CN106056140A - System and method for establishing working condition division remote damage assessment of different vehicle types based on artificial intelligence supervised learning linear regression method - Google Patents

Abstract

The invention relates to a system and method for establishing working condition division remote damage assessment of different vehicle types based on an artificial intelligence supervised learning linear regression method and belongs to the vehicle damage assessment field. The objective of the invention is to solve problems in working condition detection after a vehicle collision. According to the technical schemes of the invention, a working condition detection subsystem is adopted to judge all working condition information in the vehicle collision; and the working condition detection subsystem learns working condition training data so as to generate a working condition model, wherein the working condition model is built by adopting the supervised learning linear regression method. With the system and method provided by the technical schemes of the invention adopted, working condition detection in the vehicle collision can be realized; and a machine learning method is used in the remote damage assessment technical field, so that the accuracy of judgment in a damage assessment process can be improved.

Description

Different automobile types divided working status is set up based on artificial intelligence's supervised learning linear regression method Remotely loss assessment system and method

Technical field

The invention belongs to car damage identification field, relate to a kind of based on the foundation of artificial intelligence's supervised learning linear regression method The long-range loss assessment system of different automobile types divided working status and method.

Background technology

During low-speed motion (including low speed links traveling, vehicle parking etc.), take place frequently collision accident for vehicle and lead The Claims Resolution dispute problem caused, long-range setting loss technology by multi-signal in collection vehicle driving process (as speed, acceleration, Angular velocity, sound etc.) and analyzed with signal processing and machine learning techniques, to judge whether collision occurs and collision rift The damage situation of vehicle.

After vehicle crashes, headend equipment is capable of detecting when the generation of collision and intercepts the signal of collision process, Being sent to high in the clouds by wireless network, remote server extracts the eigenvalue of design in advance from the signal received, and uses machine Learning algorithm is analyzed, and first judges the accuracy of crash data, then judges collision object and operating mode situation, to determine Collision Number Create the damage of which kind of grade according to what part of set pair, then go out with reference to amount for which loss settled concurrent according to part injury rating calculation Deliver to insurance company.The detection for vehicle, operating mode, target, part and region can be related to during this.

Summary of the invention

After solving vehicle collision, for the problem of operating mode detection, the present invention proposes has supervision based on artificial intelligence Study linear regression method sets up the long-range loss assessment system of different automobile types divided working status and method, to realize the operating mode inspection during setting loss Survey.

In order to solve above-mentioned technical problem, the technical scheme that the present invention provides is characterized by: including:

Truck type choice subsystem, selects the model data corresponding to vehicle as total data set；

Data classification subsystem, reads CAE emulation data and real vehicle data, and classifies data accordingly；

Collision detection subsystem, it is judged that the most whether vehicle collides；Described collision detection subsystem pair Collision training data carries out learning thus generates collision model, and described collision model is set up and used supervised learning linear regression side Method；

Operating mode detection subsystem, it is judged that all work informations that collision occurs；Operating mode is instructed by described operating mode detection subsystem White silk data carry out learning thus generate condition model, and described condition model is set up and used supervised learning linear regression method.

Beneficial effect: technique scheme, it is possible to achieve the operating mode for vehicle collision detects, at this of long-range setting loss Technical field employs the method for machine learning, for machine learning method, during setting loss, it determines accuracy rate on To promote；The present invention is by selecting vehicle to import the data corresponding to this vehicle, and data classification is then for model training With the purpose of test and the step that adds；The detection of operating mode is the purpose that the program realizes, and is to obtain through sequence of operations The result arrived.

Accompanying drawing explanation

Fig. 1 is the structural schematic block diagram of system of the present invention.

Detailed description of the invention

In order to the present invention being made apparent explanation, below the technical term that the present invention relates to is made definitions:

Operating mode: all collision informations such as collision angle, direction, target, region；

Vehicle: automobile model；

Target: collision target；

Region: position of collision；

Part: auto parts；

Operating mode detects: detect all collision informations such as this car collision angle, direction, target, region；

Vehicle detects: the automobile model that detection collides with this car；

Target detection: detect this car collision target；

Region detection: detect this car position of collision；

Piece test: detect this car auto parts.

Embodiment 1:

One sets up the long-range loss assessment system of different automobile types divided working status based on artificial intelligence's supervised learning linear regression method, Including:

Truck type choice subsystem, selects the model data corresponding to vehicle as total data set；

Data classification subsystem, reads CAE emulation data and real vehicle data, and classifies data accordingly；

Collision detection subsystem, it is judged that the most whether vehicle collides；Described collision detection subsystem pair Collision training data carries out learning thus generates collision model, and described collision model is set up and used supervised learning linear regression side Method；

Operating mode detection subsystem, it is judged that all work informations that collision occurs；Operating mode is instructed by described operating mode detection subsystem White silk data carry out learning thus generate condition model, and described condition model is set up and used supervised learning linear regression method.

Described collision detection subsystem includes, collision training module, crash tests module, collision authentication module, described in touch Hitting training module and generate collision model for learning collision training data, crash tests module is for surveying collision Examination data bring the result detecting collision model in collision model into, and collision authentication module uses true sport car data verification collision mould The reliability of type and accuracy rate；

Described operating mode detection subsystem includes, operating mode training module, working condition measurement module, operating mode authentication module, described work Condition training module generates condition model for learning operating mode training data, and described working condition measurement module is for by work Condition test data bring the result detecting condition model in model into, and operating mode authentication module uses true sport car data verification operating mode mould The reliability of type and accuracy rate.

Described supervised learning linear regression method includes:

S1. input variable x is characterized, and the predictive value y of output is desired value；The curve table of matching is shown as y=h (x)；

S2. output y is the linear function of x, and is expressed as matrix form；

S3. introduce cost function, use gradient descent algorithm, after initializing learning parameter, repeat renewal learning parameter Value, to obtain minimum side's more new regulation.

Described gradient descent algorithm is: batch gradient declines and/or statistical gradient declines.

Further, said method particularly as follows:

Output y is that the linear function of x is:

h_θ(x)=θ₀+θ₁x₁+θ₂x₂

Here, θ i is parameter, and n is Grad, it is assumed that x₀=1, above formula is expressed as matrix form:

θ and x is column vector, and m is Grad, certain training set given, introduces cost function, its definition

As follows:

By initial guess initiation parameter θ, the most constantly change the value of parameter θ so that parameter J (θ) is the least, directly To finally giving the J (θ) minimized, use gradient descent algorithm, after initiation parameter θ, repeat following renewal equation, with The value of undated parameter θ, J (θ) is cost function, and j is the subscript of parameter, and θ j is parameter (asking for an interview h (x) function above to explain)；

α represents learning rate,

Thus, update equation to be reduced to:

Embodiment 2:

One sets up the long-range damage identification method of different automobile types divided working status based on artificial intelligence's supervised learning linear regression method, Comprise the following steps:

Step one. select the model data corresponding to vehicle as total data set；

Step 2. read CAE emulation data and real vehicle data, and accordingly data are classified；

Step 3. judge the most whether vehicle collides；Described collision detection subsystem is to collision training Data carry out learning thus generate collision model, and described collision model is set up and used supervised learning linear regression method；

Step 4. judge all work informations that collision occurs；Operating mode training data is entered by described operating mode detection subsystem Row learns thus generates condition model, and described condition model is set up and used supervised learning linear regression method.

Comprise the concrete steps that:

Step 3 includes:

S3.1. use collision detection subsystem that CAE collision simulation data are processed, then classify to produce collision to it Training data and crash tests data；

S3.2. in collision training module, collision training data learnt and produces collision model, carrying out simulated crash The effect of training data；

S3.3. crash tests data are used to carry out the result of test collisions model in crash tests module；

S3.4. use true sport car data as collision checking data and to bring collision authentication module into, verify collision mould The accuracy of type；

Step 4 includes:

S4.1. CAE operating mode emulation data are processed by applying working condition detection subsystem, then it is carried out classification generation operating mode instruction Practice data and working condition measurement data；

S4.2. in operating mode training module, operating mode training data learnt and produce condition model, carrying out simulated condition The effect of training data；

S4.3. in working condition measurement module, applying working condition test data carry out the result of measurement condition model；

S4.4. use true sport car data as operating mode checking data and to bring operating mode authentication module into, verify operating mode mould The accuracy of type.

Described supervised learning linear regression method includes:

S1. input variable x is characterized, and the predictive value y of output is desired value；The curve table of matching is shown as y=h (x)；

S2. output y is the linear function of x, and is expressed as matrix form；

S3. introduce cost function, use gradient descent algorithm, after initializing learning parameter, repeat renewal learning parameter Value, to obtain minimum side's more new regulation.

Described gradient descent algorithm is: batch gradient declines and/or statistical gradient declines.

Further, said method particularly as follows:

Output y is that the linear function of x is:

h_θ(x)=θ₀+θ₁x₁+θ₂x₂

Here, θ i is parameter, and n is Grad, it is assumed that x₀=1, above formula is expressed as matrix form:

θ and x is column vector, and m is Grad, certain training set given, introduces cost function, and it is defined as follows:

By initial guess initiation parameter θ, the most constantly change the value of parameter θ so that parameter J (θ) is the least, directly To finally giving the J (θ) minimized, use gradient descent algorithm, after initiation parameter θ, repeat following renewal equation, with The value of undated parameter θ, J (θ) is cost function, and j is the subscript of parameter, and θ j is parameter (asking for an interview h (x) function above to explain)；

α represents learning rate,

Thus, update equation to be reduced to:

Embodiment 3:

Supplementing as embodiment 1 or 2, uses the linear regression method of supervised learning, in statistics, linear regression It is to utilize the least square function being referred to as equation of linear regression that relation between one or more independent variables and dependent variable is built A kind of regression analysis of mould.The target of regression problem is that given D ties up input variable x, and each input vector x has correspondence Value y, it is desirable to for continuous print desired value t of new its correspondence of data prediction.In the present system, x refers to preprocessed The signal data that module extracts, y refers to impairment scale label/operating mode label/collision labels.

Essential Terms in above-mentioned learning process: overall data set is referred to as training set；Input variable x is characterized；Output Predictive value y be desired value；The curve of matching, is typically expressed as y=h (x), referred to as hypothesized model.

In order to utilize supervised learning, it would be desirable to the form of determining function h.As an initial selected, we can be false Surely output y is the linear function of x.That is:

h_θ(x)=θ₀+θ₁x₁+θ₂x₂

Here, for parameter, also referred to as weights (weights).It is assumed that=1.The most above-mentioned can be expressed as rectangular Formula:

In above formula, n represents the number of input variable, and is all column vector.By formulation above, the most just have a talk about, give After certain training set fixed, our work is how to select learning parameter θ in other words.In order to define prediction output with corresponding Difference between true output, next we introduce cost function (costfunction), and it is defined as follows:

Our target is, by the value of Selection parameter θ, the value making above-mentioned cost function as far as possible is minimum.Can Carry out initiation parameter θ first passing through initial guess (initialguess), the most constantly change the value of parameter θ so that value Parameter J (θ) is the least, minimizes until finally giving

J(θ).Particularly, it is considered to a Gradient declines (gradientdescent) algorithm.After it initializes θ, then weigh Perform equation below again and update the value of θ.

Here, α represents learning rate (learningrate).It is noted that the least meeting of α value makes finally to converge Very slow, i.e. gradient declines very slow, and value is too big, can make finally to converge quickly, i.e. gradient declines quickly；And it is different Initial value α, it is possible to cause the value of parameter θ minimum for the J (θ) finally made to differ.Next we need it is of concern that The solving of that partial derivative on the right, for our example above-mentioned, we can obtain:

So renewal equation above can be reduced to:

Above formula is exactly our known lowest mean square (Leastmeansquares, LMS) more new regulation, or also referred to as Widrow-Hoff learning rules.By above formula it may be seen that the amplitude proportional updated is in error term (that in bracket , it represents the difference between actual value and predictive value) value.

The relational expression only for a training set example that we obtain above, during for multiple training set example Under, we also need to make above formula some amendment, have two kinds of methods: (1) batch gradient declines (batchgradientdescent)；(2) statistical gradient declines (stochasticgradientdescent).The most first be given The algorithm steps of the two.

Declining for batch gradient, its algorithm is:

Declining for statistical gradient, its algorithm is:

Under the most at first sight, do not think what difference the two has, and essence is quite different.For batch gradient descent algorithm, It will search for training dataset really in renewal each time；And for statistical gradient descent algorithm, train for each The single-instance of collection, it is all made renewal, i.e. can make a response at once.So generally we can select statistical gradient to decline Algorithm.

By running above-mentioned algorithm, we can obtain finding the value of parameter θ to meet given training set.Obtain this letter After number h, for the example being newly added, its output valve just can be done a prediction by us.A place is had to it should be noted that The value of parameter θ likely can not converge, and simply makes what J (θ) minimized to oscillate about, but it practice, near this value θ Require that through meeting.So this situation should be taken in program realizes when.

Embodiment 4:Have and the identical technical scheme of embodiment 1 or 2 or 3, more specifically:

Conceptual data collection in such scheme: be entirely CAE emulation data and sport car data；Be divided into three parts as follows

1. training dataset: be used to training pattern or determine model parameter (CAE emulation data and sport car data).

2. checking data set: be used to do Model Selection (modelselection), i.e. does the final optimization pass and really of model Fixed (CAE emulation data and sport car data).

3. test data set: the Generalization Ability of the model then trained for purely test.(CAE emulates data With sport car data).

In the present embodiment also to the filtering related to during setting loss, weighting choose, feature extraction, normalization, eigentransformation Have been described.

1. wave filter technology: the filtering method realized includes FIR filtering, FIR Chebyshev approximation, Chebyshev's filter Ripple, butterworth filter etc., the Filtering.m file in mastery routine realizes.Each wave filter is common wave filter, Matlab has corresponding function to realize, and specific algorithm refers to signal processing professional book.Provide FIR filter herein Content and the introduction of flow process.

Limited impulse response digital filter (FIR, FiniteImpulseResponse) is the system of a kind of full zero point, The design of FIR filter is ensureing that amplitude characteristic meets the colleague that technology requires, it is easy to accomplish strict linear phase characteristic, So being the outstanding advantages of FIR filter according to there being stable and linear phase characteristic.Chebyshev approximation is that the ripples such as one are forced Nearly method, it is possible to make error frequency band be uniformly distributed, to same technical specification, this filter order sending out needs shoulder to shoulder is low, right In the wave filter of same exponent number, this approximatioss maximum error is minimum, and the key step of its design is as follows:

Step 1: the setting of filter parameter

The parameter of wave filter includes: cut-off frequecy of passband, stopband cut-off frequency, passband maximum attenuation and stopband minimum decline Subtract；

Step 2: be arranged on passband and the amplitude-frequency response of stopband coideal

Step 3: be scheduled on the weighting on cut-off frequecy of passband and stopband cut-off frequency point

Step 4: utilize Equation for Calculating Chebyshev approximation filter coefficient

Step 5: preserve coefficient

Step 6: extraction coefficient carries out data filtering

Wherein: the guarantee signal that is disposed to of filter parameter does not haves distortion now during processing As, the cut-off frequency of filtered signal and sample frequency need to meet Nyquist's theorem, the most after the filtering signal Highest frequency not can exceed that the 1/2 of original signal sample frequency, otherwise arises that Lou frequency phenomenon.According to the signal in current project The sample frequency of collection plate is mainly 50Hz and 1KHz, according to formula F as a example by 50Hz_Cut-off< 50/2, therefore select filter cutoff Frequency is below 25.

2. Feature Extraction Technology: feature extraction is carried out on collision alarm.Judge that the feature that collision uses includes window Acceleration in difference between acceleration maxima and minima, window in the maximum of acceleration absolute value, window in mouthful In average energy (in window the quadratic sum of acceleration a little divided by counting), window, the absolute value of each point slope is average Value.

Judge the feature that part category is used include the average energy between speed, acceleration peak to peak, Amplitude between maximum and minima/width between the two, acceleration maximum, acceleration minima, maximum place The width of half-wave, minima place half-wave width, maximum and minima between difference, between peak to peak Span, the meansigma methods of absolute value of each point slope, signal carry out each of the signal after Fourier transform in 0～38 frequency ranges The amplitude of individual frequency component.

3. normalization technology: that causes classification task to eliminate the dimension between feature or order of magnitude difference is unfavorable Impact, needs to be normalized characteristic so that have comparability between each eigenvalue, it is to avoid the spy that numerical value is bigger Levy and flood the feature that numerical value is less.Original characteristic is after normalized, and each feature is in identical codomain scope. Owing to the performance of Z-Score is more preferable, use Z-Score as method for normalizing.

4. feature transform technique: in the case of feature is more, for the dependency eliminated between feature and reduce redundancy Feature, needs to convert feature, carrys out reflected sample information with the fewest new feature.In the less situation of experiment sample Under (practical situation of this project) reduce too much intrinsic dimensionality, moreover it is possible to avoid sending out of over-fitting or poor fitting to a certain extent Raw.According to actual needs, the eigentransformation the most realized is PCA.Being found through experiments, PCA divides for improving this project Class performance there is no help, has declined, and this is that the feature owing to being used at present is less, does not has redundancy feature, therefore PCA wouldn't be used, but be as being stepped up of subsequent characteristics, however not excluded that use the probability of PCA later.

In accompanying drawing 1, record: the Truck type choice subsystem that Truck type choice is in the present invention；Data categorization module is Data classification subsystem in the present invention；The collision detection subsystem that collision judgment module is in the present invention；Operating mode detection mould Block is the operating mode detection subsystem of the present invention；Vehicle detection module is the vehicle detection subsystem of the present invention；Piece test Module i.e. piece test subsystem；Module of target detection is the target detection subsystem of the present invention, and region detection module is The region detection subsystem of the present invention.

The above, only the invention preferably detailed description of the invention, but the protection domain of the invention is not Being confined to this, any those familiar with the art is in the technical scope that the invention discloses, according to the present invention The technical scheme created and inventive concept thereof in addition equivalent or change, all should contain the invention protection domain it In.

Claims (8)

1. set up the long-range loss assessment system of different automobile types divided working status based on artificial intelligence's supervised learning linear regression method, its It is characterised by, including:

Truck type choice subsystem, selects the model data corresponding to vehicle as total data set；

Data classification subsystem, reads CAE emulation data and real vehicle data, and classifies data accordingly；

Collision detection subsystem, it is judged that the most whether vehicle collides；Described collision detection subsystem is to collision Training data carries out learning thus generates collision model, and described collision model is set up and used supervised learning linear regression method；

Operating mode detection subsystem, it is judged that all work informations that collision occurs；Operating mode is trained number by described operating mode detection subsystem Generating condition model according to carrying out learning, described condition model is set up and is used supervised learning linear regression method.

2. set up different automobile types divided working status based on artificial intelligence's supervised learning linear regression method as claimed in claim 1 remote Journey loss assessment system, it is characterised in that

Described collision detection subsystem includes, collision training module, crash tests module, collision authentication module, described collision is instructed Practicing module and generate collision model for learning collision training data, crash tests module is for by crash tests number According to bringing the result detecting collision model in collision model into, collision authentication module uses true sport car data verification collision model Reliability and accuracy rate；

Described operating mode detection subsystem includes, operating mode training module, working condition measurement module, operating mode authentication module, and described operating mode is instructed Practicing module and generate condition model for learning operating mode training data, described working condition measurement module is for surveying operating mode Examination data bring the result detecting condition model in model into, and operating mode authentication module uses true sport car data verification condition model Reliability and accuracy rate.

3. such as claim 1 or 2, to set up different automobile types divided working status based on artificial intelligence's supervised learning linear regression method long-range Loss assessment system, it is characterised in that described supervised learning linear regression method, including:

S1. input variable x is characterized, and the predictive value y of output is desired value；The curve table of matching is shown as y=h (x)；

S2. output y is the linear function of x, and is expressed as matrix form；

S3. introduce cost function, use gradient descent algorithm, after initializing learning parameter, repeat the value of renewal learning parameter, To obtain minimum side's more new regulation.

4. set up different automobile types divided working status based on artificial intelligence's supervised learning linear regression method as claimed in claim 3 remote Journey loss assessment system, it is characterised in that described descent algorithm is: batch gradient declines and/or statistical gradient declines, described in have supervision The specific algorithm of study linear regression method is:

Output y is that the linear function of x is:

h_θ(x)=θ₀+θ₁x₁+θ₂x₂

θ i is parameter, and n is Grad, it is assumed that x₀=1, above formula is expressed as matrix form:

$h\left(x\right)=\underset{i=0}{\overset{n}{\&Sigma;}}{\mathrm{\&theta;}}_i{x}_i={\mathrm{\&theta;}}^{T}x,$

θ and x is column vector, and m is Grad, certain training set given, introduces cost function, and it is defined as follows:

$J\left(\mathrm{\&theta;}\right)=\frac{1}{2}\underset{i=1}{\overset{m}{\&Sigma;}}{\left(h_{\mathrm{\&theta;}}\right(x^{\left(i\right)})-y^{\left(i\right)})}^2.$

By initial guess initiation parameter θ, the most constantly change the value of parameter θ so that parameter J (θ) is the least, until Obtain the J (θ) minimized eventually, use gradient descent algorithm, after initiation parameter θ, repeat following renewal equation, to update The value of parameter θ, J (θ) is cost function, and j is the subscript of parameter, and θ j is parameter；

${\mathrm{\&theta;}}_j:={\mathrm{\&theta;}}_j-\mathrm{\&alpha;}\frac{\&part;}{\&part;{\mathrm{\&theta;}}_j}J\left(\mathrm{\&theta;}\right).$

α represents learning rate

$\begin{array}{c}\frac{\&part;}{\&part;{\mathrm{\&theta;}}_j}J\left(\mathrm{\&theta;}\right)=\frac{\&part;}{\&part;{\mathrm{\&theta;}}_j}\frac{1}{2}{\left(h_{\mathrm{\&theta;}}\right(x)-y)}^2\\ =2\&CenterDot;\frac{1}{2}\left(h_{\mathrm{\&theta;}}\right(x)-y)\&CenterDot;\frac{\&part;}{\&part;{\mathrm{\&theta;}}_j}\left(h_{\mathrm{\&theta;}}\right(x)-y)\\ =\left(h_{\mathrm{\&theta;}}\right(x)-y)\&CenterDot;\frac{\&part;}{\&part;{\mathrm{\&theta;}}_j}(\underset{i=0}{\overset{n}{\&Sigma;}}{\mathrm{\&theta;}}_i{x}_i-y)\\ =\left(h_{\mathrm{\&theta;}}\right(x)-y)x_j\end{array}$

Thus, update equation to be reduced to:

${\mathrm{\&theta;}}_j:={\mathrm{\&theta;}}_j+\mathrm{\&alpha;}(y^{\left(i\right)}-h_{\mathrm{\&theta;}}(x^{\left(i\right)}\left)\right)x_j^{\left(i\right)}\mathrm{..}$

5. set up the long-range damage identification method of different automobile types divided working status based on artificial intelligence's supervised learning linear regression method, its It is characterised by, comprises the following steps:

Step one. select the model data corresponding to vehicle as total data set；

Step 2. read CAE emulation data and real vehicle data, and accordingly data are classified；

Step 3. judge the most whether vehicle collides；Collision training data is learnt thus generates and touch Hitting model, described collision model is set up and is used supervised learning linear regression method；

Step 4. judge all work informations that collision occurs；Operating mode training data is learnt thus generates condition model, Described condition model is set up and is used supervised learning linear regression method.

6. set up different automobile types divided working status based on artificial intelligence's supervised learning linear regression method as claimed in claim 5 remote Journey damage identification method, it is characterised in that comprise the concrete steps that:

Step 3 includes:

S3.1. use collision detection subsystem that CAE collision simulation data are processed, then classify to produce collision training to it Data and crash tests data；

S3.2. in collision training module, collision training data learnt and produces collision model, carrying out simulated crash training The effect of data；

S3.3. crash tests data are used to carry out the result of test collisions model in crash tests module；

S3.4. use true sport car data as collision checking data and to bring collision authentication module into, verify collision model Accuracy；

Step 4 includes:

S4.1. CAE operating mode emulation data are processed by applying working condition detection subsystem, then it is carried out classification generation operating mode training number According to working condition measurement data；

S4.2. in operating mode training module, operating mode training data learnt and produce condition model, carrying out simulated condition training The effect of data；

S4.3. in working condition measurement module, applying working condition test data carry out the result of measurement condition model；

S4.4. use true sport car data as operating mode checking data and to bring operating mode authentication module into, verify condition model Accuracy.

7. as described in claim 5 or 6 based on artificial intelligence's supervised learning linear regression method set up different automobile types the division of labor The long-range damage identification method of condition, it is characterised in that described supervised learning linear regression method includes:

S1. input variable x is characterized, and the predictive value y of output is desired value；The curve table of matching is shown as y=h (x)；

S2. output y is the linear function of x, and is expressed as matrix form；

S3. introduce cost function, use gradient descent algorithm, after initializing learning parameter, repeat the value of renewal learning parameter, To obtain minimum side's more new regulation.

8. set up different automobile types divided working status based on artificial intelligence's supervised learning linear regression method as claimed in claim 7 remote Journey damage identification method, it is characterised in that described descent algorithm is: batch gradient declines and/or statistical gradient declines, described in have supervision The specific algorithm of study linear regression method is:

Output y is that the linear function of x is:

h_θ(x)=θ₀+θ₁x₁+θ₂x₂

θ i is parameter, and n is Grad, it is assumed that x₀=1 above formula is expressed as matrix form:

$h\left(x\right)=\underset{i=0}{\overset{n}{\&Sigma;}}{\mathrm{\&theta;}}_i{x}_i={\mathrm{\&theta;}}^{T}x,$

θ and x is column vector, and m is Grad, certain training set given, introduces cost function, and it is defined as follows:

$J\left(\mathrm{\&theta;}\right)=\frac{1}{2}\underset{i=1}{\overset{m}{\&Sigma;}}{\left(h_{\mathrm{\&theta;}}\right(x^{\left(i\right)})-y^{\left(i\right)})}^2.$

By initial guess initiation parameter θ, the most constantly change the value of parameter θ so that parameter J (θ) is the least, until Obtain the J (θ) minimized eventually, use gradient descent algorithm, after initiation parameter θ, repeat following renewal equation, to update The value of parameter θ, J (θ) is cost function, and j is the subscript of parameter, and θ j is parameter；

${\mathrm{\&theta;}}_j:={\mathrm{\&theta;}}_j-\mathrm{\&alpha;}\frac{\&part;}{\&part;{\mathrm{\&theta;}}_j}J\left(\mathrm{\&theta;}\right).$

α represents learning rate

$\begin{array}{c}\frac{\&part;}{\&part;{\mathrm{\&theta;}}_j}J\left(\mathrm{\&theta;}\right)=\frac{\&part;}{\&part;{\mathrm{\&theta;}}_j}\frac{1}{2}{\left(h_{\mathrm{\&theta;}}\right(x)-y)}^2\\ =2\&CenterDot;\frac{1}{2}\left(h_{\mathrm{\&theta;}}\right(x)-y)\&CenterDot;\frac{\&part;}{\&part;{\mathrm{\&theta;}}_j}\left(h_{\mathrm{\&theta;}}\right(x)-y)\\ =\left(h_{\mathrm{\&theta;}}\right(x)-y)\&CenterDot;\frac{\&part;}{\&part;{\mathrm{\&theta;}}_j}(\underset{i=0}{\overset{n}{\&Sigma;}}{\mathrm{\&theta;}}_i{x}_i-y)\\ =\left(h_{\mathrm{\&theta;}}\right(x)-y)x_j\end{array}$

Thus, update equation to be reduced to:

${\mathrm{\&theta;}}_j:={\mathrm{\&theta;}}_j+\mathrm{\&alpha;}(y^{\left(i\right)}-h_{\mathrm{\&theta;}}(x^{\left(i\right)}\left)\right)x_j^{\left(i\right)}\mathrm{..}$