CN110260925B - Method and system for detecting quality of driver parking technology, intelligent recommendation method and electronic equipment - Google Patents
Method and system for detecting quality of driver parking technology, intelligent recommendation method and electronic equipment Download PDFInfo
- Publication number
- CN110260925B CN110260925B CN201910631451.8A CN201910631451A CN110260925B CN 110260925 B CN110260925 B CN 110260925B CN 201910631451 A CN201910631451 A CN 201910631451A CN 110260925 B CN110260925 B CN 110260925B
- Authority
- CN
- China
- Prior art keywords
- driver
- parking
- quality
- time sequence
- detecting
- Prior art date
- Legal status (The legal status 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 status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention provides a method for detecting the quality of a driver parking technology, which is based on a plurality of sensors, obtains action data of a steering wheel, a brake and the like in the parking process of a driver and obtains a multi-mode time sequence in a time sequence mode; further, feature extraction is carried out on the multi-mode time sequence based on a neural network architecture of a convolution variational self-encoder, and a simulation sample is generated; and carrying out reconstruction probability detection based on the simulation sample, and judging whether the driver parking technology is good or bad based on the detection result. The intelligent recommendation method provided by the invention can recommend the automatic parking function or other executable functions to the driver when needed based on the judgment of the quality of the driver parking technology. The invention also provides a detection system and electronic equipment for detecting the quality of the driver parking technology, and the detection system and the electronic equipment have the same beneficial effects as the method.
Description
[ technical field ] A method for producing a semiconductor device
The invention relates to the field of mobile data detection, in particular to a method and a system for detecting the quality of a driver parking technology, an intelligent recommendation method and electronic equipment.
[ background of the invention ]
With the satellite positioning technology, the vehicle-mounted sensing technology is mature and popular. It becomes possible to acquire accurate data of the operation and position of the vehicle and to improve the level of intelligence of the vehicle. For people living in urban space, parking in a narrow space is almost a necessary skill. Automatic parking technology is a high-level function currently provided for many types of automobiles, and is particularly practical for many people, particularly for driving novices.
However, as the level of automobile intelligence increases, the number of vehicle-mounted intelligent devices increases and the vehicle-mounted intelligent devices become more and more complex. Often, people do not have time to carefully study each function of the vehicle he/she purchases, and then the functions are automatically awakened properly, so that the automatic vehicle-mounted function recommendation is formed, and the vehicle-mounted function recommendation is very practical.
[ summary of the invention ]
The invention provides a detection method and a system thereof for detecting the quality of a driver parking technology, an intelligent recommendation method and electronic equipment, aiming at solving the technical problem that the quality of the driver driving operation is difficult to accurately confirm.
In order to solve the technical problems, the invention provides the following technical scheme: a detection method for the quality of a driver parking technology comprises the following steps: step S1, processing the vehicle sensing data in the driver parking operation process in a time series mode to obtain a multi-modal time series; step S2, carrying out convolution coding on the multi-modal time sequence based on the neural network architecture of the convolution variational self-coder, extracting the features in the multi-modal time sequence, mapping vector values and carrying out feature generalization based on the extracted features, generating an implicit variable vector space, and carrying out convolution decoding on the implicit variable vector space to generate a simulation sample; and step S3, carrying out reconstruction probability detection based on the simulation sample, and judging whether the driver parking technology is good or bad based on the detection result.
Preferably, in the step S1, the vehicle sensing data includes data obtained based on a sensor for a braking action, a shifting action, a reversing action or a GPS position during a driver' S parking operation.
Preferably, in step S1, the processing in a time-series manner specifically includes providing a plurality of continuous time points, and recording data acquired by a plurality of sensors in a time point sequence.
Preferably, the neural network architecture of the convolutional variational self-encoder comprises an encoding module and a decoding module.
Preferably, the above mapping vector values and performing feature generalization based on the extracted features, and generating the hidden variable vector space further includes the following steps: mapping the extracted features of the multi-modal time sequence into a mean vector space and a variance vector space; and adding Gaussian noise data to linearly generate a corresponding hidden variable vector space.
Preferably, the convolution decoding the hidden variable vector space in the convolution decoding of the hidden variable vector space to generate the simulation sample further includes obtaining a mean vector and a variance vector, and generating the corresponding simulation sample accordingly.
In order to solve the above technical problems, the present invention provides another technical solution as follows: a detection system for detecting whether a driver parking technique is good or bad comprises: a plurality of sensors for detecting driver operations and obtaining a plurality of sets of vehicle sensing data; the time sequence acquisition module is used for processing a plurality of groups of vehicle sensing data in the parking operation process of a driver in a time sequence mode to obtain a multi-mode time sequence; the sample generation module is used for carrying out feature extraction on the multi-modal time sequence based on a neural network architecture of the convolution variational self-encoder and generating a simulation sample; the reconstruction probability module is used for carrying out reconstruction probability detection based on the simulation sample and judging whether the driver parking technology is good or bad based on the detection result; wherein the time series obtaining module further comprises: the network architecture construction module is used for constructing a neural network architecture of the convolution variational self-encoder; the neural network architecture of the convolutional variational self-encoder comprises an encoding module and a decoding module; the characteristic extraction module is used for carrying out convolutional coding on the multi-modal time sequence by utilizing a neural network architecture of a convolutional variational self-encoder and extracting characteristics in the multi-modal time sequence; the hidden variable acquisition module is used for mapping vector values and performing feature generalization based on the extracted features, and generating a hidden variable vector space; and the simulation sample generation module is used for decoding the hidden variable vector space through convolution so as to generate a simulation sample.
In order to solve the above technical problems, the present invention provides another technical solution as follows: an intelligent recommendation method is based on the detection method of the quality of the driver parking technology to obtain the grade of the quality of the driver parking technology, and selects and recommends the automatic parking function based on the grade of the quality of the driver parking technology.
In order to solve the above technical problems, the present invention provides another technical solution as follows: an electronic device comprises a storage unit for storing a computer program and a processing unit for executing the method for detecting the superiority and inferiority of the driver parking technique as described above by means of the computer program stored in the storage unit.
Compared with the prior art, the method and the system for detecting the quality of the driver parking technology, the intelligent recommendation method and the electronic equipment have the following beneficial effects:
the invention provides a method for detecting the quality of a driver parking technology, which is different from the existing behavior detection method and can effectively solve the problem of detection errors caused by uncertainty of the operation behavior of the driver. Specifically, in the invention, a multi-mode time sequence is obtained by a plurality of groups of vehicle sensing data in the parking operation process of a driver in a time sequence mode; the method further comprises the steps of carrying out feature extraction on the multi-mode time sequence based on a neural network architecture of a convolution variational self-encoder, and generating a simulation sample, so that dominant and recessive features can be obtained, and therefore big data analysis can be realized, and a more accurate detection result can be obtained; and carrying out reconstruction probability detection based on the simulation sample, and judging whether the driver parking technology is good or bad based on the detection result. The method provided by the invention aims to obtain and correspondingly process driving action data such as a steering wheel, a brake and the like in the parking process of a driver based on various sensors, so that the judgment on the quality of the parking technology of the driver can be realized, and further, the automatic parking function or other executable functions can be recommended to the driver when needed.
In order to better obtain data from the multi-modal time sequence, a neural network framework of a convolutional variational self-encoder is constructed in the invention and comprises an encoding module and a decoding module, the multi-modal time sequence is further subjected to convolutional encoding by utilizing the neural network framework of the convolutional variational self-encoder, and the characteristics in the multi-modal time sequence are extracted; based on the extracted features, mapping vector values, performing feature generalization and generating a hidden variable vector space; and decoding the latent variable vector space by convolution to generate a simulated sample. Based on the encoding process and the decoding process, the characteristics of the multi-modal time sequence can be extracted, and the generalized processing of the characteristics can be realized.
Furthermore, the method for detecting the quality of the driver parking technology further comprises the steps of obtaining a mean vector space and a variance vector space by mapping the extracted features of the multi-modal time sequence, facilitating feature expansion based on the extracted features, obtaining more new feature data according with the original feature rule by adding Gaussian noise data, and generating a corresponding hidden variable vector space linearly based on the new feature data. Based on the steps, the anomaly detection can be carried out on the multi-dimensional complex multi-modal time sequence on the basis of a small number of samples.
Further, in the present invention, the hidden variable vector space is decoded by convolution, and the method further comprises obtaining a mean vector and a variance vector, and generating a corresponding simulation sample. Based on the steps, the obtained simulation sample can meet the characteristic rule of the original multi-modal time sequence, and the problem of data deviation can be reduced.
In the invention, the analog sample is detected by using an anomaly detection method introduced by the reconstruction probability from the variational automatic encoder in the reconstruction probability detection based on the analog sample, so that the problem of poor multi-dimensional data detection effect when the existing anomaly detection technology is carried out based on a distance method, a density method and a clustering method can be solved.
The vehicle-mounted intelligent recommendation method provided by the invention can overcome the problem that the existing technology is difficult to accurately evaluate the driver parking technology by adopting the detection method of the quality of the driver parking technology, and the parking auxiliary technology is a high-grade function which is equipped for many types of automobiles at present, and is particularly practical for many people, particularly for driving novices. However, as the level of automobile intelligence increases, the number of vehicle-mounted intelligent devices increases and the vehicle-mounted intelligent devices become more and more complex. Therefore, the method for detecting the quality of the driver parking technology according to the present invention can provide a technical solution that can automatically recommend automatic parking based on the driver's driving technology.
The invention also provides a system for detecting the quality of the driver parking technology and electronic equipment, which have the same beneficial effects as the method for detecting the quality of the driver parking technology, and can realize the acquisition and corresponding processing of action data of a steering wheel, a brake and the like in the driver parking process based on various sensors, thereby realizing the judgment of the quality of the driver parking technology.
[ description of the drawings ]
Fig. 1 is a flowchart illustrating steps of a method for detecting whether a driver's parking skill is good or bad according to a first embodiment of the present invention.
Fig. 2 is a distribution table of sensor sample data in time series.
Fig. 3 is a flowchart illustrating the specific steps in step S2 shown in fig. 1.
Fig. 4 is a schematic diagram of an encoder convolutional network structure.
Fig. 5 is a schematic diagram of a decoder convolutional network structure.
Fig. 6 is a flowchart illustrating a detailed step of step S23 shown in fig. 3.
FIG. 7 is a functional block diagram of a system for detecting whether the driver parking skill provided in the second embodiment of the present invention is good or bad.
Fig. 8 is a schematic diagram of specific functional units of the address word segmentation module shown in fig. 7.
Fig. 9 is a schematic diagram of specific functional modules of the hidden variable acquiring module shown in fig. 8.
Fig. 10 is a flowchart illustrating steps of an intelligent recommendation method according to a third embodiment of the present invention.
Fig. 11 is a functional block diagram of an electronic device provided in a fourth embodiment of the present invention.
The attached drawings indicate the following:
20. a detection system for detecting the quality of the driver parking technology; 21. a sensor; 22. a time sequence acquisition module; 23. a sample generation module; 24. a reconstruction probability module; 231. a network architecture construction unit; 232. a feature extraction unit; 233. a hidden variable acquisition unit; 234. a simulation sample generation unit; 2331. a feature mapping unit; 2332. generating a hidden variable unit;
30. an electronic device; 31. a storage unit; 32. and a processing unit.
[ detailed description ] embodiments
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Referring to fig. 1, a first embodiment of the present invention provides a method 10 for detecting whether a driver's parking skill is good or bad, which includes the following steps:
step S1, obtaining a multi-modal time sequence in a time sequence mode from a plurality of groups of vehicle sensing data in the parking operation process of a driver;
step S2, performing feature extraction on the multi-modal time sequence based on the neural network architecture of the convolution variational self-encoder, and generating a simulation sample; and
and step S3, carrying out reconstruction probability detection based on the simulation sample, and judging whether the driver parking technology is good or bad based on the detection result.
Specifically, the plurality of sets of vehicle sensing data described in the above step S1 may be obtained based on sensors installed in the vehicle. The sensor can be installed in a vehicle body, and can acquire information such as braking action, gear shifting action, reversing action, GPS position and the like in the parking operation process of a driver based on the sensor, and a multi-mode time sequence is formed in a time sequence mode.
The multi-modal time series specifically includes a plurality of sensor corresponding values in the same time series.
In the step S1, the processing in a time-series manner is to provide a plurality of continuous time points and record the data acquired by the plurality of sensors in a time-point sequence.
In the present embodiment, the collected sensor-related data is closely related to the driving operation action of the driver. For example, in one embodiment, in order to better record the specific operation of the driver during the backing process, in the present invention, the specific operation of the driver can be recorded according to the sensor provided in the vehicle body. For example, the starting time of the sensor to acquire the corresponding operation may be set to be 5 minutes backward from the end of parking, wherein the frequency of the sensor sampling is 1 second sampling, as shown in fig. 2, the horizontal axis represents the time point t (specifically, may be represented as t1, t2 … … tn), and the vertical axis represents the value s sensed by the sensor (specifically, may be represented as s1, s2 … … sn); wherein, referring to fig. 2, t1, t2, t3 … … tn respectively represent 5 minutes, 4 minutes 59 seconds, 4 minutes 58 seconds, 4 minutes 57 seconds, and so on. While the corresponding s1 may represent data acquired for a brake corresponding sensor, s2 may represent data acquired for a shift corresponding sensor, s3 may represent data acquired for a steering corresponding sensor, s4 may represent data acquired for a GPS positioning system sensor, and so on. Each value s can be represented as a response value of the corresponding sensor at the time point, and for the convenience of calculation, the value detected by the sensor can be determined as a real value of 0-1, for example, a value obtained by normalizing the angle value obtained by the steering sensor. It will be appreciated that for ease of calculation, the appropriate data processing scheme may also be matched based on the amount of data collected.
Therefore, if existing unsupervised (or semi-supervised) type anomaly detection algorithms are chosen, there will be a lack of clear guidance labels. It is difficult to achieve superior results with multiple sensors and fine time granularity operations using manual extraction of features.
In order to better solve the above problem, as shown in fig. 3, in step S2, the method further defines a neural network architecture based on a convolutional variational self-encoder to perform feature extraction on the multi-modal time series, and generates a simulation sample, which specifically includes the following steps:
step S21, constructing a neural network framework of the convolution variational self-encoder; the neural network architecture of the convolutional variational self-encoder comprises an encoding module and a decoding module;
step S22, carrying out convolutional coding on the multi-modal time sequence by utilizing a neural network architecture of a convolutional variational self-coder, and extracting the characteristics in the multi-modal time sequence;
step S23, based on the extracted features, mapping Vector values and performing feature generalization, and generating a Latent Variable Vector Space (Latent Variable Vector Space); and
step S24, the hidden variable vector space is decoded by convolution to generate a simulation sample.
Based on the steps, the rule hidden behind the specific data in the multi-modal time sequence can be obtained, so that more data conforming to the rule can be obtained, and the problem that in the prior art, analysis is only performed on data with a large repetition degree, and analysis of the rule of data with small repeatability is ignored is solved.
In step S21, an encoder convolutional network structure (as shown in fig. 4) and a decoder convolutional network structure (as shown in fig. 5) in the neural network architecture of the convolutional variational self-encoder are specifically constructed, where the corresponding symbols "c", "S", and "u" are respectively expressed as convolution, downsampling, and union. The overall structure of the neural network architecture of the convolution variational self-encoder can be specifically based on an MC-1D-CVAE (multi-channel one-dimensional convolution variational self-encoder) as a feature extractor and an anomaly judgment algorithm framework. Specifically, the construction process comprises the step of deconvolving a convolution kernel into an encoding module and a decoding module of the variational self-encoder.
Furthermore, in order to better construct the neural network architecture of the convolutional variational self-encoder, in some embodiments of the present invention, a deep learning framework BigDL is used to build and train a network model. And (3) taking a certain data period as a reference (for example, the month can be taken as the data period), carrying out technical judgment on each time of parking of each driver, and finally marking a label on the corresponding driver, wherein if the parking technology is the excellent grade or the parking technology is the inferior grade, the intelligent recommendation system can be further used for an on-board intelligent recommendation system to carry out automatic parking function use reminding on an unskilled parking vehicle owner.
The self-encoder principle is an unsupervised learning model of a neural network version, and the neural network can be trained on unlabeled samples. The network training principle can enable the network output to learn the implicit characteristics of the data (usually expressed as low-dimensional and basic hidden unit characteristics) under the condition of being as close to the input sample as possible. The variational self-encoder is a probability model of a neural network, the variational self-encoder realizes the generation of samples by learning the distribution of the samples, and the samples are abundant in quantity, so that the generated samples can better embody the characteristics of original data.
As shown in fig. 6, in step S23, based on the extracted features, vector values are mapped and feature generalization is performed, and a hidden variable vector space is linearly generated, which may be further subdivided into the following steps:
step S231, mapping the extracted features of the multi-modal time sequence into a mean vector space and a variance vector space; and
step S232, Gaussian noise data is added to linearly generate a corresponding hidden variable vector space.
In the above steps S231 to S232, after the features are extracted, noise is added to make the features more generalized, so that more and richer virtual samples can be generated, thereby enhancing the samples.
Further, the step S24 may be further subdivided into: and carrying out convolution decoding on the implicit variable vector space to obtain a mean vector and a variance vector, and generating a corresponding simulation sample according to the mean vector and the variance vector.
In the above steps, the hidden variable vector space may be understood as a set of hidden variables, and the set is closed to addition and multiplication of the hidden variables, that is, the vector space closed to addition and multiplication is also referred to as a linear space.
In the step S3, the reconstruction probability detection is performed based on the simulation sample, and the quality of the driver parking technique is determined based on the detection result, specifically including detecting the simulation sample by using an anomaly detection method introduced by the reconstruction probability from the variational automatic encoder. Wherein the reconstruction probability can be combined with the probability characteristics of the variational auto-encoder by considering the concept of variability.
In the above embodiments of the present invention, the purpose of anomaly detection is to automatically find sample data or patterns of significant differences from data, and usually such samples only account for a small part of the total samples. The traditional anomaly detection technology comprises a distance-based method, a density-based method and a clustering method, and multi-dimensional data processing cannot be carried out. And the neural network is used for carrying out abnormity detection, the strategy is basically to learn the overall characteristics of the sample by using the neural network, and detect the sample with significant difference compared with the overall characteristics. Thus, the self-encoder is a very natural anomaly detector. In a variational autoencoder, the optimized objective function (also called evidence lower bound):
wherein ELBO (evidence Lower bound) is expressed in the functional qθMaximize the lower bound of evidence. Wherein x isiIs fixed, we can define xiDistribution under conditions q (z | x)i). This allows us to select a different p (z) for each x, the first term of the above equation, which is the probability that the sample is reconstructed, can be used to construct an anomaly score (AnomalyScore):
it can be seen that, in the abnormality detection process in step S3 described above, the determination may be made based on this abnormality score. And if the abnormal value is large, the corresponding multi-modal time sequence is judged to be abnormal. Reconstruction probability is a probability measure that takes into account the variability of the distribution of variables. By using the generation characteristics of the variational autoencoder, data reconstruction can be deduced to analyze the root cause of the anomaly.
In this embodiment, for convenience of determination, an abnormality threshold may be set, and when the abnormality score exceeds the abnormality threshold, the multimodal time series is determined as abnormal, and at this time, the driver corresponding to the parking driving operation is determined as the inferior grade.
Based on the detection method for detecting the quality of the driver parking technology, provided by the invention, the action of the parking operation of the driver can be identified by using sensor data, so that the parking technology of the driver can be judged by responding to the driving action data such as steering, braking and the like of a steering wheel in the parking process of the driver, a driver with unskilled parking technology can be judged, and the automatic parking function can be recommended under a proper situation.
By utilizing the neural network architecture of the convolutional variational self-encoder, the whole sample size can be more and richer in convolutional encoding. Compared with other model architectures, the characteristics, especially the extraction of the recessive characteristics, can be realized by using a small amount of samples, so that the driving operation of a driver with higher behavior dimensionality is met, and the quick and accurate judgment on the quality of the driver parking technology can be improved.
In the method, based on the acquisition of the hidden variable vector space, the problem that some driving operation actions of a driver are low in frequency and cannot be effectively detected can be solved, and the hidden features can be correspondingly detected by using the method provided by the embodiment.
Compared with the prior art that other encoders need massive samples to perform corresponding operation or simulation, the neural network architecture of the convolutional variational self-encoder provided by the embodiment can realize feature extraction and sample generation by using a variational self-encoder, and can achieve a better anomaly detection effect after a small amount of samples are labeled.
Referring to fig. 7, a second embodiment of the present invention provides a system 20 for detecting whether a driver parking skill is good or bad, comprising:
a plurality of sensors 21 for detecting driver operations and obtaining a plurality of sets of vehicle sensing data; the sensor can record action responses such as steering wheel rotation, braking action, reversing speed and the like of the vehicle.
The time sequence acquisition module 22 is used for acquiring a multi-modal time sequence from a plurality of groups of vehicle sensing data in the parking operation process of a driver in a time sequence mode;
the sample generation module 23 is configured to perform feature extraction on the multi-modal time sequence based on a neural network architecture of the convolutional variational self-encoder, and generate a simulation sample; and
and the reconstruction probability module 24 is used for performing reconstruction probability detection based on the simulation sample and judging whether the driver parking technology is good or bad based on the detection result.
Wherein, the sensor 21 includes a combination of any several of an angle sensor, a pressure sensor, a flow sensor, a speed sensor, an acceleration sensor, a levelness sensor, a transmission sensor or a distance sensor. Thus, the steering angle of the steering wheel of the vehicle body, the speed and the acceleration of the vehicle body in the parking process and the like can be detected.
In order to overcome the defect of processing the multi-modal time sequence by the existing anomaly detection algorithm and the manual feature extraction method, in this embodiment, please refer to fig. 8, the time sequence obtaining module 23 further includes:
a network architecture construction unit 231 for constructing a neural network architecture of the convolutional variational self-encoder; the neural network architecture of the convolutional variational self-encoder comprises an encoding module and a decoding module;
the feature extraction unit 232 is configured to perform convolutional coding on the multi-modal time sequence by using a neural network architecture of a convolutional variational self-encoder, and extract features in the multi-modal time sequence;
a hidden variable obtaining unit 233, configured to map a vector value based on the extracted features, perform feature generalization, and generate a hidden variable vector space; and
a simulation sample generating unit 234, configured to decode the hidden variable vector space by convolution to generate a simulation sample.
Referring to fig. 9, in the present embodiment, the hidden variable acquiring module 233 further includes:
a feature mapping unit 2331 for mapping the extracted features of the multi-modal time series to a mean vector space and a variance vector space; and
a hidden variable generation unit 2332 for adding gaussian noise data to linearly generate a corresponding hidden variable vector space.
In this embodiment, the specific contents of the neural network architecture related to the convolution variational self-encoder are consistent with the related descriptions in the first embodiment, and are not repeated herein.
Referring to fig. 10, a third embodiment of the present invention provides an intelligent recommendation method P10, which includes the following steps:
a step P01 of obtaining a grade of the driver's parking skill based on the detection method of the driver's parking skill in the first embodiment; and
and step P02, selecting and recommending an automatic parking function based on the quality level of the parking technology.
Furthermore, the intelligent recommendation method 30 may also recommend different service items based on the quality level of the parking technology, such as easier parking space or parking lot recommendation, insurance purchase recommendation, and the like.
In the above step P02, specifically, the level of the driver's parking skill based on the detection method of the driver's parking skill in the above first embodiment may be divided into a plurality of levels, and different recommendation functions may be matched based on different levels.
The related limitations of the method for detecting the quality of the driver parking technique are the same as those of the first embodiment, and are not repeated herein.
Referring to fig. 11, a fourth embodiment of the present invention provides an electronic device 30, where the electronic device 30 includes a storage unit 31 and a processing unit 32, the storage unit 31 is used for storing a computer program, and the processing unit 32 is used for executing the specific steps of the data verification method for detecting abnormal values in the first embodiment by using the computer program stored in the storage unit 31.
In some specific embodiments of the present invention, the electronic device 30 may be hardware or software. When the electronic device 30 is hardware, it may be various electronic devices having a display screen and supporting video playing, including but not limited to a smart phone, a tablet computer, an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III, motion Picture Experts Group Audio Layer 3), an MP4 player (Moving Picture Experts Group Audio Layer IV, motion Picture Experts Group Audio Layer 4), a laptop computer, a desktop computer, and the like. When the electronic device 30 is software, it can be installed in the electronic devices listed above. It may be implemented as multiple pieces of software or software modules (e.g., multiple pieces of software or software modules to provide distributed services) or as a single piece of software or software module. And is not particularly limited herein.
The storage unit 31 includes a storage portion of a Read Only Memory (ROM), a Random Access Memory (RAM), a hard disk, and the like, and the processing unit 32 may perform various appropriate actions and processes according to a program stored in the Read Only Memory (ROM) or a program loaded into the Random Access Memory (RAM). In a Random Access Memory (RAM), various programs and data necessary for the operation of the electronic device 30 are also stored.
The electronic device 30 may further include an input portion (not shown) of a keyboard, a mouse, and the like; the electronic device 30 may further include an output portion (not shown) such as a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, and a speaker and the like; and the electronic device 30 may further include a communication part (not shown) of a network interface card such as a LAN card, a modem, and the like. The communication section performs communication processing via a network such as the internet.
In particular, according to the embodiments of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, the disclosed embodiments of the invention may include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method illustrated in the flowchart. In such an embodiment, the computer program may be downloaded and installed from a network via the communication section.
When executed by the processing unit 32, the computer program performs the above-described functions defined in the method for training a neural network model with an anti-counterfeiting function of the present application. It should be noted that the computer readable medium described herein can be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
In the present application, a computer readable storage medium may also be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In this application, however, a computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, fiber optic cable, RF, etc., or any suitable combination of the foregoing.
Computer program code for carrying out operations for aspects of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).
The flowchart and block diagrams in the figures of the present application illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will be understood that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The units involved in the embodiments of the present invention may be implemented by software or hardware. The described units may also be located in the processor.
As another aspect, a fifth embodiment of the present invention also provides a computer-readable medium, which may be contained in the apparatus described in the above embodiments; or may be present separately and not assembled into the device. The computer-readable medium carries one or more programs, which specifically include: obtaining a multi-modal time sequence from a plurality of groups of vehicle sensing data in the parking operation process of a driver in a time sequence mode; performing feature extraction on the multi-modal time sequence based on a neural network architecture of a convolution variational self-encoder, and generating a simulation sample; and carrying out reconstruction probability detection based on the simulation sample, and judging whether the driver parking technology is good or bad based on the detection result.
Compared with the prior art, the method and the system for detecting the quality of the driver parking technology, the intelligent recommendation method and the electronic equipment have the following beneficial effects:
the invention provides a method for detecting the quality of a driver parking technology, which is different from the existing behavior detection method and can effectively solve the problem of detection errors caused by uncertainty of the operation behavior of the driver. Specifically, in the invention, a multi-mode time sequence is obtained by a plurality of groups of vehicle sensing data in the parking operation process of a driver in a time sequence mode; the method further comprises the steps of carrying out feature extraction on the multi-mode time sequence based on a neural network architecture of a convolution variational self-encoder, and generating a simulation sample, so that dominant and recessive features can be obtained, and therefore big data analysis can be realized, and a more accurate detection result can be obtained; and carrying out reconstruction probability detection based on the simulation sample, and judging whether the driver parking technology is good or bad based on the detection result. The method provided by the invention aims to obtain and correspondingly process driving action data such as a steering wheel, a brake and the like in the parking process of a driver based on various sensors, so that the judgment on the quality of the parking technology of the driver can be realized, and further, the automatic parking function or other executable functions can be recommended to the driver when needed.
In the invention, the analog sample is detected by using an anomaly detection method introduced by the reconstruction probability from the variational automatic encoder in the reconstruction probability detection based on the analog sample, so that the problem of poor multi-dimensional data detection effect when the existing anomaly detection technology is carried out based on a distance method, a density method and a clustering method can be solved.
The vehicle-mounted intelligent recommendation method provided by the invention can overcome the problem that the existing technology is difficult to accurately evaluate the driver parking technology by adopting the detection method of the quality of the driver parking technology, and the parking auxiliary technology is a high-grade function which is equipped for many types of automobiles at present, and is particularly practical for many people, particularly for driving novices. However, as the level of automobile intelligence increases, the number of vehicle-mounted intelligent devices increases and the vehicle-mounted intelligent devices become more and more complex. Therefore, the method for detecting the quality of the driver parking technology according to the present invention can provide a technical solution that can automatically recommend automatic parking based on the driver's driving technology.
The invention also provides a system for detecting the quality of the driver parking technology and electronic equipment, which have the same beneficial effects as the method for detecting the quality of the driver parking technology, and can realize the acquisition and corresponding processing of action data of a steering wheel, a brake and the like in the driver parking process based on various sensors, thereby realizing the judgment of the quality of the driver parking technology.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit of the present invention are intended to be included within the scope of the present invention.
Claims (9)
1. A detection method for the quality of a driver parking technology is characterized in that: the detection method for the quality of the driver parking technology comprises the following steps:
step S1, processing the vehicle sensing data in the driver parking operation process in a time series mode to obtain a multi-modal time series;
step S2, carrying out convolution coding on the multi-modal time sequence based on the neural network architecture of the convolution variational self-coder, extracting the features in the multi-modal time sequence, mapping vector values and carrying out feature generalization based on the extracted features, generating an implicit variable vector space, and carrying out convolution decoding on the implicit variable vector space to generate a simulation sample; and
and step S3, carrying out reconstruction probability detection based on the simulation sample, and judging whether the driver parking technology is good or bad based on the detection result.
2. The method for detecting the quality of driver parking skill as set forth in claim 1, wherein: in the above step S1, the vehicle sensing data includes data obtained based on the sensor for a braking action, a shifting action, a reversing action or a GPS position during a driver' S parking operation.
3. The method for detecting the quality of driver parking skill as set forth in claim 1, wherein: in step S1, the processing in a time-series manner specifically includes providing a plurality of consecutive time points, and recording data acquired by the plurality of sensors in a time-point sequence.
4. The method for detecting the quality of driver parking skill as set forth in claim 1, wherein: the neural network architecture of the convolutional variational self-encoder comprises an encoding module and a decoding module.
5. The method for detecting the quality of driver's parking skill as set forth in claim 4, wherein: the step of mapping vector values and performing feature generalization based on the extracted features and generating a hidden variable vector space further comprises the following steps: mapping the extracted features of the multi-modal time sequence into a mean vector space and a variance vector space; and adding Gaussian noise data to linearly generate a corresponding hidden variable vector space.
6. The method for detecting the quality of driver's parking skill as set forth in claim 4, wherein: after the implicit variable vector space is decoded through convolution, a mean vector and a variance vector are obtained, and a corresponding simulation sample is generated according to the mean vector and the variance vector.
7. The utility model provides a detecting system of driver parking technique goodness which characterized in that: the detection system of driver parking technique goodness includes:
a plurality of sensors for detecting driver operations and obtaining a plurality of sets of vehicle sensing data;
the time sequence acquisition module is used for processing a plurality of groups of vehicle sensing data in the parking operation process of a driver in a time sequence mode to obtain a multi-mode time sequence;
the sample generation module is used for carrying out feature extraction on the multi-modal time sequence based on a neural network architecture of the convolution variational self-encoder and generating a simulation sample; and
the reconstruction probability module is used for carrying out reconstruction probability detection based on the simulation sample and judging whether the driver parking technology is good or bad based on the detection result; wherein the time series obtaining module further comprises:
the network architecture construction module is used for constructing a neural network architecture of the convolution variational self-encoder; the neural network architecture of the convolutional variational self-encoder comprises an encoding module and a decoding module;
the characteristic extraction module is used for carrying out convolutional coding on the multi-modal time sequence by utilizing a neural network architecture of a convolutional variational self-encoder and extracting characteristics in the multi-modal time sequence;
the hidden variable acquisition module is used for mapping vector values and performing feature generalization based on the extracted features, and generating a hidden variable vector space; and
and the simulation sample generation module is used for decoding the hidden variable vector space through convolution so as to generate a simulation sample.
8. An intelligent recommendation method is characterized in that: the method for detecting the quality of the driver's parking technology based on any one of claims 1 to 6 is used for obtaining the grade of the quality of the driver's parking technology and selecting and recommending the automatic parking function based on the grade of the quality of the parking technology.
9. An electronic device, characterized in that: the electronic device comprises a storage unit for storing a computer program and a processing unit for executing the method for detecting the superiority and inferiority of the driver's parking technique according to any one of claims 1 to 6 by means of the computer program stored in the storage unit.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910631451.8A CN110260925B (en) | 2019-07-12 | 2019-07-12 | Method and system for detecting quality of driver parking technology, intelligent recommendation method and electronic equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910631451.8A CN110260925B (en) | 2019-07-12 | 2019-07-12 | Method and system for detecting quality of driver parking technology, intelligent recommendation method and electronic equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110260925A CN110260925A (en) | 2019-09-20 |
CN110260925B true CN110260925B (en) | 2021-06-25 |
Family
ID=67925888
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910631451.8A Active CN110260925B (en) | 2019-07-12 | 2019-07-12 | Method and system for detecting quality of driver parking technology, intelligent recommendation method and electronic equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110260925B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112508058B (en) * | 2020-11-17 | 2023-11-14 | 安徽继远软件有限公司 | Transformer fault diagnosis method and device based on audio feature analysis |
Citations (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6208958B1 (en) * | 1998-04-16 | 2001-03-27 | Samsung Electronics Co., Ltd. | Pitch determination apparatus and method using spectro-temporal autocorrelation |
CN106710306A (en) * | 2016-12-09 | 2017-05-24 | 深圳市元征科技股份有限公司 | Vehicle driving behavior monitoring method and vehicle driving behavior monitoring device |
CN107492251A (en) * | 2017-08-23 | 2017-12-19 | 武汉大学 | It is a kind of to be identified and driving condition supervision method based on the driver identity of machine learning and deep learning |
CN108009520A (en) * | 2017-12-21 | 2018-05-08 | 东南大学 | A kind of finger vein identification method and system based on convolution variation self-encoding encoder neutral net |
CN108182452A (en) * | 2017-12-29 | 2018-06-19 | 哈尔滨工业大学(威海) | Aero-engine fault detection method and system based on grouping convolution self-encoding encoder |
CN108216252A (en) * | 2017-12-29 | 2018-06-29 | 中车工业研究院有限公司 | A kind of subway driver vehicle carried driving behavior analysis method, car-mounted terminal and system |
CN108710865A (en) * | 2018-05-28 | 2018-10-26 | 电子科技大学 | A kind of driver's anomaly detection method based on neural network |
CN108725437A (en) * | 2018-07-17 | 2018-11-02 | 广东工业大学 | A kind of automatic stopping storage method and device |
CN108764064A (en) * | 2018-05-07 | 2018-11-06 | 西北工业大学 | SAR Target Recognition Algorithms based on Steerable filter device and self-encoding encoder |
CN108881196A (en) * | 2018-06-07 | 2018-11-23 | 中国民航大学 | The semi-supervised intrusion detection method of model is generated based on depth |
CN109034953A (en) * | 2018-07-02 | 2018-12-18 | 西南交通大学 | A kind of film recommended method |
WO2019051356A1 (en) * | 2017-09-08 | 2019-03-14 | The General Hospital Corporation | A system and method for automated labeling and annotating unstructured medical datasets |
CN109543100A (en) * | 2018-10-31 | 2019-03-29 | 上海交通大学 | User interest modeling method and system based on Cooperative Study |
CN109559209A (en) * | 2019-01-18 | 2019-04-02 | 深圳创新奇智科技有限公司 | A kind of electric business clothes based on multi-modal information, which are worn, takes recommended method |
CN109726795A (en) * | 2017-10-30 | 2019-05-07 | 罗伯特·博世有限公司 | Method for the central artificial intelligence module of training |
CN109800767A (en) * | 2018-12-12 | 2019-05-24 | 天津津航技术物理研究所 | Object detection method based on HOG feature and self-encoding encoder |
CN109800658A (en) * | 2018-12-26 | 2019-05-24 | 中汽研(天津)汽车工程研究院有限公司 | Parking position type online recognition neural network based and positioning system and method |
CN109910901A (en) * | 2019-02-15 | 2019-06-21 | 苏州工业园区职业技术学院 | A kind of intelligent DAS (Driver Assistant System) with driving behavior research and application function |
CN109919189A (en) * | 2019-01-29 | 2019-06-21 | 华南理工大学 | A kind of depth K mean cluster method towards time series data |
JP2019109306A (en) * | 2017-12-15 | 2019-07-04 | 日本電信電話株式会社 | Voice conversion device, voice conversion method and program |
CN109978217A (en) * | 2018-11-09 | 2019-07-05 | 国网浙江省电力有限公司绍兴供电公司 | Methods of electric load forecasting based on singular spectrum analysis and memory network |
-
2019
- 2019-07-12 CN CN201910631451.8A patent/CN110260925B/en active Active
Patent Citations (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6208958B1 (en) * | 1998-04-16 | 2001-03-27 | Samsung Electronics Co., Ltd. | Pitch determination apparatus and method using spectro-temporal autocorrelation |
CN106710306A (en) * | 2016-12-09 | 2017-05-24 | 深圳市元征科技股份有限公司 | Vehicle driving behavior monitoring method and vehicle driving behavior monitoring device |
CN107492251A (en) * | 2017-08-23 | 2017-12-19 | 武汉大学 | It is a kind of to be identified and driving condition supervision method based on the driver identity of machine learning and deep learning |
WO2019051356A1 (en) * | 2017-09-08 | 2019-03-14 | The General Hospital Corporation | A system and method for automated labeling and annotating unstructured medical datasets |
CN109726795A (en) * | 2017-10-30 | 2019-05-07 | 罗伯特·博世有限公司 | Method for the central artificial intelligence module of training |
JP2019109306A (en) * | 2017-12-15 | 2019-07-04 | 日本電信電話株式会社 | Voice conversion device, voice conversion method and program |
CN108009520A (en) * | 2017-12-21 | 2018-05-08 | 东南大学 | A kind of finger vein identification method and system based on convolution variation self-encoding encoder neutral net |
CN108182452A (en) * | 2017-12-29 | 2018-06-19 | 哈尔滨工业大学(威海) | Aero-engine fault detection method and system based on grouping convolution self-encoding encoder |
CN108216252A (en) * | 2017-12-29 | 2018-06-29 | 中车工业研究院有限公司 | A kind of subway driver vehicle carried driving behavior analysis method, car-mounted terminal and system |
CN108764064A (en) * | 2018-05-07 | 2018-11-06 | 西北工业大学 | SAR Target Recognition Algorithms based on Steerable filter device and self-encoding encoder |
CN108710865A (en) * | 2018-05-28 | 2018-10-26 | 电子科技大学 | A kind of driver's anomaly detection method based on neural network |
CN108881196A (en) * | 2018-06-07 | 2018-11-23 | 中国民航大学 | The semi-supervised intrusion detection method of model is generated based on depth |
CN109034953A (en) * | 2018-07-02 | 2018-12-18 | 西南交通大学 | A kind of film recommended method |
CN108725437A (en) * | 2018-07-17 | 2018-11-02 | 广东工业大学 | A kind of automatic stopping storage method and device |
CN109543100A (en) * | 2018-10-31 | 2019-03-29 | 上海交通大学 | User interest modeling method and system based on Cooperative Study |
CN109978217A (en) * | 2018-11-09 | 2019-07-05 | 国网浙江省电力有限公司绍兴供电公司 | Methods of electric load forecasting based on singular spectrum analysis and memory network |
CN109800767A (en) * | 2018-12-12 | 2019-05-24 | 天津津航技术物理研究所 | Object detection method based on HOG feature and self-encoding encoder |
CN109800658A (en) * | 2018-12-26 | 2019-05-24 | 中汽研(天津)汽车工程研究院有限公司 | Parking position type online recognition neural network based and positioning system and method |
CN109559209A (en) * | 2019-01-18 | 2019-04-02 | 深圳创新奇智科技有限公司 | A kind of electric business clothes based on multi-modal information, which are worn, takes recommended method |
CN109919189A (en) * | 2019-01-29 | 2019-06-21 | 华南理工大学 | A kind of depth K mean cluster method towards time series data |
CN109910901A (en) * | 2019-02-15 | 2019-06-21 | 苏州工业园区职业技术学院 | A kind of intelligent DAS (Driver Assistant System) with driving behavior research and application function |
Non-Patent Citations (4)
Title |
---|
Object tracking from image sequences using adaptive models in fuzzy particle filter;Changyong Yoon 等;《Information Sciences》;20131231;第253卷;第74-99页 * |
变分自编码器模型综述;翟正利 等;《计算机工程与应用》;20190228;第55卷(第3期);第1-9页 * |
商用车辆异常驾驶行为检测算法研究;贾硕 等;《公路交通科技》;20171231;第34卷;第29-36页 * |
基于计算机视觉和深度学习的自动驾驶方法研究;白辰甲;《中国优秀硕士论文全文数据库 工程科技II辑》;20180215(第2期);第C035-129页 * |
Also Published As
Publication number | Publication date |
---|---|
CN110260925A (en) | 2019-09-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Yang et al. | Label-driven reconstruction for domain adaptation in semantic segmentation | |
CN111539942B (en) | Method for detecting face depth tampered image based on multi-scale depth feature fusion | |
CN111078940B (en) | Image processing method, device, computer storage medium and electronic equipment | |
US10073908B2 (en) | Functional space-time trajectory clustering | |
CN110674979A (en) | Risk prediction model training method, prediction device, medium and equipment | |
CN111199459A (en) | Commodity recommendation method and device, electronic equipment and storage medium | |
Gao et al. | A data-driven lane-changing behavior detection system based on sequence learning | |
CN112734873A (en) | Image attribute editing method, device, equipment and medium for resisting generation network | |
CN110260925B (en) | Method and system for detecting quality of driver parking technology, intelligent recommendation method and electronic equipment | |
CN113762454A (en) | Track abnormity detection method and device | |
CN117217368A (en) | Training method, device, equipment, medium and program product of prediction model | |
CN111291715A (en) | Vehicle type identification method based on multi-scale convolutional neural network, electronic device and storage medium | |
Yang et al. | A survey on diffusion models for time series and spatio-temporal data | |
CN116665110B (en) | Video action recognition method and device | |
Gao et al. | Multi-scale space-time transformer for driving behavior detection | |
Zhang | Research on satellite selection algorithm in ship positioning based on both geometry and geometric dilution of precision contribution | |
CN115067945A (en) | Fatigue detection method, device, equipment and storage medium | |
CN113822293A (en) | Model processing method, device and equipment for graph data and storage medium | |
CN113516315A (en) | Wind power generation power interval prediction method, device and medium | |
CN111784181A (en) | Criminal reconstruction quality evaluation system evaluation result interpretation method | |
Zhang et al. | Intelligent vehicle modeling design based on image processing | |
CN115689648B (en) | User information processing method and system applied to directional delivery | |
CN117349677B (en) | Training method, device, equipment, medium and program product of pavement recognition model | |
CN113919893B (en) | Information pushing method and device, electronic equipment and readable medium | |
CN116308763B (en) | Vehicle lending post-lending risk prediction method and system based on convolution self-encoder |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
TA01 | Transfer of patent application right | ||
TA01 | Transfer of patent application right |
Effective date of registration: 20210518 Address after: Room 306, 3rd floor, building 9, Chongqing enlightenment science park, 69 Huoju Avenue, Jiulongpo District, Chongqing 400000 Applicant after: Chongqing qisaidi Technology Co.,Ltd. Address before: Room 201, building 9, Chongqing Qidi science and Technology Park, 69 torch Avenue, Jiulongpo District, Chongqing 400000 Applicant before: AINNOVATION (CHONGQING) TECHNOLOGY Co.,Ltd. |
|
GR01 | Patent grant | ||
GR01 | Patent grant |