c ₃ as the output variables;

(3.2) Set the discourse domain of the fuzzy model input variable Δs to [-50, 50], the discourse domain of the input variable v _r to [-20, 20], and the quantization factors are all set to 1; the output variables c ₁ , c ₂ , The universe of c ₃ is set to [0,50], and the scale factor is set to 1;

(3.3) Divide the fuzzy subsets of the difference Δs between the actual distance between the two vehicles and the expected safety distance and the relative speed v _r of the two vehicles into 7 language variable levels, and the fuzzy sets are all set to {NB, NM, NS, ZO , PS, PM, PB}, respectively represent {negative large, negative medium, negative small, zero, positive small, positive medium, positive large}; set the controller parameters c ₁ , c ₂ , c ₃ as five language variable levels , the fuzzy sets are all {ZO, PS, PM, PB, VB}, respectively representing {zero, positive small, positive, positive large, very large}; choose Gaussian function as the membership function;

(3.4) The fuzzy model control rule is: when the difference Δs between the actual distance and the expected safety distance of the two vehicles and the relative speed v _r of the two vehicles are both PB (positive), the motor needs to output a large driving torque, and the output variable c is taken at this time ₁ , c ₂ , c ₃ are ZO; when the difference Δs between the actual distance and the expected safety distance of the two vehicles and the relative speed v _r of the two vehicles are both PB, the motor needs to output a large braking torque, and the output variable c is also taken at this time ₁ , c ₂ , c ₃ are ZO; when the difference Δs between the actual distance and the expected safety distance of the two vehicles and the relative speed v _r of the two vehicles are both ZO, the control converges at this time, and the output variables c ₁ , c ₂ , c ₃ are taken is VB; according to the same logic, the control strategy is summarized into 49 control rules;

(4) The adjustment of the membership function in the fuzzy system is a process that needs a lot of experience and experimental data. The self-learning function of the neural network can be used to integrate the neural network with the fuzzy system, so that the membership function in the fuzzy system can be adjusted. The selection is continuously improved, so a five-layer neural network fuzzy system is constructed, and the multi-layer forward neural network is used to complete the function of each step of the fuzzy model. Each layer is defined as follows:

The first layer is the input layer, which transmits the input value to the next layer;

Each node of the second layer represents a linguistic variable value, such as PB, ZO, etc., and is used to calculate the membership function that each input component belongs to the fuzzy set of each linguistic variable value;

Each node in the third layer represents a fuzzy rule, which is used to match the antecedent of the fuzzy rule and calculate the applicability of each rule;

The fourth layer completes the aftermath of the rules, performs fuzzy reasoning and outputs fuzzy quantities;

The fifth layer is the output layer, which realizes clear calculation and outputs control amount.

The technical effects of the present invention are as follows: a safety distance control strategy and a driving/braking switching strategy are proposed, and a neurofuzzy-based inversion sliding mode adaptive cruise tracking mode control method can solve the nonlinearity of the speed control of the adaptive cruise system of an electric vehicle The strong coupling between the problem and the system state ensures the ability of the vehicle to track the vehicle ahead during adaptive cruise driving, and improves the utilization rate of traffic roads and the safety and comfort of vehicle driving.

Description of drawings

FIG. 1 is a schematic diagram of the structural composition of an embodiment of an adaptive cruise control system for an intelligent electric vehicle according to the present invention.

Fig. 2 is a flow chart of an embodiment of an adaptive cruise control method for an intelligent electric vehicle according to the present invention.

Figure 3 is a diagram of the driving/braking switching strategy.

Figure 4 is a schematic diagram of the structure of the neural network fuzzy system.

Detailed ways

The present invention will be further described in detail below in conjunction with the accompanying drawings.

As shown in Figure 1, the intelligent electric vehicle adaptive cruise control system of the present invention includes an information acquisition module 1, a working mode selection module 2, a control action switching module 3, a desired torque calculation module 4, a converter module 5 and an actuator module 6. The information acquisition module 1 includes a vehicle-vehicle communication module 11, a vehicle-road communication module 12, and a vehicle-mounted information sensor 13, and the input terminal of the working mode selection module 2 communicates with the vehicle-vehicle communication module 11 and the vehicle-road communication module. The module 12 is connected to the output end of the vehicle information sensor 13, the output end of the working mode selection module 2 is connected to the input end of the control action switching module 3, and the control action switching module 3 includes a drive control module 31 and a braking control module 32; expected torque calculation Module 4 includes a drive torque calculation module 41 and a braking torque calculation module 42, the output terminal of the drive control module 31 is connected to the drive torque calculation module 41, and the output terminal of the braking control module 32 is connected to the braking torque calculation module 42; the expected torque calculation module 4 The output end is connected to the input end of the converter module 5 , and the output end of the converter module 5 is connected to the input end of the actuator module 6 .

As shown in the flow chart of Figure 2, the embodiment of the intelligent electric vehicle adaptive cruise control method of the present invention includes the following steps:

Step 1: After the driver sets the cruise speed and activates the adaptive cruise switch, the cruise starts;

Step 2: The information acquisition module of the adaptive cruise system collects the driving state information of the vehicle and the surrounding environment information in real time. The vehicle speed sensor measures the driving speed of the vehicle, the radar measures the distance between the vehicle and the vehicle in front, and the vehicle-vehicle/vehicle-road communication The system obtains the speed of the vehicle in front, etc.;

Step 3: Based on the information obtained in real time, the control system working mode selection module selects the cruise control mode or the tracking mode. If the radar detects that there is no vehicle in front, it enters the cruise control mode, otherwise it enters the automatic tracking vehicle ahead mode; If the driver intervenes during the cruise, the cruise will end;

According to the measured vehicle information, the expected safety distance of the tracking mode is calculated. The expected vehicle distance adopts the fixed time distance control strategy (CTG), and its expression is s _d =τv _x +d ₀ , where s _d is the expected vehicle distance , τ is the time distance between vehicles, v _x is the speed of the vehicle, and d ₀ is the set minimum safe vehicle distance.

Step 4: If the vehicle enters the mode of automatically tracking the vehicle in front, the system control role switching module selects the driving or braking mode according to the driving/braking switching strategy.

Let s be the actual distance between two vehicles, then the distance error is expressed as Δs=ss _d , and the relative speed of the two vehicles (ie the difference between the speed of the preceding vehicle and the speed of the own vehicle) is represented by v _r . The invention proposes a driving/braking switching strategy according to the relationship between the vehicle distance error and the relative speed, so as to ensure the comfort of the vehicle during driving.

As shown in Figure 3, taking the vehicle distance error Δs and the relative speed _vr of the two vehicles as the coordinate axes, set the maximum distance that the radar can detect as L, and take ±L as the upper and lower limits of Δs. The whole area is divided into three control areas, which are driving area, braking area, and transition area. In the driving area, the vehicle runs in driving mode; in the braking area, the vehicle runs in braking mode; the slope of the dotted line in the figure is -45 degrees, and the area formed by moving δ up and down on the dotted line (δ is a value greater than zero and very small) is In the transition zone, the vehicle distance error Δs and the relative speed v _r of the two vehicles are in this zone, when the vehicle neither drives nor brakes, it is called coasting mode. This transition zone is set up to reduce the frequent switching of driving and braking and improve the driving comfort of the vehicle.

Step 5: When the vehicle enters the driving mode or braking mode, the expected torque calculation module calculates the corresponding expected control torque with Δs, v _r infinitely approaching zero as the control target. The algorithm includes the following steps:

Step 5.1: Build the car-following mode model of the adaptive cruise system

Considering that the longitudinal slip rate is very small under adaptive cruise driving conditions, assuming that the tire longitudinal force is proportional to the slip rate, a longitudinal model is established, and the state equations in driving mode and braking mode are respectively:

(brake mode)

(drive mode)

Among them, the state variable x=[x ₁ ,x ₂ ,x ₃ ] ^T , x ₁ =s represents the actual distance between the vehicle in front and the vehicle in front, x ₂ =v _x represents the speed of the vehicle, x ₃ =ω _w represents the wheel speed; v _l is the speed of the front vehicle; k is the cornering stiffness of the tire; m is the vehicle mass; r _eff is the effective radius of the tire; f is the rolling resistance coefficient; g is the acceleration of gravity; C _d is the air resistance coefficient; A vehicle frontal area J wheel Moment of inertia; u is the input of the system model, which is taken as the torque T _wheel transmitted to the wheel. When T _wheel is positive, it represents the driving torque, and when it is negative, it represents the braking torque; ΔE(t) is the uncertainty and external disturbance, t is time.

Step 5.2: The speed control of the longitudinal dynamic model of the above-mentioned adaptive cruise system has nonlinear problems and there is a strong coupling between the system state variables. Therefore, the present invention uses an inverse sliding mode controller, and uses the braking mode as an example to illustrate the braking Calculation of the torque sliding mode control law:

Step 5.2.1: Take the expected distance signal s _d , set z ₁ =ss _d , then

Step 5.2.2: Define the Lyapunov function: but Pick where c ₁ is a normal number greater than 0, z ₂ is the virtual control quantity, namely

but

Step 5.2.3: Define the Lyapunov function: because

but:

Pick:

Among them, c ₂ is a normal number greater than 0,

but:

Step 5.2.4: Take the sliding mode switching function as s=z ₃ , and define the Lyapunov function: because

but:

Step 5.2.5: To make In braking mode, the braking torque sliding mode control law is:

Wherein, c ₃ is a normal number greater than 0; η is a sliding mode switching gain;

In the driving mode, the calculation method of the sliding mode control law of the driving torque is the same as the above steps, and the sliding mode control law of the driving torque is:

Step 5.3: Adjusting the values of parameters c ₁ , c ₂ , and c ₃ in the control law can change the dynamic characteristics of the system. The present invention uses fuzzy logic to realize the self-adjustment of parameters c ₁ , c ₂ , and c ₃ of the inversion sliding mode controller .

Step 5.3.1: Select the difference Δs between the actual distance of the two vehicles and the expected safety distance and the relative speed v _r of the two vehicles as the input variables of the fuzzy model, and the controller parameters c ₁ , c ₂ , c ₃ as the output variables.

Step 5.3.2: Set the discourse domain of the fuzzy model input variable Δs to [-50, 50], the discourse domain of the input variable v _r to [-20, 20], and the quantization factors are all set to 1; the output variables c ₁ , c ₂ and c ₃ are both set to [0,50], and the scale factor is set to 1.

Step 5.3.3: Divide the fuzzy subsets of the difference Δs between the actual distance and the expected safety distance between the two vehicles and the relative speed v _r of the two vehicles into 7 language variable levels, and the fuzzy sets are all set to {NB, NM, NS , ZO, PS, PM, PB} respectively represent {negative big, negative middle, negative small, zero, positive small, positive middle, positive big}; set controller parameters c ₁ , c ₂ , c ₃ to 5 languages respectively Variable grades and fuzzy sets are {ZO, PS, PM, PB, VB}, respectively representing {zero, positive small, medium, positive large, very large}; the Gaussian function is selected as the membership function.

Step 5.3.4: The fuzzy rule control table is shown in Table 1. The determination logic of the fuzzy model control rule is: when the difference Δs between the actual distance between the two vehicles and the expected safety distance and the relative speed v _r of the two vehicles are both PB (Zhengda) , the motor needs to output a large driving torque. At this time, the output variables c ₁ , c ₂ , and c ₃ are taken as ZO; when the difference Δs between the actual distance between the two vehicles and the expected safety distance and the relative speed v _r of the two vehicles are both PB, the expected The torque is a large braking torque. At this time, the output variables c ₁ , c ₂ , and c ₃ are also taken as ZO; when the difference Δs between the actual distance between the two vehicles and the expected safety distance and the relative speed v _r of the two vehicles are both ZO, the When the control converges, the output variables c ₁ , c ₂ , and c ₃ are taken as VB; according to the same logic, the control strategy is summarized into 49 control rules.

Table 1

Step 5.4: Integrate the neural network with the fuzzy system, so that the selection of the membership function in the fuzzy system can be continuously improved. Therefore, the present invention constructs a five-layer neural network fuzzy system, as shown in Fig. 4, a schematic structural diagram of the neural network fuzzy system, and adopts a multi-layer forward neural network to complete each step function of the fuzzy model.

Step 5.4.1: The node function of each layer is given below.

The first layer: the input layer, which transmits the input value to the next layer. The number of nodes N ₁ =2,

The second layer: the total number of nodes N ₂ =14, each node represents a language variable value, such as PB, ZO, etc., used to calculate the membership function that each input component belongs to the fuzzy set of each language variable value Among them, c _ij and σ _ij represent the center and width of the membership function respectively, i=1, 2 is the dimension of the input quantity, and j=1, 2...7 is the rank number of the language variable.

The third layer: each node represents a fuzzy rule, which is used to match the antecedent of the fuzzy rule and calculate the applicability of each rule. The number of nodes N ₃ =49 represents 49 fuzzy rules, where i ₁ ∈ {1, 2, ..., 7}, i ₂ ∈ {1, 2, ..., 7}, j=1, 2, ..., 49;

The fourth layer: complete the aftermath of the rules, perform fuzzy reasoning and output fuzzy quantities. The number of nodes N ₄ =N ₃ =49, which realizes normalized calculation, namely

The fifth layer: the output layer, which realizes clear calculation and outputs control amount. The number of nodes N ₅ =3, representing 3 output variables, Where ω _ij is the connection weight of the network.

Step 5.4.2: In the above-mentioned network node function, the main parameters that need to be learned are the connection weight ω _ij of the last layer and the center c _ij and width σ _ij of the membership function. The present invention adopts the method of BP network error backpropagation method to design and adjust the parameters of the learning algorithm.

Define the objective function as where r _i and y _i represent the desired input and actual output, respectively.

The learning algorithm for parameter adjustment is:

Among them, β>0 is the learning rate.

Step 6: The converter module converts the driving or braking torque calculated by the expected torque calculation module into a corresponding pedal signal, and outputs it to the actuator module to complete the adaptive cruise control.

Claims (2)

1. a kind of intelligent electric automobile adaptive cruise control system, it is characterised in that including data obtaining module, operating mode Selecting module, it is expected torque computing module, conversion module and executor module at control action handover module；Described information obtains Modulus block includes Che-vehicle communication module, Che-road communication module and on-vehicle information sensor, the operating mode selecting module Input terminal is connect with the output end of Che-vehicle communication module, Che-road communication module and on-vehicle information sensor, operating mode selection The output end of module is connect with control action handover module input terminal, and control action handover module includes drive control module and system Dynamic control module；It is expected that torque computing module includes driving moment computing module and braking moment computing module, drive control mould Block output termination driving moment computing module, brake control module output termination braking moment computing module；It is expected that Calculating Torque during Rotary The output end of module switches through parallel operation module input, the input terminal of conversion module output termination executor module.

2. intelligent electric automobile self-adapting cruise control method, it is characterised in that include the following steps：

1) after driver sets cruising speed and adaptive cruise is activated to switch, cruise starts；

2) self-adaption cruise system data obtaining module acquires this vehicle traveling movement state information and ambient condition information in real time, main This vehicle travel speed, this vehicle of radar surveying and leading vehicle distance are measured by vehicle speed sensor, Che-vehicle/Che-road communication system obtains Front truck travel speed；

3) information obtained in real time according to step 2), control system operating mode selecting module is to cruise pattern or tracking mould Formula makes a choice, if detections of radar does not have vehicle to front, enters cruise pattern；Otherwise enter from motion tracking front truck mould Formula；If driver intervenes during cruise, cruise terminates；

The tracing mode it is expected when spacing control strategy uses fixed away from control strategy, it is expected that spacing changes with speed linearity, Its expression formula is s_d=τ v_x+d₀, wherein s_dIt is expected spacing, τ is headway, v_xFor this vehicle speed, d₀For the minimum peace of setting Full spacing；

4) if vehicle enters from motion tracking front truck pattern, control action handover module is according to driving/braking switchover policy, to driving Dynamic or braking mode makes a choice；

If s is two vehicle actual vehicle spacing, then spacing error is expressed as △ s=s-s_d, two vehicle relative velocity v_rIt indicates, according to Spacing error and the relationship of relative velocity propose driving/braking switchover policy；

5) it is expected that torque computing module calculates the desired control torque of tracking front truck, the desired control torque includes driving force Square or braking moment；

6) desired control torque signals are converted into driving pedal signal or brake pedal signal by conversion module, are output to and are held Row device module controls respective actuators, completes the control of adaptive cruise；

Control targe in the tracking mode is △ s → 0, v_r→0；

For the control targe, electric vehicle tracing mode control method includes the following steps：

(1) intelligent electric automobile adaptive cruise Longitudinal Dynamic Model is established；

(2) method for using back-stepping design to be combined with sliding formwork control proposes that tracking front truck it is expected the inverting of wheel control moment Sliding formwork control ratio；

(3) fuzzy logic is used to realize important parameter c in back-stepping sliding mode control rule₁, c₂, c₃Self-adjusting, the specific method is as follows：

(3.1) the difference △ s and two vehicle relative velocity v of two vehicle actual ranges and desired safe distance are selected_rAs the defeated of fuzzy model Enter variable, control law parameter c₁, c₂, c₃As output variable；

(3.2) domain of fuzzy model input variable △ s is set as [- 50,50], input variable v_rDomain be [- 20,20], amount Change the factor and all takes 1；Output variable c₁, c₂, c₃Domain be set to [0,50], scale factor all takes 1；

(3.3) by the difference △ s and two vehicle relative velocity v of the two vehicles actual range and desired safe distance_rFuzzy subset all divide For 7 linguistic variable grades, fuzzy set is set as { NB, NM, NS, ZO, PS, PM, PB }, indicate respectively negative big, it is negative small in bearing, Zero, just small, center is honest }；By controller parameter c₁, c₂, c₃Be set to 5 linguistic variable grades, fuzzy set be all ZO, PS, PM, PB, VB }, it indicates respectively { zero, just small, center is honest, very greatly }；Select Gaussian function as membership function；

(3.4) fuzzy model control rule is：As the difference △ s and two vehicle relative velocities of two vehicle actual ranges and desired safe distance v_rWhen being all PB, needs motor to export big driving moment, take output variable c at this time₁, c₂, c₃For ZO；When two vehicle actual ranges And the difference △ s of desired safe distance and two vehicle relative velocity v_rWhen being all PB, motor is needed to export big braking moment, at this time Take output variable c₁, c₂, c₃For ZO；As the difference △ s and two vehicle relative velocity v of two vehicle actual ranges and desired safe distance_rAll it is When ZO, control convergence, takes output variable c at this time₁, c₂, c₃For VB；According to identity logic, control strategy is summarized and is controlled for 49 Rule；

(4) neural network is by the adjustment of membership function with fuzzy using the self-learning function of neural network in fuzzy system System fusion, constructs five layers of Neural Fuzzy system, and each step work(of fuzzy model is completed using multilayer feedforward neural network Can, each layer is defined as follows：

First layer is input layer, and input value is sent to next layer；

One linguistic variable value of each node on behalf of the second layer, act as calculating each input component and belongs to each linguistic variable value mould Paste the membership function of set；The linguistic variable value includes PB, ZO；

One fuzzy rule of each node on behalf of third layer, the former piece for matching fuzzy rule calculate every rule Relevance grade；

The 4th layer of consequent for completing rule carries out fuzzy reasoning and exports fuzzy quantity；

Layer 5 is output layer, realizes that sharpening calculates, exports controlled quentity controlled variable.