CN116720045A - Abnormal unloading behavior identification method and device for transport vehicle and transport vehicle - Google Patents
Abnormal unloading behavior identification method and device for transport vehicle and transport vehicle Download PDFInfo
- Publication number
- CN116720045A CN116720045A CN202311000089.7A CN202311000089A CN116720045A CN 116720045 A CN116720045 A CN 116720045A CN 202311000089 A CN202311000089 A CN 202311000089A CN 116720045 A CN116720045 A CN 116720045A
- Authority
- CN
- China
- Prior art keywords
- transport vehicle
- data
- state
- current state
- representing
- 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.)
- Pending
Links
- 230000002159 abnormal effect Effects 0.000 title claims abstract description 48
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000001514 detection method Methods 0.000 claims abstract description 41
- 230000006399 behavior Effects 0.000 claims abstract description 24
- 238000012549 training Methods 0.000 claims abstract description 18
- 210000002569 neuron Anatomy 0.000 claims description 29
- 238000001914 filtration Methods 0.000 claims description 19
- 230000006870 function Effects 0.000 claims description 16
- 230000009467 reduction Effects 0.000 claims description 16
- 238000004364 calculation method Methods 0.000 claims description 15
- 238000013528 artificial neural network Methods 0.000 claims description 9
- 230000004913 activation Effects 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 8
- 238000006243 chemical reaction Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 4
- 230000001537 neural effect Effects 0.000 claims description 3
- 230000008859 change Effects 0.000 abstract description 13
- 230000001133 acceleration Effects 0.000 description 10
- 230000008569 process Effects 0.000 description 8
- 239000004568 cement Substances 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 4
- 239000003245 coal Substances 0.000 description 4
- 239000004576 sand Substances 0.000 description 4
- 239000004566 building material Substances 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 239000004575 stone Substances 0.000 description 3
- 238000009434 installation Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 238000012512 characterization method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/10—Pre-processing; Data cleansing
- G06F18/15—Statistical pre-processing, e.g. techniques for normalisation or restoring missing data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- 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
- G06N3/047—Probabilistic or stochastic networks
-
- 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
- G06N3/048—Activation functions
-
- 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
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
- G06F2218/04—Denoising
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Abstract
The invention provides a method and a device for identifying abnormal unloading behaviors of a transport vehicle and the transport vehicle, wherein the method comprises the following steps: acquiring state data of the transport vehicle based on a plurality of inertial sensors, acquiring strain data of the transport vehicle based on a plurality of resistance strain gauges, and acquiring position information of the transport vehicle based on a GPS sensor; combining the state data with the strain data to obtain combined data; inputting the combined data into a state detection model obtained by training in advance, and detecting the current state of the transport vehicle; wherein the current state includes a discharge state and a non-discharge state; transmitting state data, strain data and current state of the transport vehicle and position information corresponding to the transport vehicle to a cloud server for storage; according to the invention, the inertial sensor is combined with the resistance strain gauge to acquire signal data of the transport vehicle, so that the small weight change of the vehicle is detected, and whether the vehicle is in a unloading state or not is judged, the detection accuracy is high, and the cost is low.
Description
Technical Field
The present invention relates to the field of data processing, and in particular, to a method and an apparatus for identifying abnormal unloading behavior of a transport vehicle, and a transport vehicle.
Background
Cement, sand, coal and the like are important basic raw materials for national economy construction, and no other materials can replace the materials at home and abroad at present. Along with the high-speed development of the economy in China, the roles of cement, sand and stone and coal in the national economy are increasingly larger. The use amount of cement, sand and stone and coal in the past ten years is always high, and the transportation mode of the cement, sand and stone and coal is land transportation and water transportation, and the building materials are transported by adopting four-axis, six-axis and other multi-axis dump trucks in general land transportation.
Based on the goods value of the cement and other materials, the cement can be transported more than the goods value if transported for a long distance, and the product competitiveness is completely lost, so that the land transportation with the common cement transportation radius is 200km, and the water transportation is 500km. Because of the strong dependence of building materials on areas, common transportation companies will divide different prices according to sales areas, and thus the actual flow of building materials to areas becomes particularly important, which is a focus of attention for various large manufacturers.
In the current technical scheme for detecting abnormal unloading of a transport vehicle, whether the abnormal unloading of the transport vehicle occurs is mainly determined by identifying the weight change of the vehicle-mounted goods. For example, the vehicle weight and the vehicle position are monitored using data (rotation speed, torque, etc.) of the vehicle engine and GPS information, and the weight change position of the transport vehicle are detected from these data. For another example, the system is utilized to automatically match and compare the shipping order information of the vehicle to find abnormal vehicle orders to detect vehicles with abnormal discharge.
The above prior art scheme can effectively detect abnormal unloading behavior of the transport vehicle, but can only effectively detect the change of a large weight (the weight exceeds 3 tons) under the condition of more abnormal unloading, and hardly effectively detect the change of a small weight (the weight change is lower than 3 tons).
At present, in a transport vehicle, abnormal unloading is often not the unloading of the whole vehicle, but artificial or semi-mechanical unloading of small-weight cargoes within hundreds of kilograms to tons, and for this reason, the accurate detection cannot be achieved by the existing abnormal unloading detection scheme.
Disclosure of Invention
The invention mainly aims to provide an abnormal unloading behavior identification method and device for a transport vehicle and the transport vehicle, and aims to overcome the defect that unloading of small-weight goods cannot be accurately detected in the prior art.
In order to achieve the above object, the present invention provides a method for identifying abnormal unloading behavior of a transport vehicle, which is characterized by comprising the following steps:
acquiring state data of the transport vehicle based on a plurality of inertial sensors, acquiring strain data of the transport vehicle based on a plurality of resistance strain gauges, and acquiring position information of the transport vehicle based on a GPS sensor; wherein, the state data is vibration signal data of the transport vehicle;
combining the state data with the strain data to obtain combined data; inputting the combined data into a state detection model obtained by training in advance, and detecting the current state of the transport vehicle; wherein the current state includes a discharge state and a non-discharge state;
the state data, the strain data and the current state of the transport vehicle and the position information corresponding to the transport vehicle are sent to a cloud server for storage; the cloud server judges whether the transport vehicle is in an abnormal unloading state or not based on the current state of the transport vehicle and position information corresponding to the transport vehicle.
Further, the state data collected by each inertial sensor comprises signal data collected by a triaxial accelerometer and a triaxial gyroscope respectively; before said combining said state data with said strain data, comprising:
respectively carrying out sliding average value noise reduction on signal data collected by the triaxial accelerometer and signal data collected by the triaxial gyroscope at the same moment;
the sliding average value noise reduction calculation method comprises the following steps:
where x (i) is the data value at the i-th time, x (i-5) is the data value at the i-5-th time, x (i+5) is the data value at the i+5-th time, and 11 is the size of the mean window.
Further, before the combining the state data with the strain data, the method includes:
carrying out noise reduction treatment on the strain data by adopting median filtering; the method specifically comprises the following steps: arranging the data in the median filtering window from small to large, and then selecting a point at a sixth position as a filtered data point, thereby completing median filtering; wherein the window of median filtering is 11.
Further, the state detection model is obtained based on feedforward neural network training; the feedforward neural network comprises an input layer, a hidden layer and an output layer;
and the combined data is input by the input layer, subjected to feature extraction and data conversion by the hidden layer, and finally output by the output layer, the corresponding state classification and the corresponding probability.
Further, the formula from the input layer to the hidden layer is:
wherein ,representing the weight between the i-th neuron of the input layer and the j-th neuron of the hidden layer,/and>a bias term representing the jth neuron of the hidden layer, f representing the activation function, x (i) being the data value at the ith instant,/v>Weighted input representing hidden layer neuron j, < +.>Representing neural elements of the hidden layer.
Further, the calculation formula from the hidden layer to the output layer is as follows:
wherein ,representing the weight between the jth neuron of the hidden layer and the kth neuron of the output layer,/>A bias term representing the kth neuron of the output layer, f representing the activation function, +.>Representing the weighted output of hidden layer neuron k, < +.>Representing the output layer neuron k output.
Further, the cloud server is specifically configured to:
if the current state of the transport vehicle is a discharge state, judging whether the current state is a preset discharge position or not according to the position information at the moment; if not, judging that the transport vehicle is in an abnormal unloading state currently; if yes, judging that the transport vehicle is in a normal unloading state currently.
Further, the number of the inertial sensors is three, and the inertial sensors are uniformly arranged at the bottom of the frame of the transport vehicle; the resistance strain gauges are eight and are respectively arranged at axle positions of the transport vehicle.
The invention also provides an abnormal unloading behavior recognition device of the transport vehicle, which comprises:
the system comprises an acquisition unit, a control unit and a control unit, wherein the acquisition unit is used for acquiring state data of a transport vehicle based on a plurality of inertial sensors, acquiring strain data of the transport vehicle based on a plurality of resistance strain gauges and acquiring position information of the transport vehicle based on a GPS sensor; wherein, the state data is vibration signal data of the transport vehicle;
the detection unit is used for combining the state data with the strain data to obtain combined data; inputting the combined data into a state detection model obtained by training in advance, and detecting the current state of the transport vehicle; wherein the current state includes a discharge state and a non-discharge state;
the transmitting unit is used for transmitting the state data, the strain data, the current state of the transport vehicle and the position information corresponding to the transport vehicle to the cloud server for storage; the cloud server judges whether the transport vehicle is in an abnormal unloading state or not based on the current state of the transport vehicle and position information corresponding to the transport vehicle.
The invention also provides a transport vehicle comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the methods when executing the computer program.
The invention provides a method and a device for identifying abnormal unloading behaviors of a transport vehicle and the transport vehicle, wherein the method comprises the following steps: acquiring state data of the transport vehicle based on a plurality of inertial sensors, acquiring strain data of the transport vehicle based on a plurality of resistance strain gauges, and acquiring position information of the transport vehicle based on a GPS sensor; wherein, the state data is vibration signal data of the transport vehicle; combining the state data with the strain data to obtain combined data; inputting the combined data into a state detection model obtained by training in advance, and detecting the current state of the transport vehicle; wherein the current state includes a discharge state and a non-discharge state; the state data, the strain data and the current state of the transport vehicle and the position information corresponding to the transport vehicle are sent to a cloud server for storage; the cloud server judges whether the transport vehicle is in an abnormal unloading state or not based on the current state of the transport vehicle and position information corresponding to the transport vehicle. According to the invention, the inertial sensor is combined with the resistance strain gauge to acquire signal data of the transport vehicle, so that the small weight change of the vehicle is detected, and whether the vehicle is in a unloading state or not is judged, the detection accuracy is high, and the cost is low.
Drawings
FIG. 1 is a schematic diagram showing steps of a method for identifying abnormal unloading behavior of a transport vehicle according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of the installation of an inertial sensor in an embodiment of the invention;
FIG. 3 is a schematic diagram illustrating the installation of a resistance strain gauge according to an embodiment of the present invention;
FIG. 4 is a block diagram showing an abnormal unloading behavior recognition apparatus of a transport vehicle according to an embodiment of the present invention;
fig. 5 is a schematic block diagram of a transport vehicle according to an embodiment of the present invention.
The achievement of the objects, functional features and advantages of the present invention will be further described with reference to the accompanying drawings, in conjunction with the embodiments.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
Referring to fig. 1, in one embodiment of the present invention, a method for identifying abnormal unloading behavior of a transport vehicle is provided, including the following steps:
step S1, acquiring state data of a transport vehicle based on a plurality of inertial sensors, acquiring strain data of the transport vehicle based on a plurality of resistance strain gauges, and acquiring position information of the transport vehicle based on a GPS sensor; wherein, the state data is vibration signal data of the transport vehicle;
step S2, combining the state data with the strain data to obtain combined data; inputting the combined data into a state detection model obtained by training in advance, and detecting the current state of the transport vehicle; wherein the current state includes a discharge state and a non-discharge state;
step S3, the state data, the strain data, the current state of the transport vehicle and the position information corresponding to the transport vehicle are sent to a cloud server for storage; the cloud server judges whether the transport vehicle is in an abnormal unloading state or not based on the current state of the transport vehicle and position information corresponding to the transport vehicle.
In this embodiment, the above scheme is applied to abnormal unloading behavior recognition of the transport vehicle, and is different from the scheme adopting vehicle engine data in the traditional scheme, and in this embodiment, the scheme adopts an inertial sensor to collect vibration signal data of the transport vehicle, and combines a resistance strain gauge to collect strain data of the transport vehicle, so as to detect small weight change of the vehicle, further judge whether to be in an unloading state, and has high detection accuracy and low cost.
As described in the above step S1, the inertial sensor is mounted at the bottom of the frame of the transport vehicle (see fig. 2), and is capable of collecting state data of the transport vehicle, and the resistive strain gauges are eight and are mounted at axle positions of the transport vehicle (see fig. 3). The GPS sensor is arranged on the transport vehicle and used for collecting position information of the transport vehicle. In this embodiment, the sensing terminal (inertial sensor, resistance strain gauge, GPS sensor, etc.), host terminal, cloud server, and mobile terminal are included. Specifically, at the sensing end, inertial sensors (IMUs, inertial Measurement Unit), resistive strain gauges, and GPS sensors are mainly utilized. And the inertial sensor (model MPU 6050) is provided with a triaxial accelerometer and a triaxial gyroscope and is used for detecting vibration signal data generated by a frame when the transport vehicle is loaded and unloaded. When the existing transport vehicle is unloaded, severe collision can be generated to the bottom of the carriage, and then the inertial sensor can acquire high-frequency vibration signals. In order for the inertial sensors to better receive car status data, the three inertial sensors need to be uniformly mounted at the bottom of the car and connected to the host. The GPS sensor is mainly used for detecting the position information of the transport vehicle in real time and recording the position coordinates of the state change of the transport vehicle. The resistance strain gauge is a sensor for measuring deformation of an object, and when the shape of the resistance strain gauge is changed, the resistance value is changed. By measuring the change in the resistance value, the degree of deformation of the object can be known. The above-mentioned resistance strain gauge is mounted at an axle of the carrier vehicle, and as shown in fig. 3, 8 resistance strain gauges are mounted in total. When the transportation vehicle is loaded, obvious deformation can occur to the axletree, and then resistance strain gauge can have corresponding numerical variation. When the transport vehicle is unloaded, the vehicle sound load is relieved, and the resistance strain count value has corresponding value change.
As described in the above steps S2-S3, a processing chip (STM 32 chip) and a machine learning algorithm are provided on the host computer to perform state recognition on the signal data transmitted from the inertial sensor and the resistance strain gauge, and then output 2 state results of whether the current transport vehicle is unloaded or not. And finally, sending the state data of the transport vehicle, the strain data of the transport vehicle, the current state of the transport vehicle and the position information corresponding to the transport vehicle to a cloud server for storage, so that the current monitored cargo flow direction information of the transport vehicle can be queried on a mobile terminal such as an android or a computer webpage.
On the cloud: the method is mainly used for storing and calculating whether the current monitored transport vehicle is unloaded at a preset position, and sending out an alarm when the current monitored transport vehicle is unloaded at an unscheduled position, so that the transport vehicle is illegally unloaded at the moment, and the transport vehicle is monitored.
In this embodiment, vibration signal data of the transport vehicle is collected by adopting a plurality of inertial sensors, and strain data of the transport vehicle is collected by combining the resistance strain gauge, so that small weight change of the vehicle is detected, and whether the vehicle is in a unloading state is judged, the detection accuracy is high, and the cost is low. The key point of the embodiment of the invention is that the abnormal unloading behavior of the transport vehicle is monitored in a whole process by fusing the data of the inertial sensor and the resistance strain gauge, and the monitored abnormal unloading weight can be as low as hundreds of kilograms of small-weight abnormal unloading to tens of tons of large-weight abnormal unloading.
In one embodiment, the state data collected by each inertial sensor includes signal data collected by a tri-axis accelerometer and a tri-axis gyroscope, respectively; before said combining said state data with said strain data, comprising:
respectively carrying out sliding average value noise reduction on signal data collected by the triaxial accelerometer and signal data collected by the triaxial gyroscope at the same moment;
for the obtained state data of the inertial sensor, because the obtained state data contains other noise, preliminary noise reduction is needed, and a sliding average noise reduction method is adopted to process the state data.
The sliding average value noise reduction calculation method comprises the following steps:
where x (i) is the data value at the i-th time, x (i-5) is the data value at the i-5-th time, x (i+5) is the data value at the i+5-th time, and 11 is the size of the mean window.
The status data collected by each inertial sensor is mainly 1 inertial sensor data, which, considering the collection frequency of 20Hz, will obtain 3 accelerometer values and 3 gyroscope values at each time stamp. The values of the three axes of the three-axis accelerometer are respectively marked as Acc_x, acc_y and Acc_z, the values of the three axes of the three-axis gyroscope are marked as Gro_x, gro_y and Gro_z, and the data size of the inertial sensor is 20, 18, namely 20 pieces of data are acquired every second. The 8 resistance strain gauge sensor will also acquire 20 data points per second, so the data size acquired is [20,8]. Because the data acquired by the inertial sensor and the resistance strain gauge belong to time sequence data, the original data is required to be obtained by adopting a sliding window, wherein the size of the sliding window is 90, the sliding step length is 30, so that the IMU data size in the window is [90, 18], and the strain gauge data size is [90,8].
The output size of the 18 time series data of the three inertial sensors is still 90, 18 after average filtering and noise reduction.
In an embodiment, a normalization process may be performed, where the normalized calculation formula is:
wherein ,mean value of data axis, +.>Refers to the variance of the data axis.
In an embodiment, the state data collected by the inertial sensor includes signal data collected by a tri-axis accelerometer and a tri-axis gyroscope respectively;
before said combining said state data with said strain data, comprising:
respectively carrying out characteristic calculation on signal data collected by the triaxial accelerometer and signal data collected by the triaxial gyroscope at the same moment to obtain corresponding acceleration energy value characteristics and gyroscope energy value characteristics;
carrying out combined acceleration characteristic calculation on signal data corresponding to the triaxial accelerometer at the same moment;
carrying out combined gyroscope characteristic calculation on signal data corresponding to the triaxial gyroscope at the same moment;
in this embodiment, after the state data collected by the inertial sensor is obtained, it is required to perform feature engineering, which refers to extracting meaningful and usable features from the state data by using domain knowledge and data processing skills, and using the extracted meaningful and usable features for training and prediction of a machine learning model. Further characterization engineering was done here on 6 data points per time stamp in order to improve the performance of the model, reduce the risk of the model being over fitted in the training data, speed up model training, and enhance the interpretability of the model.
The method specifically comprises the following steps:
and calculating the energy value characteristics, namely the acceleration energy value characteristics and the gyroscope energy value characteristics, wherein the characteristics are obtained by taking data of the three-axis accelerometer and the three-axis gyroscope at the moment i as an absolute value.
Combined acceleration characteristics: and squaring and then squaring three signal data of the triaxial acceleration of each time stamp to obtain the combined acceleration characteristic.
Characteristics of the combined gyroscope: and squaring and then squaring three signal data of the three-axis gyroscope with each time stamp to obtain the characteristics of the combined gyroscope. After the feature engineering stage, 6 energy value features, 1 combined acceleration feature, 1 combined gyroscope feature and the original 6 data are extracted in total to obtain 14 features for model training.
In an embodiment, the calculation formula of the acceleration energy value characteristic is:
the calculation formula of the gyroscope energy value characteristic is as follows:
acc_x_abs (i) refers to the acceleration energy value of the x-axis of the three-axis accelerometer after the absolute value of the moment i, gro_x_abs (i) refers to the gyroscope energy value characteristic of the x-axis of the three-axis gyroscope after the absolute value of the moment i, abs refers to the absolute value function;
the calculation formula of the combined acceleration characteristic is as follows:
the calculation formula of the synthetic gyroscope features is as follows:
wherein x, y, z are three axes of the tri-axis accelerometer and the tri-axis gyroscope;
acc_x (i), acc_y (i) and Acc_z (i) are respectively signal data of x, y and z axes of the triaxial accelerometer at the moment i;
gro_x (i), gro_y (i), gro_z (i) are signal data of the x, y, z axes of the three-axis gyroscope at the moment i respectively.
In an embodiment, before said combining said state data with said strain data, comprising:
carrying out noise reduction treatment on the strain data by adopting median filtering; the method specifically comprises the following steps: arranging the data in the median filtering window from small to large, and then selecting a point at a sixth position as a filtered data point, thereby completing median filtering; wherein the window of median filtering is 11.
For the data acquired by 8 strain gauges, it is necessary to make preliminary noise reduction using median filtering. The window of median filtering is 11, and median filtering noise reduction is respectively carried out on 8 pieces of data acquired by the sliding window, and the noise reduction calculation method comprises the following steps: firstly, arranging data in a median filtering window from small to large, and then selecting a point at a sixth position as a filtered data point, thereby completing median filtering.
Assuming that x (j) is the data value at the j-th time, x (j-5) refers to the data value at the j-5-th time, x (j+5) refers to the data value at the j+5-th time, and 11 refers to the size of the median window. Therefore, the output size of the 8 pieces of time series data of the 8 strain gauge sensors after median filtering noise reduction is still [90,8].
And at the end of data preprocessing, combining the data of the 8 resistance strain gauges and the data of the 3 inertial sensors to obtain data of the [90, 26] size, inputting the data into a state detection model for identification and classification, and obtaining a corresponding detection result, namely the current state of the transport vehicle, wherein the state comprises a unloading state and a non-unloading state.
The state detection model is obtained based on feed-forward neural network training (FNN, factorisation Machine supported Neural Network); the feedforward neural network comprises an input layer, a hidden layer and an output layer;
and the combined data is input by the input layer, subjected to feature extraction and data conversion by the hidden layer, and finally output by the output layer, the corresponding state classification and the corresponding probability.
In this embodiment, the parameter of the input layer is set to 26, that is, the data of each row of the sliding window is taken as input, and the predicted value 0/1 of each row is sequentially output, wherein 0 represents normal unloading, and 1 represents abnormal unloading.
The formula from the input layer to the hidden layer is as follows:
wherein ,representing the weight between the i-th neuron of the input layer and the j-th neuron of the hidden layer,/and>a bias term representing the jth neuron of the hidden layer, f representing the activation function, x (i) being the data value at the ith instant,/v>Weighted input representing hidden layer neuron j, < +.>Representing hidden layersA neural unit. Wherein the activation function is typically a sigmoid function or a ReLU function, which is selected in this embodiment.
The hidden layer mainly serves to perform feature extraction and data conversion, and in particular, the hidden layer can extract higher-level features from input data through a combination of a plurality of neurons. These features may be used for subsequent tasks such as classification, regression, pattern recognition, etc. In this embodiment, the hidden layer is provided with 3 layers, a first layer having a size (32, 48), a second layer having a size (48, 16), and a third layer having a size (16, 2).
The calculation formula from the hidden layer to the output layer is as follows:
wherein ,representing the weight between the jth neuron of the hidden layer and the kth neuron of the output layer,/>A bias term representing the kth neuron of the output layer, f representing the activation function, +.>Representing the weighted output of hidden layer neuron k, < +.>Representing the output layer neuron k output.
In this embodiment, the activation function is a sigmoid function, since here the output is only two classifications.
The output layer is (2, 1) and the network output is a class 2, so two data values will be output. The output layer can obtain probability values of two categories, namely 0 and 1 by using a sigmoid function, and the probability value is higher when the probability value is larger, so that the probability value is higher. 0 denotes non-discharge and 1 denotes discharge.
In this embodiment, the foregoing training process of the feedforward neural network generally employs a back-propagation algorithm, also referred to as an error back-propagation algorithm, to optimize the weights and bias terms in the network by minimizing the loss function. The basic idea of the back propagation algorithm is to propagate the errors of the output layer forward layer by the chain law, calculate the error gradient of each neuron, and thus update the weights and bias terms in the network. The feedforward neural network adopts cross entropy as a loss function, so that the model can better approximate training data to obtain the minimization of the loss function.
In an embodiment, the cloud server is specifically configured to:
if the current state of the transport vehicle is a discharge state, judging whether the current state is a preset discharge position or not according to the position information at the moment; if not, judging that the transport vehicle is in an abnormal unloading state currently; if yes, judging that the transport vehicle is in a normal unloading state currently. When abnormal unloading occurs, an alarm can be sent out.
In one embodiment, each inertial sensor is an identical sensor and each resistive strain gauge is an identical sensor; the step of inputting the combined data into a pre-trained state detection model and detecting the current state of the transport vehicle further comprises the following steps:
acquiring the type of an inertial sensor and the type of a resistance strain gauge;
calling a database; the database comprises a plurality of folders, each folder is provided with a unique identification number, and each folder is stored with a plurality of corresponding preselected state detection models;
acquiring a first character corresponding to the type of the inertial sensor, acquiring a second character corresponding to the type of the resistance strain gauge, and combining the first character with the second character to obtain a combined character; wherein, the corresponding relation between the characters and the types of the sensors is stored in the database.
Sequentially inputting each character of the combined characters into a first set; wherein each character is an element in the first set;
for each folder, sequentially inputting identification characters in the unique identification number of each folder into a second set to obtain second sets respectively corresponding to the folders; wherein each identification character is an element in the second set;
respectively calculating intersection elements of a first set and each second set, acquiring the second set with the largest intersection elements of the first set as a target set, and taking folders corresponding to the target set as target folders;
acquiring identification information of each pre-selected state detection model stored in a target folder;
combining numbers corresponding to the numbers of the inertial sensors and the resistance strain gauges to obtain combined numbers, and encoding the combined numbers to obtain encoded information; for example, if the number of inertial sensors and resistance strain gauges is 3 and 8, respectively, the combination number is 38, and the encoding may be performed using a Base 16 encoding table.
Selecting a preselected state detection model with the same identification information as the coding information from the preselected state detection models as a state detection model; the state detection model is used for detecting the current state of the transport vehicle.
In the present embodiment, the type of inertial sensor and the type of resistance strain gauge are acquired: before starting to detect the current state of the transport vehicle, it is first necessary to acquire the type of inertial sensor and resistance strain gauge. These two sensors are typically used to measure specific physical quantities of the transporter, and play an important role in status detection. The database contains a plurality of folders, each folder has a unique identification number, and a corresponding plurality of preselected state detection models are stored. By accessing the database, a preselected state detection model associated with the transporter can be obtained.
Constructing a combined character set and a first set, and a second set: and respectively acquiring a first character and a second character corresponding to the inertial sensor and the resistance strain gauge according to the types of the inertial sensor and the resistance strain gauge. The two characters are combined and each of the combined characters is taken as an element of the first set. Meanwhile, the identification characters in the unique identification numbers of each folder are sequentially used as one element of the second set.
Calculating intersection elements of the first set and the second set, and finding a target set and a target folder: and respectively calculating intersection elements of the first set and each second set, and finding the second set with the most intersection elements of the first set as a target set. From this target set, the corresponding folder is determined as the target folder for use in subsequent steps.
Acquiring identification information of each pre-selected state detection model in the target folder: once the target folder is determined, the system obtains the stored identification information for each of the preselected state detection models from the folder. Such identification information is typically used to identify and distinguish between different state detection models.
Combining the digital codes and the identification information of the preselected state detection model for matching: based on the number of the inertial sensors and the resistance strain gauges, corresponding numbers are combined and encoded to obtain encoded information. And then selecting a model matched with the coded information from the pre-selected state detection models as a final state detection model. This state detection model will be used to detect the current state of the transporter.
Referring to fig. 4, in an embodiment of the present invention, there is further provided an abnormal unloading behavior recognition apparatus for a transport vehicle, including:
the system comprises an acquisition unit, a control unit and a control unit, wherein the acquisition unit is used for acquiring state data of a transport vehicle based on a plurality of inertial sensors, acquiring strain data of the transport vehicle based on a plurality of resistance strain gauges and acquiring position information of the transport vehicle based on a GPS sensor; wherein, the state data is vibration signal data of the transport vehicle;
the detection unit is used for combining the state data with the strain data to obtain combined data; inputting the combined data into a state detection model obtained by training in advance, and detecting the current state of the transport vehicle; wherein the current state includes a discharge state and a non-discharge state;
the transmitting unit is used for transmitting the state data, the strain data, the current state of the transport vehicle and the position information corresponding to the transport vehicle to the cloud server for storage; the cloud server judges whether the transport vehicle is in an abnormal unloading state or not based on the current state of the transport vehicle and position information corresponding to the transport vehicle.
In this embodiment, for specific implementation of each unit in the above embodiment of the apparatus, please refer to the description in the above embodiment of the method, and no further description is given here.
Referring to fig. 5, there is further provided a transporter, which may be a server, and an internal structure of which may be as shown in fig. 5. The transport vehicle comprises a processor, a memory, a display screen, an input device, a network interface and a database which are connected through a system bus. Wherein the computer is configured to provide computing and control capabilities. The memory of the transport vehicle comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the transport vehicle is used for storing the corresponding data in the embodiment. The network interface of the transport vehicle is used for communicating with an external terminal through network connection. Which computer program, when being executed by a processor, carries out the above-mentioned method.
It will be appreciated by those skilled in the art that the structure shown in fig. 5 is merely a block diagram of a portion of the structure associated with the present invention and is not intended to limit the transportation vehicle to which the present invention is applied.
In summary, the method and the device for identifying abnormal unloading behavior of a transport vehicle and the transport vehicle provided in the embodiments of the present invention include: acquiring state data of the transport vehicle based on a plurality of inertial sensors, acquiring strain data of the transport vehicle based on a plurality of resistance strain gauges, and acquiring position information of the transport vehicle based on a GPS sensor; wherein, the state data is vibration signal data of the transport vehicle; combining the state data with the strain data to obtain combined data; inputting the combined data into a state detection model obtained by training in advance, and detecting the current state of the transport vehicle; wherein the current state includes a discharge state and a non-discharge state; the state data, the strain data and the current state of the transport vehicle and the position information corresponding to the transport vehicle are sent to a cloud server for storage; the cloud server judges whether the transport vehicle is in an abnormal unloading state or not based on the current state of the transport vehicle and position information corresponding to the transport vehicle. According to the invention, the inertial sensor is combined with the resistance strain gauge to acquire signal data of the transport vehicle, so that the small weight change of the vehicle is detected, and whether the vehicle is in a unloading state or not is judged, the detection accuracy is high, and the cost is low.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium provided by the present invention and used in embodiments may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), memory bus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM, among others.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, apparatus, article, or method. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, apparatus, article or method that comprises the element.
The foregoing description is only of the preferred embodiments of the present invention and is not intended to limit the scope of the invention, and all equivalent structures or equivalent processes using the descriptions and drawings of the present invention or direct or indirect application in other related technical fields are included in the scope of the present invention.
Claims (10)
1. The abnormal unloading behavior identification method of the transport vehicle is characterized by comprising the following steps of:
acquiring state data of the transport vehicle based on a plurality of inertial sensors, acquiring strain data of the transport vehicle based on a plurality of resistance strain gauges, and acquiring position information of the transport vehicle based on a GPS sensor; wherein, the state data is vibration signal data of the transport vehicle;
combining the state data with the strain data to obtain combined data; inputting the combined data into a state detection model obtained by training in advance, and detecting the current state of the transport vehicle; wherein the current state includes a discharge state and a non-discharge state;
the state data, the strain data and the current state of the transport vehicle and the position information corresponding to the transport vehicle are sent to a cloud server for storage; the cloud server judges whether the transport vehicle is in an abnormal unloading state or not based on the current state of the transport vehicle and position information corresponding to the transport vehicle.
2. The method for identifying abnormal unloading behavior of a transport vehicle according to claim 1, wherein the state data collected by each inertial sensor includes signal data collected by a three-axis accelerometer and a three-axis gyroscope, respectively; before said combining said state data with said strain data, comprising:
respectively carrying out sliding average value noise reduction on signal data collected by the triaxial accelerometer and signal data collected by the triaxial gyroscope at the same moment;
the sliding average value noise reduction calculation method comprises the following steps:
where x (i) is the data value at the i-th time, x (i-5) is the data value at the i-5-th time, x (i+5) is the data value at the i+5-th time, and 11 is the size of the mean window.
3. The method of identifying abnormal discharge behavior of a vehicle according to claim 1, wherein prior to said combining said status data with said strain data, comprising:
carrying out noise reduction treatment on the strain data by adopting median filtering; the method specifically comprises the following steps: arranging the data in the median filtering window from small to large, and then selecting a point at a sixth position as a filtered data point, thereby completing median filtering; wherein the window of median filtering is 11.
4. The method for identifying abnormal unloading behaviors of a transport vehicle according to claim 1, wherein the state detection model is obtained based on feedforward neural network training; the feedforward neural network comprises an input layer, a hidden layer and an output layer;
and the combined data is input by the input layer, subjected to feature extraction and data conversion by the hidden layer, and finally output by the output layer, the corresponding state classification and the corresponding probability.
5. The method for identifying abnormal unloading behavior of a transportation vehicle according to claim 4, wherein the formula from the input layer to the hidden layer is:
wherein ,representing the weight between the i-th neuron of the input layer and the j-th neuron of the hidden layer,/and>a bias term representing the jth neuron of the hidden layer, f representing the activation function, x (i) being the data value at the ith instant,/v>Weighted input representing hidden layer neuron j, < +.>Representing neural elements of the hidden layer.
6. The method for identifying abnormal unloading behavior of a transport vehicle according to claim 5, wherein the hidden layer to output layer calculation formula is:
wherein ,representing the weight between the jth neuron of the hidden layer and the kth neuron of the output layer,/>A bias term representing the kth neuron of the output layer, f representing the activation function, +.>Representing the weighted output of hidden layer neuron k, < +.>Representing the output layer neuron k output.
7. The method for identifying abnormal unloading behavior of a transport vehicle according to claim 1, wherein the cloud server is specifically configured to:
if the current state of the transport vehicle is a discharge state, judging whether the current state is a preset discharge position or not according to the position information at the moment; if not, judging that the transport vehicle is in an abnormal unloading state currently; if yes, judging that the transport vehicle is in a normal unloading state currently.
8. The method for identifying abnormal unloading behavior of a transport vehicle according to claim 1, wherein the inertial sensors are three and are uniformly installed at the bottom of a vehicle frame of the transport vehicle; the resistance strain gauges are eight and are respectively arranged at axle positions of the transport vehicle.
9. An abnormal discharge behavior recognition device for a transport vehicle, comprising:
the system comprises an acquisition unit, a control unit and a control unit, wherein the acquisition unit is used for acquiring state data of a transport vehicle based on a plurality of inertial sensors, acquiring strain data of the transport vehicle based on a plurality of resistance strain gauges and acquiring position information of the transport vehicle based on a GPS sensor; wherein, the state data is vibration signal data of the transport vehicle;
the detection unit is used for combining the state data with the strain data to obtain combined data; inputting the combined data into a state detection model obtained by training in advance, and detecting the current state of the transport vehicle; wherein the current state includes a discharge state and a non-discharge state;
the transmitting unit is used for transmitting the state data, the strain data, the current state of the transport vehicle and the position information corresponding to the transport vehicle to the cloud server for storage; the cloud server judges whether the transport vehicle is in an abnormal unloading state or not based on the current state of the transport vehicle and position information corresponding to the transport vehicle.
10. A transporter comprising a memory and a processor, the memory having stored therein a computer program, characterized in that the processor, when executing the computer program, carries out the steps of the method according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311000089.7A CN116720045A (en) | 2023-08-10 | 2023-08-10 | Abnormal unloading behavior identification method and device for transport vehicle and transport vehicle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311000089.7A CN116720045A (en) | 2023-08-10 | 2023-08-10 | Abnormal unloading behavior identification method and device for transport vehicle and transport vehicle |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116720045A true CN116720045A (en) | 2023-09-08 |
Family
ID=87875588
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311000089.7A Pending CN116720045A (en) | 2023-08-10 | 2023-08-10 | Abnormal unloading behavior identification method and device for transport vehicle and transport vehicle |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116720045A (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108760003A (en) * | 2018-07-27 | 2018-11-06 | 杭州网源科技有限公司 | A kind of vehicle load measuring system |
| CN110379041A (en) * | 2019-07-18 | 2019-10-25 | 亳州职业技术学院 | A kind of travel position intelligence automatic identifying method based on vehicle location |
| CN113091866A (en) * | 2021-04-06 | 2021-07-09 | 深圳市汉德网络科技有限公司 | Method and device for measuring load mass of automobile in real time |
| CN113807470A (en) * | 2021-11-17 | 2021-12-17 | 腾讯科技（深圳）有限公司 | Vehicle driving state determination method and related device |
-
2023
- 2023-08-10 CN CN202311000089.7A patent/CN116720045A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108760003A (en) * | 2018-07-27 | 2018-11-06 | 杭州网源科技有限公司 | A kind of vehicle load measuring system |
| CN110379041A (en) * | 2019-07-18 | 2019-10-25 | 亳州职业技术学院 | A kind of travel position intelligence automatic identifying method based on vehicle location |
| CN113091866A (en) * | 2021-04-06 | 2021-07-09 | 深圳市汉德网络科技有限公司 | Method and device for measuring load mass of automobile in real time |
| CN113807470A (en) * | 2021-11-17 | 2021-12-17 | 腾讯科技（深圳）有限公司 | Vehicle driving state determination method and related device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113091866B (en) | Method and device for measuring loading quality of automobile in real time | |
| CN108694408B (en) | Driving behavior recognition method based on deep sparse filtering convolutional neural network | |
| CN111976389A (en) | Tire wear degree identification method and device | |
| CN111144485B (en) | Vehicle accident judging method and system based on xgboost classification algorithm | |
| CN112200231A (en) | Dangerous goods vehicle identification method, system, device and medium | |
| CN110986938A (en) | Bumpy road identification method and device and electronic equipment | |
| CN109543710B (en) | Model training method and related equipment | |
| CN110263836B (en) | Bad driving state identification method based on multi-feature convolutional neural network | |
| CN113514109A (en) | Unmanned aerial vehicle flight fault detection method and system | |
| Zhang et al. | Road surface defects detection based on imu sensor | |
| CN114863170A (en) | Deep learning-based new energy vehicle battery spontaneous combustion early warning method and device | |
| CN107187443A (en) | Vehicle unstability early warning system and method | |
| CN116090938B (en) | Method for identifying load state of rear loading vehicle | |
| CN116070985B (en) | Dangerous chemical vehicle loading and unloading process identification method | |
| CN116720045A (en) | Abnormal unloading behavior identification method and device for transport vehicle and transport vehicle | |
| CN113095197A (en) | Vehicle driving state identification method and device, electronic equipment and readable storage medium | |
| CN116664028B (en) | Cargo flow direction control method and device of transport vehicle and transport vehicle | |
| Yang et al. | An intelligent tyre system for road condition perception | |
| Hamad et al. | A new approach for varied speed weigh-in-motion vehicle based on smartphone inertial sensors | |
| CN114676479A (en) | Bridge damage identification method and calculation system under random vehicle load effect | |
| CN115238904A (en) | Method, apparatus and computer program for creating training data in a vehicle | |
| CN114964776A (en) | Wheel set bearing fault diagnosis method based on MSE and PSO-SVM | |
| Schmidt | Tire Pressure Monitoring using Weighted Horizontal Visibility Graphs | |
| CN113954850B (en) | Wheel center load estimation method and system based on edge calculation and soft measurement | |
| CN114078227B (en) | Intelligent AI recognition alarm system and method |
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 |