CN102745194B - Self-adaption alarming method for preventing tailgating with front car on expressway - Google Patents

Self-adaption alarming method for preventing tailgating with front car on expressway Download PDF

Info

Publication number
CN102745194B
CN102745194B CN201210203507.8A CN201210203507A CN102745194B CN 102745194 B CN102745194 B CN 102745194B CN 201210203507 A CN201210203507 A CN 201210203507A CN 102745194 B CN102745194 B CN 102745194B
Authority
CN
China
Prior art keywords
time
car
driver
chaufeur
reaction
Prior art date
Application number
CN201210203507.8A
Other languages
Chinese (zh)
Other versions
CN102745194A (en
Inventor
李旭
宋翔
张为公
Original Assignee
东南大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 东南大学 filed Critical 东南大学
Priority to CN201210203507.8A priority Critical patent/CN102745194B/en
Publication of CN102745194A publication Critical patent/CN102745194A/en
Application granted granted Critical
Publication of CN102745194B publication Critical patent/CN102745194B/en

Links

Abstract

The invention discloses a self-adaption alarming method for preventing tailgating with a front car on an expressway. The method comprises the following steps of: adapting to driver characteristic and driving environment based on a safety time logic algorithm; and performing self-adaption adjustment on a safety time threshold value, namely an alarming threshold to effectively avoid false-alarm. According to the algorithm, an alarm inhibition strategy is set by determining the intension of a driver to reduce false-alarm rate according to the selection of an initial mode and vehicle driving relevant information acquired by a vehicle-mounted sensor so as to calculate collision time (TTC) of the own car and the front car; the current alarm safety threshold value (TW) is calculated in real time according to the driver characteristic and the current driving environment; and finally, whether the potential danger of tailgating with the front car exists is determined according to the calculated collision time (TTC) of the own car and the front car and the alarm safety threshold value (TW); and an alarm instruction is sent when the potential danger exists. By utilizing the method, the potential tailgating danger can be accurately found out in real time according to the driving age of the driver, the driving environment, the vehicle brake type and the like so as to alarm the driver; and therefore, the occurrence of the tailgating accidents of the own car and the front car on the expressway can be effectively prevented.

Description

A kind of highway automobile is prevented knocking into the back the adaptive alarm method of front truck

Technical field

The present invention relates to the research that a kind of highway automobile travels the adaptive alarm method of the anti-front truck that knocks into the back, particularly relate to the foundation that a kind of highway automobile that can react chaufeur and running environment characteristic based on safety time logical algorithm travels rear-end impact Early-warning Model, its object is in vehicle traveling process, can find that potential rear-end impact is dangerous timely, and give chaufeur with warning, thus avoid knocking into the back front truck from car, belong to field of traffic safety.

Background technology

The statistical information of traffic accident shows, the rear-end impact accident of vehicle on expressway is topmost Crash characteristics.The research of Daimler-benz (Daimler-Benz) company shows, if chaufeur can occur to obtain early warning in first 0.5 second in collision, just can avoid the collision case of 50%; If obtained early warning before 1 second, then can avoid the collision case of 90%.Therefore, develop Warning Device of Preventing Rear End Collision for Motor Vehicle, before rear-end impact occurs, give driver alert, remind chaufeur take braking or the measure such as to change, prevent from knocking into the back front truck from car, for raising ride safety of automobile, reduce the generation of rear-end collision, have great importance.Setting up the alarm method of the rear-end impact of science and model is then the core studied of rear-end impact warning device and gordian technique place.Only have and set up comparatively accurate, reasonable and comprehensive math modeling in theory, warning device could conform to the practical operation situation of vehicle, just has practicality.

At present, preventing collision alarm method both domestic and external and model mainly comprise safety distance logical algorithm and safety time logical algorithm two class.Under safety distance logical algorithm refers to the conditions present by calculating running car, the minor increment between the vehicle of rear-end impact needs maintenance and obstacle avoided by vehicle, and compares with safety distance threshold value, determines safe condition.Safety time logical algorithm is that collision time by calculating two workshops and safety time threshold value compare, and determines safe condition.More for the research of Safety distance model both at home and abroad, also form fixing spacing model, the model that fixing headway model and kinematics model etc. are multi-form.But in two the important parameter relative distances affecting rear-end impact and relative velocity, what Safety distance model was paid close attention to is this parameter of relative distance, and have ignored this key parameter of relative velocity, thus have impact on the accuracy of Early-warning Model to a great extent.Meanwhile, Safety distance model also also exists very large deficiency on the comformability for chaufeur individuality.

Based on such consideration, what the present invention adopted is rear-end impact alarm method based on safety time logical algorithm.But existing safety time logical algorithm alarm method and model also exist larger problem and shortage.First, safety time logical algorithm is that collision time by calculating two workshops and safety time threshold value compare to determine precarious position, and in existing research at home and abroad, no matter be what takes report to the police, choose several safety time threshold value, its safety time threshold value all adopts empirical value and constant, the difference of different driving environment and different chaufeur individuality cannot be embodied, flase alarm under some driving cycles without the need to warning, and do not report to the police under other driving cycles that need report to the police, cause false alarm rate and miss alert rate higher.Secondly, existing safety time logical algorithm cannot discover the driving intention of chaufeur, even if discover to some extent collision risk at chaufeur, and when having taked braking or the behavior such as to have changed, still report to the police, this creates impact to the normal running of chaufeur to a certain extent, easily causes the dislike of chaufeur, serves the effect of running counter to desire.Moreover, existing safety time logical algorithm is when the collision time of calculating two car, the method of calculating taked is divided by two car relative velocities by relative distance, two car relative velocities are less trend towards zero time, the collision time value calculated is comparatively large, can not produce warning, if but now two car relative spacings are less, then can give chaufeur a kind of unsafe sensation, be in comparatively dangerous situation equally, chaufeur should be reminded to be noted.Finally, present safety time logical algorithm cannot ensure that chaufeur is after acceptance is reported to the police, and has sufficient time proper operation vehicle, prevents the generation of rear-end collision.

Summary of the invention

For the problems referred to above with not enough, the object of this invention is to provide a kind of highway automobile based on safety time logical algorithm to prevent knocking into the back the adaptive alarm method of front truck, monitor this car with the relative motion state of front truck in track, effectively prevention to be knocked into the back front truck from car.It is individual that this algorithm can adapt to different running environment and different chaufeur, real-time adjustment safety time threshold value, and formulate in conjunction with driver intention shielding strategy of reporting to the police, and effectively reduces false alarm rate and alert rate by mistake.When two workshop relative velocities are zero, keep certain vehiclar gap to reduce the sense of insecurity of chaufeur, ensure that chaufeur is after acceptance is reported to the police simultaneously, have sufficient time proper operation vehicle to prevent the generation of knocking into the back.

For achieving the above object, the present invention takes following technical scheme, and a kind of highway automobile is prevented knocking into the back the adaptive alarm method of front truck, it is characterized in that, comprises the following steps:

1) Current vehicle type of retardation, road surface types and chaufeur driving age is selected

Before driving, select Current vehicle type of retardation, road surface types and chaufeur driving age, wherein, car brakeing type comprises vapour-pressure type and fluid pressure type, road surface types comprises dry pavement, wet road surface, snowy road surface and ice-patch surface, the chaufeur driving age is the chaufeur actual driving age, and option is divided into and is less than 1 year, 1 ~ 3 year and be greater than 3 years three classes;

2) front truck is gathered relative to from the relative distance of car, relative velocity and the speed from car, accelerator pedal signal, steering indicating light signal, brake pedal signal and current visbility signal

Utilize millimeter wave radar sensor or vision sensor gather front truck with from car relative distance and relative velocity; Gather from vehicle speed sensor from vehicle speed Information Pull; Das Gaspedal force snesor and brake pedal force sensor is utilized to obtain Das Gaspedal and brake pedal signal respectively, and by information that whether automobile CAN-bus Real-time Obtaining steering indicating light is opened; Visibility sensor is utilized to obtain current visibility information;

3) according to the information gathered in step 2, setting shielding is reported to the police tactful

According to gather in step 2 from the speed V of car relative to front truck r, unit: m/s, if V r≤ 0, then collisionless is dangerous, and shielding is reported to the police, otherwise, according to the brake pedal, Das Gaspedal and the steering indicating light signal that gather in step 2, judge that driver is intended to, following shielding strategy of reporting to the police is set: when chaufeur steps on braking, shielding of reporting to the police; During chaufeur accelerator releasing, shielding of reporting to the police; When chaufeur opens steering indicating light, shielding of reporting to the police;

4) according to the information gathered in step 2 and step 1, when from car relative to front vehicle speed V rduring > 0, calculate the collision time TTC from car and front truck and alarm security threshold T w

The described collision time TTC from car and front truck is calculated as follows:

TTC = L V r

In formula, L be front truck relative to the distance from car, unit: m, V rfor from the speed of car relative to front truck, unit: m/s, both obtain by step 2;

Described alarm security threshold T wbe calculated as follows:

T W=T 1+T 2+T 3+T 4

In formula, T 1for time of driver's reaction, namely accept from chaufeur the time that early warning takes brake measure to experience to it, unit: s;

T 2for the brake coordination time, namely from the time that chaufeur takes brake measure to experience to brake system generation braking force, unit: s;

T 3for deceleration/decel build up time, namely braking deceleration rises to the time of maximum braking deceleration from 0, unit: s;

T 4for the continuous braking time, to Ben Che and front truck geo-stationary required time when namely braking deceleration reaches maxim, unit: s;

Described time of driver's reaction T 1span statistical value be 0.3 ~ 1s, it is calculated as follows:

The reaction time of chaufeur is determined by the method setting up a time of driver's reaction identification BP artificial neural net (ANN), artificial neural net (ANN) comprises input layer, hidden layer and output layer, wherein input layer receives the input vector calculated required for time of driver's reaction, this input vector is 4 dimensional vectors, and each dimension is respectively from the car speed of a motor vehicle, visbility, road surface types and chaufeur driving age;

Hidden layer comprises 6 neuron nodes, and each neuron node can calculate according to the input vector of input layer, and the output valve of calculating calculates the input value of time of driver's reaction as output layer; Neuron node is tansig function according to the function that input vector carries out calculating: tansig (x)=1/ (1+e -x), concrete, the expression formula of this tansig function as shown in the formula:

m j = tan sig [ Σ i = 1 4 ( w i × p i ) + b j ]

In above formula, i=1,2 ... 4, be input vector dimension, p ibe i-th input vector; J=1,2 ... 6, be the number of hidden layer neuron, m jfor a hidden layer jth neuronic output valve, w ifor p corresponding in hidden layer inetwork weight coefficient, b jfor a hidden layer jth threshold coefficient;

Output layer comprises an output layer neuron node, and the result that this neuron node calculates according to hidden layer calculates time of driver's reaction, and output layer neuron node function is purelin function: purelin (x)=kx; Concrete, the function that output layer neuron node calculates time of driver's reaction is shown in following formula:

m = purelin [ Σ j = 1 6 ( w j × m j ) + b ]

In formula, m jfor a hidden layer jth neuronic output valve, w jfor m corresponding in output layer jnetwork weight coefficient,

B is the threshold coefficient of output layer, and m is the neuronic output valve of output layer, i.e. output vector;

The foundation of whole BP artificial neural net (ANN) have employed matlab function newff, sees following formula:

net=newff(min max(p),[6,1],{′tansig′,′purlin′},′trainlm′)

In above formula, net is neural network, and p is input vector, and minmax is that matlab obtains Vector Extremal function, and trainlm is learning algorithm;

Input vector p according to above-mentioned artificial neural net (ANN) gathers, and comprises and measures difference from the car speed of a motor vehicle, different visbility, different road surface types, different chaufeur driving age and time of driver's reaction; Utilize above-mentioned collection value off-line training time of driver's reaction identification BP artificial neural net (ANN), obtain the network weight coefficient w of hidden layer iwith threshold coefficient b j, and the network weight coefficient w of output layer jwith threshold coefficient b;

The input vector of time of driver's reaction identification BP artificial neural net (ANN) is defined as follows: be actual measured value from car speed of a motor vehicle input value; Visbility input value is also actual measured value; Road surface types is divided into dry pavement, wet road surface, snowy road surface and ice-patch surface; Dry pavement input value is 1, and wet road surface input value is 2, and snowy road surface input value is 3, and ice-patch surface input value is 4; The chaufeur driving age is divided into and is less than 1 year, 1 ~ 3 year and be greater than 3 years; To be less than 1 year input value be 1,1 ~ 3 year input value is 2, and being greater than 3 years input values is 3;

The output vector of BP artificial neural net (ANN) is defined as follows: output vector value scope is 0 ~ 3, represent the statistical value 0.3 ~ 1.0s of time of driver's reaction, 0 to represent the chaufeur real reaction time short, get 0.3s, 1 to represent time of driver's reaction shorter, gets 0.5s, 2 represent chaufeur reflects that reality is longer, get 0.8s, 3 to represent time of driver's reaction long, gets 1s;

The main purpose of BP artificial neural net (ANN) training obtains network weight coefficient w i, w jwith threshold coefficient b j, b, make output vector meet certain precision; The training of BP artificial neural net (ANN) directly adopts matlab function train to carry out off-line training, sees following formula:

net=train(net,p,t)

In above formula: net is BP artificial neural net (ANN), p is training sample input vector, and t is training sample output vector;

The network weight coefficient w that above-mentioned training is obtained i, w jwith threshold coefficient b j, b substitutes in neural network, according to reality input from the car speed of a motor vehicle, visbility, road surface types and driver person's driving age, calculate output vector;

According to the output vector calculated, and according to the output vector defined, judge time of driver's reaction.BP neural network 1 output vector close to 0, then judges that time of driver's reaction is 0.3s; Output vector close to 1, then judges that time of driver's reaction is 0.5s; Output vector close to 2, then judges that time of driver's reaction is 0.8s; Output vector close to 3, then judges that time of driver's reaction is 1s;

Described brake coordination time T 2, according to related data, getting empirical value is 0.2s;

Described deceleration/decel build up time T 3according to the type of retardation in step 1 selected by chaufeur, hydraulic brake system and Pneumatic braking system get 0.2s, 0.7s respectively;

Described continuous braking time T 4be calculated as follows:

T 4 = V r μ × g

In formula, V rfor from the speed of car relative to front truck, unit: m/s, obtains by step 2; G is acceleration due to gravity, and its value is 9.8m/s 2; μ is coefficient of road adhesion, determines: the adhesion value of dry pavement gets 0.61, and the adhesion value of wet road surface gets 0.46, and the adhesion value of snowy road surface gets 0.20, and the adhesion value of ice-patch surface gets 0.07 according to road surface types in step 1;

5) according to the collision time TTC from car and front truck calculated and alarm security threshold T w, differentiate whether there is potential rear-end impact danger, thus determine whether to need to report to the police

Described identification tactic is as follows, as TTC > T wduring+τ, be determined as safety, do not send alarm command; As TTC≤T wduring+τ, be determined as danger, send alarm command and remind chaufeur; Type of alarm can take the different type of alarms such as voice prompt alarm, does not limit at this;

Wherein, TTC is the collision time from car and front truck, unit: s, T wfor alarm security threshold value unit: s, τ are the time of giving warning in advance [s]; The described time that gives warning in advance τ is calculated as follows:

τ=d/V

In formula, d is safe stopping distance [m], and getting empirical value is 3m; V is this vehicle speed [m/s], obtains by step 2.

The present invention is owing to adopting above technical scheme, and it has the following advantages:

(1) the present invention adopts the preventing collision algorithm based on safety time logic, take into full account two key factors affecting rear-end impact: relative distance and relative velocity, can calculate Ben Che and front truck relative motion state and judge precarious position in real time, thus when there is the potential front truck that knocks into the back dangerous, generate alarm command, actv. prevents and to knock into the back front truck from car.

(2) the present invention adopts the method for BP neural network, determines the reaction time of chaufeur, adapts to different chaufeurs and different driving cycles, effectively prevent the reaction time to obtain long or too short, avoids alert and false-alarm by mistake.

(3) the present invention is according to chaufeur driving age and concrete running environment, car brakeing type etc., real-time computationally secure time gate limit value, achieve adaptation that is individual to different chaufeur and different running environment, solve the problem that orthodox method adopts constant threshold comformability deficiency.

(4) the present invention is according to judging that the driving intention setting shielding of chaufeur is reported to the police tactful, is judging that chaufeur has discovered danger and under the operating mode of taking measures, shielding is reported to the police, and actv. avoids false-alarm.

(5) setting of safe stopping distance, actv. avoid relative velocity less when two car relative distances less dangerous working conditions.Meanwhile, the setting of the time that gives warning in advance, also ensures that chaufeur is after accepting early warning, has sufficient time proper operation vehicle.

Accompanying drawing explanation

Fig. 1 is the workflow diagram of alarm method method of the present invention.

Fig. 2 is the structured flowchart of alarm method of the present invention.

Fig. 3 is time of driver's reaction identification BP artificial neural network structure.

Detailed description of the invention

Highway rear-end collision prevention front truck early warning and alarming method is the core technology of highway rear-end collision prevention warning device, only have and set up comparatively accurate, reasonable and comprehensive alarm method and math modeling in theory, the anti-warning device that knocks into the back could conform to the practical operation situation of vehicle, just there is practicality, really could play and prevent from knocking into the back front truck from car, improve ride safety of automobile, reduce traffic accident, improve the effect of safety traffic.

Below in conjunction with embodiment, the present invention is described in detail.

As shown in Figure 1, the present invention includes following steps:

1) chaufeur manually selects Current vehicle type of retardation, road surface types and chaufeur driving age.

As shown in Figure 2, the originate mode module 21 preset, before driving, manually selects Current vehicle type of retardation, road surface types and chaufeur driving age by chaufeur.Car brakeing type comprises vapour-pressure type and fluid pressure type, road surface types comprises dry pavement, wet road surface, snowy road surface and ice-patch surface, is differentiated by chaufeur, and the chaufeur driving age is the chaufeur actual driving age, option is divided into and is less than 1 year, 1 ~ 3 year and be greater than 3 years three classes; All by chaufeur drive before manually select corresponding pattern.And selected pattern information is sent to the early warning discrimination module 23 shown in Fig. 2.

2) front truck is gathered relative to from the relative distance of car, relative velocity and from the speed of car, accelerator pedal signal, steering indicating light signal, brake pedal signal and visibility information.

As shown in Figure 2, on-vehicle information acquisition module 22, utilizes millimeter wave radar sensor or vision sensor to gather front truck and from car relative distance and relative velocity, does not limit at this.The ESR type millimeter wave radar that radar sensor can adopt Delphi (delphi) company to produce, the VS078FC type digital camera that vision sensor can adopt dimension to produce depending on (Microvision) company, gather from vehicle speed sensor from vehicle speed Information Pull, JK08/NS43NCT1 photoelectricity car speed sensor can be adopted.Das Gaspedal force snesor and brake pedal force sensor is utilized to obtain Das Gaspedal and brake pedal signal respectively, and by information that whether CAN Real-time Obtaining steering indicating light is opened.Visibility information can be obtained by FYG11 type visibility sensor.Gathered on-vehicle information is sent to the early warning discrimination module 23 shown in Fig. 2.

3) according to the information gathered in step 2, setting shielding is reported to the police tactful

In the early warning discrimination module 23 shown in Fig. 2, according to gather in step 2 from the speed V of car relative to front truck r[m/s], if V r≤ 0, then collisionless is dangerous, and shielding is reported to the police, if V r> 0, then start to start alarming logic (step 4,5).

Meanwhile, according to the brake pedal, Das Gaspedal and the steering indicating light signal that gather in step 2, judge that driver is intended to, following shielding strategy of reporting to the police is set: when chaufeur steps on braking, reaction driver energy is concentrated, and is just taking brake measure to avoid knocking into the back, shielding of reporting to the police; During chaufeur accelerator releasing, reaction driver energy is concentrated, and is just taking deceleration measure to avoid knocking into the back, shielding of reporting to the police; When chaufeur opens steering indicating light, reaction driver energy is concentrated, and just carries out conscious changing, shielding of reporting to the police.

4) according to the information gathered in step 2 and step 1, when from car relative to front vehicle speed V rduring > 0, calculate the collision time TTC from car and front truck and alarm security threshold T w.

The calculating of the described collision time TTC from car and front truck is such as formula (1):

TTC = L V r - - - ( 1 )

In formula (1), L be front truck relative to the distance [m] from car, V rfor from the speed [m/s] of car relative to front truck.

Both obtain by step 2.

Described alarm security threshold T wcalculating such as formula (2):

T W=T 1+T 2+T 3+T 4(2)

In formula (2), T 1for time of driver's reaction, namely accept from chaufeur the time [s] that early warning takes brake measure to experience to chaufeur.

T 2for the brake coordination time, namely from the time [s] that chaufeur takes brake measure to experience to brake system generation braking force.T 3for deceleration/decel build up time, namely braking deceleration rises to the time [s] of maximum braking deceleration from 0.T 4for the continuous braking time, to Ben Che and front truck geo-stationary required time [s] when namely braking deceleration reaches maxim.

1. described time of driver's reaction T 1be calculated as follows:

The skill level, degree of fatigue etc. of time of driver's reaction and chaufeur are closely bound up, and by the impact of the factors such as the speed of a motor vehicle, weather, visbility, therefore be a uncertain value, a large amount of experimental datas shows, the statistical value of reaction opening time of chaufeur is 0.3 ~ 1.0s.According to this reaction time range, the present invention determines the reaction time of chaufeur by setting up a time of driver's reaction identification BP artificial neural net (ANN) 1.

Consider that the factor affecting time of driver's reaction is from the car speed of a motor vehicle, visbility, road surface types and chaufeur proficiency, and chaufeur proficiency is the factor of a bad tolerance, it is generally acknowledged, driving age longer chaufeur often has better proficiency and reaction time faster, therefore, the present invention chooses and differentiates time of driver's reaction from the car speed of a motor vehicle, visbility, road surface types and chaufeur driving age.

As shown in Figure 3, artificial neural net (ANN) 1 comprises input layer 11, hidden layer 12 and output layer 13.

Wherein input layer 11 receives the input vector calculated required for time of driver's reaction, and this input vector is 4 dimensional vectors, and each dimension is respectively from car speed of a motor vehicle U1, visbility U2, road surface types U3 and chaufeur driving age U4.

Hidden layer 12 comprises 6 neuron nodes 121, and this neuron node 121 can calculate according to the input vector of input layer 11, and the output valve of calculating calculates the input value of time of driver's reaction as output layer 13; Neuron node 121 is tansig function according to the function that input vector carries out calculating:

Tansig (x)=1/ (1+e -x), concrete, the expression formula of this tansig function is shown in formula (3)

m j = tan sig [ Σ i = 1 4 ( w i × p i ) + b j ] - - - ( 3 )

In formula (3), i=1,2 ... 4, be input vector dimension, p ibe i-th input vector; J=1,2 ... 6, be the number of hidden layer neuron 121, m jfor the output valve of a hidden layer 12 jth neuron 121, w ifor p corresponding in hidden layer 12 inetwork weight coefficient, b jfor a hidden layer 12 jth threshold coefficient.

Output layer 13 comprises an output layer neuron node 131, and the result that this neuron node 131 calculates according to hidden layer 12 calculates time of driver's reaction, and output layer neuron node 131 function is purelin function: purelin (x)=kx.Concrete, the function that output layer neuron node 131 calculates time of driver's reaction is shown in formula (4)

m = purelin [ Σ j = 1 6 ( w j × m j ) + b ] - - - ( 4 )

In formula (4), m jfor the output valve of a hidden layer 12 jth neuron 121, w jfor m corresponding in output layer 13 jnetwork weight coefficient, b is the threshold coefficient of output layer 13, and m is the output valve of output layer neuron 131, i.e. output vector.

The foundation of whole BP artificial neural net (ANN) 1 have employed matlab function newff, sees formula (5)

net=newff(min max(p),[6,1],{′tansig′,′purlin′},′trainlm′) (5)

In formula (5), net is neural network, and p is input vector, and min max is that matlab obtains Vector Extremal function, and trainlm is learning algorithm.

Input vector p according to above-mentioned artificial neural net (ANN) 1 measures, and comprises and measures difference from the car speed of a motor vehicle, different visbility, different road surface types, different chaufeur driving age and time of driver's reaction.

Wherein, road surface types and chaufeur driving age are manually inputted before driving by chaufeur, JK08/NS43NCT1 photoelectricity speed of a motor vehicle sensor measurement is utilized from the car speed of a motor vehicle, visbility utilizes FYG11 type visibility sensor to obtain, time of driver's reaction is measured by counting machine 74LS190, also obtains by expert's questionnaire.

Utilize above-mentioned observed reading off-line training time of driver's reaction identification BP artificial neural net (ANN) 1, obtain the network weight coefficient w of hidden layer 12 iwith threshold coefficient b j, and the network weight coefficient w of output layer 13 jwith threshold coefficient b.

The input vector of time of driver's reaction identification BP artificial neural net (ANN) 1 is defined as follows: be actual measured value from car speed of a motor vehicle input value; Visbility input value is also actual measured value; Road surface types is divided into dry pavement, wet road surface, snowy road surface and ice-patch surface.Dry pavement input value is 1, and wet road surface input value is 2, and snowy road surface input value is 3, and ice-patch surface input value is 4; The chaufeur driving age is divided into and is less than 1 year, 1 ~ 3 year and be greater than 3 years.To be less than 1 year input value be 1,1 ~ 3 year input value is that 2,2 to be greater than 3 years input values be 3.

The output vector of BP artificial neural net (ANN) 1 is defined as follows: output vector value scope is 0 ~ 3, represent the statistical value 0.3 ~ 1.0s of time of driver's reaction, 0 to represent the chaufeur real reaction time short, get 0.3s, 1 to represent time of driver's reaction shorter, gets 0.5s, 2 represent chaufeur reflects that reality is longer, get 0.8s, 3 to represent time of driver's reaction long, gets 1s.

The main purpose that BP artificial neural net (ANN) 1 is trained obtains network weight coefficient w i, w jwith threshold coefficient b j, b, make output vector meet certain precision.The training of BP artificial neural net (ANN) directly adopts matlab function tranin to carry out off-line training, sees formula (6)

net=train(net,p,t) (6)

In formula (6): net is BP artificial neural net (ANN) 1, p is training sample input vector, and t is training sample output vector.

The network weight coefficient w that above-mentioned training is obtained i, w jwith threshold coefficient b j, b substitutes in neural network, according to reality input from the car speed of a motor vehicle, visbility, road surface types and driver person's driving age, and convolution (3) and (4), calculating output vector.

According to the output vector calculated, and according to the output vector defined, judge time of driver's reaction.BP neural network 1 output vector close to 0, then judges that time of driver's reaction is 0.3s; Output vector close to 1, then judges that time of driver's reaction is 0.5s; Output vector close to 2, then judges that time of driver's reaction is 0.8s; Output vector close to 3, then judges that time of driver's reaction is 1s.

2. described brake coordination time T 2, according to data, getting statistical value is 0.2s.

3. described deceleration/decel build up time T 3depend on type of retardation, hydraulic brake system is approximately 0.15 ~ 0.3s, and pneumatic brake system is approximately 0.3 ~ 0.9s.In the present invention, according to the type of retardation in step 1 selected by chaufeur, hydraulic brake system and Pneumatic braking system value 0.2s, 0.7s respectively.

4. described continuous braking time T 4be calculated as follows:

T 4 = V r μ × g - - - ( 7 )

In formula (7), V rfor from the speed [m/s] of car relative to front truck, obtain by step 2.G is acceleration due to gravity, and its value is 9.8m/s 2.μ is coefficient of road adhesion, determines: the adhesion value of dry pavement gets 0.61, and the adhesion value of wet road surface gets 0.46, and the adhesion value of snowy road surface gets 0.20, and the adhesion value of ice-patch surface gets 0.07 according to road surface types in step 1.In fact, the value that μ × g calculates is vehicle maximum braking deceleration [m/s under current road adhesion value 2].

According to the above-mentioned T calculated 1, T 3and T 4, then substitute into T 2the empirical value 0.2s got, utilizes formula (2) to calculate alarm security threshold T w.

5) according to the collision time TTC from car and front truck calculated and alarm security threshold T w, differentiate whether there is potential rear-end impact danger, thus determine whether to need to report to the police.

Each kinematic feature factor above-mentioned is analyzed, can draw the following conclusions: when this vehicle speed is less than or equal to front vehicle speed, be i.e. V rwhen≤0, also can not collide even if do not implement braking, therefore this algorithm only works as V r> 0 starts warning; If this vehicle speed is greater than front vehicle speed, i.e. V r> 0, as long as reserved sufficient chaufeur braking time, namely at TTC>=T win time, brakes, then can not collide after braking.

Visible, the limiting conditions implementing braking are TTC=T w.Travelling with car now after braking is the state that a kind of car is close to front truck, in order to not make chaufeur produce the mental burden of having a lingering fear, should keep necessary spacing d, i.e. safe stopping distance d [m] after braking between two cars.Be embodied on time quantum, namely advance time τ carries out early warning.

The computing formula of time τ of giving warning in advance is:

τ=d/V (8)

In formula (8), d is safe stopping distance [m], and getting empirical value is 3m.V is this vehicle speed [m/s], obtains by step 2.

The setting of τ, can ensure that chaufeur is after accepting early warning, has sufficient time proper operation vehicle.Now, the limiting conditions implementing braking are TTC=T w+ τ.

Therefore, described identification tactic is as follows, as TTC > T wduring+τ, be determined as safety, do not send alarm command; As TTC≤T wduring+τ, be determined as danger, send alarm command and remind chaufeur.

In the present embodiment, type of alarm can take the different type of alarms such as voice prompt alarm, does not limit at this.

Claims (1)

1. highway automobile is prevented knocking into the back an adaptive alarm method for front truck, it is characterized in that, comprises the following steps:
1) Current vehicle type of retardation, road surface types and chaufeur driving age is selected
Before driving, select Current vehicle type of retardation, road surface types and chaufeur driving age, wherein, car brakeing type comprises vapour-pressure type and fluid pressure type, road surface types comprises dry pavement, wet road surface, snowy road surface and ice-patch surface, the chaufeur driving age is the chaufeur actual driving age, and option is divided into and is less than 1 year, 1 ~ 3 year and be greater than 3 years three classes;
2) front truck is gathered relative to from the relative distance of car, relative velocity and the speed from car, accelerator pedal signal, steering indicating light signal, brake pedal signal and current visbility signal
Utilize millimeter wave radar sensor or vision sensor gather front truck with from car relative distance and relative velocity; Gather from vehicle speed sensor from vehicle speed Information Pull; Das Gaspedal force snesor and brake pedal force sensor is utilized to obtain Das Gaspedal and brake pedal signal respectively, and by information that whether automobile CAN-bus Real-time Obtaining steering indicating light is opened; Visibility sensor is utilized to obtain current visibility information;
3) according to the information gathered in step 2, setting shielding is reported to the police tactful
According to gather in step 2 from the speed V of car relative to front truck r, unit: m/s, if V r≤ 0, then collisionless is dangerous, and shielding is reported to the police, otherwise, according to the brake pedal, Das Gaspedal and the steering indicating light signal that gather in step 2, judge that driver is intended to, following shielding strategy of reporting to the police is set: when chaufeur steps on braking, shielding of reporting to the police; During chaufeur accelerator releasing, shielding of reporting to the police; When chaufeur opens steering indicating light, shielding of reporting to the police;
4) according to the information gathered in step 2 and step 1, when from car relative to front vehicle speed V rduring > 0, calculate the collision time TTC from car and front truck and alarm security threshold T w
The described collision time TTC from car and front truck is calculated as follows:
TTC = L V r
In formula, L be front truck relative to the distance from car, unit: m, V rfor from the speed of car relative to front truck, unit: m/s, both obtain by step 2;
Described alarm security threshold T wbe calculated as follows:
T W=T 1+T 2+T 3+T 4
In formula, T 1for time of driver's reaction, namely accept from chaufeur the time that early warning takes brake measure to experience to it, unit: s;
T 2for the brake coordination time, namely from the time that chaufeur takes brake measure to experience to brake system generation braking force, unit: s;
T 3for deceleration/decel build up time, namely braking deceleration rises to the time of maximum braking deceleration from 0, unit: s;
T 4for the continuous braking time, to Ben Che and front truck geo-stationary required time when namely braking deceleration reaches maxim, unit: s;
Described time of driver's reaction T 1span statistical value be 0.3 ~ 1s, it is calculated as follows:
The reaction time of chaufeur is determined by the method setting up a time of driver's reaction identification BP artificial neural net (ANN), artificial neural net (ANN) comprises input layer, hidden layer and output layer, wherein input layer receives the input vector calculated required for time of driver's reaction, this input vector is 4 dimensional vectors, and each dimension is respectively from the car speed of a motor vehicle, visbility, road surface types and chaufeur driving age;
Hidden layer comprises 6 neuron nodes, and each neuron node can calculate according to the input vector of input layer, and the output valve of calculating calculates the input value of time of driver's reaction as output layer; Neuron node is tansig function according to the function that input vector carries out calculating: tansig (x)=1/ (1+e -x), concrete, the expression formula of this tansig function as shown in the formula:
m j = tan sig [ Σ i = 1 4 ( w i × p i ) + b j ]
In above formula, i=1,2 ... 4, be input vector dimension, p ibe i-th input vector; J=1,2 ... 6, be the number of hidden layer neuron, m jfor a hidden layer jth neuronic output valve, w ifor p corresponding in hidden layer inetwork weight coefficient, b jfor a hidden layer jth threshold coefficient;
Output layer comprises an output layer neuron node, and the result that this neuron node calculates according to hidden layer calculates time of driver's reaction, and output layer neuron node function is purelin function: purelin (x)=kx; Concrete, the function that output layer neuron node calculates time of driver's reaction is shown in following formula:
m = purelin [ Σ j = 1 6 ( w j × m j ) + b ]
In formula, m jfor a hidden layer jth neuronic output valve, w jfor m corresponding in output layer jnetwork weight coefficient, b is the threshold coefficient of output layer, and m is the neuronic output valve of output layer, i.e. output vector;
The foundation of whole BP artificial neural net (ANN) have employed matlab function newff, sees following formula:
net=newff(min max(p),[6,1],{'tansig','purlin'},'trainlm')
In above formula, net is neural network, and p is input vector, and min max is that matlab obtains Vector Extremal function, and trainlm is learning algorithm;
Input vector p according to above-mentioned artificial neural net (ANN) gathers, and comprises and measures difference from the car speed of a motor vehicle, different visbility, different road surface types, different chaufeur driving age and time of driver's reaction; Utilize above-mentioned collection value off-line training time of driver's reaction identification BP artificial neural net (ANN), obtain the network weight coefficient w of hidden layer iwith threshold coefficient b j, and the network weight coefficient w of output layer jwith threshold coefficient b;
The input vector of time of driver's reaction identification BP artificial neural net (ANN) is defined as follows: be actual measured value from car speed of a motor vehicle input value; Visbility input value is also actual measured value; Road surface types is divided into dry pavement, wet road surface, snowy road surface and ice-patch surface; Dry pavement input value is 1, and wet road surface input value is 2, and snowy road surface input value is 3, and ice-patch surface input value is 4; The chaufeur driving age is divided into and is less than 1 year, 1 ~ 3 year and be greater than 3 years; To be less than 1 year input value be 1,1 ~ 3 year input value is 2, and being greater than 3 years input values is 3;
The output vector of BP artificial neural net (ANN) is defined as follows: output vector value scope is 0 ~ 3, represent the statistical value 0.3 ~ 1.0s of time of driver's reaction, 0 to represent the chaufeur real reaction time short, get 0.3s, 1 to represent time of driver's reaction shorter, gets 0.5s, 2 represent chaufeur reflects that reality is longer, get 0.8s, 3 to represent time of driver's reaction long, gets 1s;
The main purpose of BP artificial neural net (ANN) training obtains network weight coefficient w i, w jwith threshold coefficient b j, b, make output vector meet certain precision; The training of BP artificial neural net (ANN) directly adopts matlab function train to carry out off-line training, sees following formula:
net=train(net,p,t)
In above formula: net is BP artificial neural net (ANN), p is training sample input vector, and t is training sample output vector;
The network weight coefficient w that above-mentioned training is obtained i, w jwith threshold coefficient b j, b substitutes in neural network, according to reality input from the car speed of a motor vehicle, visbility, road surface types and driver person's driving age, calculate output vector;
According to the output vector calculated, and according to the output vector defined, judge time of driver's reaction, BP neural network 1 output vector close to 0, then judges that time of driver's reaction is 0.3s; Output vector close to 1, then judges that time of driver's reaction is 0.5s; Output vector close to 2, then judges that time of driver's reaction is 0.8s; Output vector close to 3, then judges that time of driver's reaction is 1s;
Described brake coordination time T 2, according to related data, getting empirical value is 0.2s;
Described deceleration/decel build up time T 3according to the type of retardation in step 1 selected by chaufeur, hydraulic brake system and Pneumatic braking system get 0.2s, 0.7s respectively;
Described continuous braking time T 4be calculated as follows:
T 4 = V r μ × g
In formula, V rfor from the speed of car relative to front truck, unit: m/s, obtains by step 2; G is acceleration due to gravity, and its value is 9.8m/s 2; μ is coefficient of road adhesion, determines: the adhesion value of dry pavement gets 0.61, and the adhesion value of wet road surface gets 0.46, and the adhesion value of snowy road surface gets 0.20, and the adhesion value of ice-patch surface gets 0.07 according to road surface types in step 1;
5) according to the collision time TTC from car and front truck calculated and alarm security threshold T w, differentiate whether there is potential rear-end impact danger, thus determine whether to need to report to the police
Described identification tactic is as follows, as TTC > T wduring+τ, be determined as safety, do not send alarm command; As TTC≤T wduring+τ, be determined as danger, send alarm command and remind chaufeur; Type of alarm takes voice prompt alarm;
Wherein, TTC is the collision time from car and front truck, unit: s, T wfor alarm security threshold value unit: s, τ are the time of giving warning in advance [s]; The described time that gives warning in advance τ is calculated as follows:
τ=d/V
In formula, d is safe stopping distance [m], and getting empirical value is 3m; V is this vehicle speed [m/s], obtains by step 2.
CN201210203507.8A 2012-06-19 2012-06-19 Self-adaption alarming method for preventing tailgating with front car on expressway CN102745194B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201210203507.8A CN102745194B (en) 2012-06-19 2012-06-19 Self-adaption alarming method for preventing tailgating with front car on expressway

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201210203507.8A CN102745194B (en) 2012-06-19 2012-06-19 Self-adaption alarming method for preventing tailgating with front car on expressway

Publications (2)

Publication Number Publication Date
CN102745194A CN102745194A (en) 2012-10-24
CN102745194B true CN102745194B (en) 2015-02-25

Family

ID=47025807

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201210203507.8A CN102745194B (en) 2012-06-19 2012-06-19 Self-adaption alarming method for preventing tailgating with front car on expressway

Country Status (1)

Country Link
CN (1) CN102745194B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107128251A (en) * 2017-05-11 2017-09-05 张家港工领信息科技有限公司 A kind of collision prevention of vehicle control method

Families Citing this family (48)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103106811A (en) * 2013-01-15 2013-05-15 东南大学 Recognition method of motor vehicle valid traffic conflict based on crashing time of two vehicles
HRP20130542A2 (en) * 2013-06-12 2013-11-22 Vojislav Iliev Audio-visual warning system for road traffic participants
CN103350670B (en) * 2013-07-16 2015-08-12 厦门金龙联合汽车工业有限公司 A kind of vehicle forward direction collision-proof alarm method based on car networking technology
KR101449326B1 (en) * 2013-09-03 2014-10-08 현대자동차주식회사 Apparatus and method for calculating concentration grade of driver, vehicle collision warning system using the same
CN103434511B (en) * 2013-09-17 2016-03-30 东南大学 The combined estimation method of a kind of speed of a motor vehicle and road-adhesion coefficient
CN104691545B (en) * 2013-12-04 2017-10-24 财团法人车辆研究测试中心 Adaptability vehicle collision avoidance method
CN103761889B (en) * 2014-02-17 2015-10-28 东南大学 A kind of front vehicles effective target defining method of preventing collision
JP5969534B2 (en) * 2014-04-21 2016-08-17 株式会社デンソー Driving support device
CN103971540B (en) * 2014-05-16 2016-08-17 长安大学 A kind of prior-warning device preventing vehicle pileup from knocking into the back and method for early warning
CN104280736A (en) * 2014-09-24 2015-01-14 天津三星电子有限公司 Mobile terminal and prompting method
CN105574552A (en) * 2014-10-09 2016-05-11 东北大学 Vehicle ranging and collision early warning method based on monocular vision
CN104346955A (en) * 2014-10-16 2015-02-11 浙江吉利汽车研究院有限公司 Man-vehicle communication-based pedestrian collision avoiding method and collision avoiding system
CN105788360B (en) * 2014-12-16 2018-10-02 中国移动通信集团公司 A kind of methods, devices and systems of vehicle collision avoidance
CN104999983A (en) * 2015-07-27 2015-10-28 苏州宏展信息科技有限公司 Anti-collision method of automobile
CN105253145B (en) * 2015-08-17 2018-01-12 深圳市美好幸福生活安全系统有限公司 Self-tuning parameter adjustment method in the advanced drive assist system of automobile
DE102015117976A1 (en) * 2015-10-22 2017-04-27 Knorr-Bremse Systeme für Nutzfahrzeuge GmbH Method and device for controlling a warning module
CN105291965B (en) * 2015-11-04 2017-12-05 石海水 A kind of vehicle anticollision method and device
CN105235681B (en) * 2015-11-11 2018-06-08 吉林大学 A kind of vehicle rear-end collision collision avoidance system and method based on pavement conditions
CN105774807A (en) * 2016-03-04 2016-07-20 苏黎 Vehicle speed control method and vehicle
CN106696966A (en) * 2016-11-18 2017-05-24 杭州好好开车科技有限公司 Alarming method for ADAS product in progress
CN106781691A (en) * 2016-11-30 2017-05-31 北京汽车集团有限公司 Drive pre-warning system and method
JP2018101402A (en) * 2016-12-21 2018-06-28 トヨタ自動車株式会社 Vehicle data recording device
CN108216098A (en) * 2016-12-22 2018-06-29 乐视汽车(北京)有限公司 Vehicle early warning threshold value update method, system and its electronic equipment
CN106846547A (en) * 2017-01-18 2017-06-13 成都锐奕信息技术有限公司 A kind of drive recorder of detecting distance and intelligent alarm
CN107170292B (en) * 2017-06-30 2019-08-20 维沃移动通信有限公司 A kind of driving safety prompt method and electronic equipment
CN107791943A (en) * 2017-10-26 2018-03-13 南京越博电驱动系统有限公司 The preceding collision warning systems and method for early warning of a kind of pure electric automobile
CN107697071B (en) * 2017-11-06 2019-07-02 东南大学 A kind of driving safety grade based on field theory determines method and device
CN109801511B (en) * 2017-11-16 2021-01-05 华为技术有限公司 Collision early warning method and device
CN107963076A (en) * 2017-12-11 2018-04-27 开沃新能源汽车集团有限公司 Based on millimetre-wave radar detection target obstacle longitudinal direction collision-proof method
CN108569268B (en) * 2017-12-15 2020-08-04 上海蔚来汽车有限公司 Vehicle anti-collision parameter calibration method and device, vehicle controller and storage medium
CN108725440B (en) * 2018-04-20 2020-11-27 深圳市商汤科技有限公司 Forward collision control method and apparatus, electronic device, program, and medium
CN108639065B (en) * 2018-05-15 2019-09-27 辽宁工业大学 A kind of vehicle safe driving control method of view-based access control model
CN111081061A (en) * 2018-10-22 2020-04-28 杭州海康威视数字技术股份有限公司 Collision early warning method and device
CN111114540A (en) * 2018-10-30 2020-05-08 长城汽车股份有限公司 Vehicle and safe driving method and device thereof
CN109300323A (en) * 2018-11-21 2019-02-01 荆门博谦信息科技有限公司 A kind of speed bootstrap technique and system based on car networking
CN109552289B (en) * 2018-11-29 2020-06-02 辽宁工业大学 Automobile self-adaptive auxiliary braking system and control method thereof
CN109572692B (en) * 2018-12-29 2020-11-06 辽宁工业大学 Control method of electric control vehicle anti-collision system
CN109741630A (en) * 2019-01-18 2019-05-10 爱驰汽车有限公司 Away from threshold value of warning adjusting method, system, equipment and storage medium when automobile
CN109910865A (en) * 2019-02-26 2019-06-21 辽宁工业大学 A kind of vehicle early warning brake method based on Internet of Things
CN110103954B (en) * 2019-03-01 2020-08-18 辽宁工业大学 Electric control-based automobile rear-end collision prevention early warning device and method
CN110155046B (en) * 2019-05-09 2020-12-29 武汉理工大学 Automatic emergency braking hierarchical control method and system
CN110329249B (en) * 2019-07-02 2020-08-07 武汉理工大学 Automobile forward collision early warning control system and method of cyclic neural network
CN110316198A (en) * 2019-07-29 2019-10-11 竹旭 A kind of safe-guard system and operation method for highway speed-raising
CN110588641A (en) * 2019-09-06 2019-12-20 爱驰汽车有限公司 Forward target crossing warning method and system
CN110648550A (en) * 2019-10-17 2020-01-03 深圳中集智能科技有限公司 Driving safety system
CN111145589A (en) * 2019-12-17 2020-05-12 北京交通大学 Vehicle omnidirectional anti-collision early warning system based on vector algorithm
CN111086513A (en) * 2019-12-24 2020-05-01 泰牛汽车技术(苏州)有限公司 Control method for automatic driving
CN111238825A (en) * 2020-01-10 2020-06-05 东南大学 Intelligent driving automatic emergency braking performance testing method for combined test pavement

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101151171A (en) * 2005-09-26 2008-03-26 丰田自动车株式会社 Vehicle surroundings information output system and method for outputting vehicle surroundings information
CN101468630A (en) * 2007-12-25 2009-07-01 三菱扶桑卡客车株式会社 Device for reducing collision damage

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1741079B1 (en) * 2004-04-08 2008-05-21 Mobileye Technologies Limited Collision warning system
DE102005011241A1 (en) * 2005-03-11 2006-09-14 Robert Bosch Gmbh Method and apparatus for collision warning
JP4907207B2 (en) * 2006-03-29 2012-03-28 本田技研工業株式会社 Vehicle travel safety device
JP4297132B2 (en) * 2006-05-12 2009-07-15 トヨタ自動車株式会社 Vehicle alarm device

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101151171A (en) * 2005-09-26 2008-03-26 丰田自动车株式会社 Vehicle surroundings information output system and method for outputting vehicle surroundings information
CN101468630A (en) * 2007-12-25 2009-07-01 三菱扶桑卡客车株式会社 Device for reducing collision damage

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107128251A (en) * 2017-05-11 2017-09-05 张家港工领信息科技有限公司 A kind of collision prevention of vehicle control method

Also Published As

Publication number Publication date
CN102745194A (en) 2012-10-24

Similar Documents

Publication Publication Date Title
Feng et al. Can vehicle longitudinal jerk be used to identify aggressive drivers? An examination using naturalistic driving data
CN105679020B (en) A kind of driving behavior analysis apparatus and method
CN104627180B (en) A kind of half cruise active control system and its method
CN103496366B (en) A kind of initiative lane change collision avoidance control method based on collaborative truck and device
CN104129377B (en) Automobile active anticollision adaptive fuzzy control method
CN102874188B (en) Driving behavior warning method based on vehicle bus data
CN105966396B (en) A kind of vehicle intelligent collision control method based on driver's collision avoidance behavior
CN105015548B (en) A kind of longitudinal collision avoidance is reminded and follows integrated system and method automatically
CN104340226B (en) System for sending alarm before motor vehicle and object may collide
CN105722738B (en) The drive dynamic control device of moving body
CN104114423B (en) Method for determining an emergency braking situation of a vehicle
CN102616235B (en) A kind of collaborative collision avoidance device based on truck traffic and collision avoidance method
CN103350670B (en) A kind of vehicle forward direction collision-proof alarm method based on car networking technology
CN105374215A (en) Highway roadside early warning device
CN107161097B (en) Vehicle running intelligent security system based on Beidou Navigation System
CN103029665B (en) Control system and control method for actively avoiding side collision of automobile
CN102774378B (en) Vehicle rear-end collision early warning and guard method, system
CN202703577U (en) Intelligent car brake control system
CN107554420A (en) A kind of safe distance between vehicles warning system based on road environment
CN102069800B (en) Vehicle running control method and control device
CN103366416B (en) A kind of vehicle data collection and dispensing device
CN105501220B (en) The method, apparatus and vehicle of vehicle collision prewarning
CN102897155B (en) Collision avoidance system and automatic braking method of electric automobile
CN105313769B (en) A kind of vehicle-mounted active forewarning method, system and device based on millimetre-wave radar
CN105118316A (en) Curved road safe speed calculating method and caution system based on vehicle infrastructure cooperation

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant