Disclosure of Invention
In order to solve the technical problems, the invention aims to provide an AGV car pedestrian automatic following method based on artificial intelligence, and the adopted technical scheme is as follows:
acquiring a plurality of frames of first images; acquiring a target position, an AGV trolley position and a personnel position according to the multi-frame first image;
acquiring the following demand degree of the personnel, the distance relationship between a first task path when the AGV trolley position reaches the target position and a second task path when the personnel position reaches the target position, and the following degree of the AGV trolley.
Obtaining a final fusion path of the AGV according to the following demand degree, the distance relation and the AGV trolley following degree, so that the AGV trolley runs along the final fusion path;
and when the personnel are identified to carry the goods, selecting the AGV trolley reaching the position of the personnel.
The invention also provides a technical scheme of the AGV trolley automatic following system based on the artificial intelligence, which comprises a memory and a processor, wherein the processor executes a computer program stored in the memory so as to realize the AGV trolley automatic following method based on the artificial intelligence.
The AGV trolley automatic following method and system have the beneficial effects that: the method comprises the steps of firstly, acquiring information inside a factory, and then obtaining a first fusion path by obtaining the degree required by a person; obtaining a second fusion path by obtaining the distance relationship between the second task path and the first task path; and obtaining a final fusion path by acquiring the AGV trolley following degree, and finally selecting the AGV trolley to be followed. The embodiment of the invention solves the problem that the AGV follows the personnel blindly, and improves the working efficiency of the AGV and the experience degree of the personnel.
Further, the method for acquiring the personnel following demand degree comprises the following steps:
determining the movement track of the person according to the positions of the person in the multiple frames of first images, and acquiring the change degree of the movement direction and the movement speed of the person;
and acquiring the personnel following demand degree according to the change degree.
Further, the method for obtaining the final fusion path comprises the following steps:
acquiring a shortest path from the AGV trolley position to the personnel position; fusing by taking the following demand degree as the weight between the shortest path and the first task path to obtain a first fused path;
fusing by taking the distance relation as the weight between the shortest path and the first fused path to obtain a second fused path;
according to the following degree of the AGV trolley; and fusing by taking the following degree as the weight between the first task path and the second fusion path to obtain a final fusion path.
Further, the method for obtaining the AGV trolley following degree comprises the following steps:
acquiring the following inhibition amplitude of each AGV according to the following demand degree, the distance relation and the number of AGV trolleys reaching the personnel position;
and acquiring the following degree according to the following inhibition amplitude and the following demand degree.
Detailed Description
To further illustrate the technical means and effects of the present invention for achieving the predetermined objects, the following detailed description of the automatic following method for AGV according to the present invention with reference to the accompanying drawings and preferred embodiments will be made as follows. In the following description, different "one embodiment" or "another embodiment" refers to not necessarily the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following describes a specific scheme of an AGV automatic following method based on artificial intelligence in detail with reference to the accompanying drawings.
In actual warehouse, some goods can't directly be assigned to the AGV dolly and carry, thereby it realizes the adjustment to the goods position to need personnel to carry out patrolling and examining of goods and carrying, for example do not put good goods or, the unable discernment's of AGV dolly goods or other special goods, all need artificial transport, in order to save staff's physical power, the AGV dolly need follow the staff before people's haulage goods, when the staff discovers the goods that need the transport, place the goods on the dolly of following, let the AGV dolly accomplish the transport.
In order to solve the problem of blind following of AGV in the warehouse, the present embodiment utilizes the path fusion means, has reached the automatic intelligent following purpose of AGV dolly, and is specific as follows:
referring to FIG. 1, a flowchart illustrating steps of an AGV based on artificial intelligence method according to an embodiment of the present invention is shown, wherein the method includes the following steps:
s001, acquiring a plurality of frames of first images; and acquiring a target position, an AGV trolley position and a personnel position according to the multi-frame first image.
Specifically, a plurality of cameras are installed at the top of the warehouse, the visual fields of the cameras face downwards, the visual fields of all the cameras cover the whole warehouse, and the target position, the AGV trolley position, the personnel position and the goods position are obtained according to the obtained multi-frame images.
It should be noted that the AGV in this embodiment is an AGV within a radius of 30 meters from the position of the person, and the AGV can share the above data, and in this embodiment, the external convex polygons of all the goods positions are used as the goods placement area.
When it is detected that the person enters the placement area, step S002 is performed.
And S002, acquiring the following demand degree of the personnel, the distance relation between a first task path of the AGV trolley position reaching the target position and a second task path of the personnel position reaching the target position, and the following degree of the AGV trolley.
(1) And acquiring the following demand degree of the personnel.
In particular, in a warehouse a person is considered to be not stopped to handle goods if the person is traveling in a fixed direction and speed or is traveling at a high speed, i.e. there is no need to follow the person. When the speed of a person is decelerated to stop, uncertainty of obtaining the speed and the moving direction of the person is large, the fact that the person may have an intention of stopping is explained, and when the intention of carrying goods exists, the AGV trolley needs to have a tendency of driving towards the person, so that when the fact that the person carries the goods is detected, the requirement of the person is responded quickly, and the person is assisted in carrying the goods.
The intention of the personnel for carrying the goods is embodied by the following demand degree of the personnel, and the following demand degree of the personnel is obtained according to the motion direction sequence and the motion speed sequence in the preset time period before the current moment.
Specifically, the sum of the stability of the motion direction sequence and the speed magnitude sequence is obtained, the larger the value is, the smaller the speed direction change is, the more impossible the personnel can carry the goods, and the AGV trolley is not required to follow. When the value is smaller, the speed and the direction of the personnel are variable, whether the personnel need to carry the goods cannot be confirmed at the moment, and the AGV trolley needs to follow the situation. Specific personnel following demand degree w1The specific calculation method is as follows:
let the result of the normalization of the human motion direction sequence be p1 ═ { p11,...p1i,...p1NThe result of normalization of the velocity magnitude sequence is p2 ═ p21,...p2i,...p2NAnd N is the length of the sequence.
Wherein the content of the first and second substances,
represents the mean of the velocity sequence, P1 represents the degree of smoothness of the sequence P1, P2 represents the degree of smoothness of the sequence P2, wherein:
(2) and obtaining the distance relation between the first task path and the second task path.
Assuming that the second task path is G3, the path length of G1 is L1, and the sum of the path lengths of G3 and G2 is L2, a correction weight is calculated:
w2for the first task path and the second taskThe distance relationship of the path, the smaller the difference between L1 and L2 and the larger L1, the larger w2The larger.
(3) Obtaining the following degree of the AGV
Regard as the node with every AGV dolly, every two AGV dollies acquire an limit weight, and the limit weight is the negative exponential power of the difference of the length of the shortest path G2 of two AGV dollies, and the limit weight explains that the time difference that two AGV dollies reachd personnel department is more big the less, acquires the degree of every AGV dolly, and the degree of AGV dolly is rather than the limit weight sum of adjacent all AGV dollies, and is concrete:
the method comprises the following steps of (1) setting the degree of the AGV car as d, setting the node size of the AGV car as q, and specifically calculating the following inhibition amplitude of the AGV car as follows:
wherein the node size q ═ w2(w1-1)+1]And d is the AGV trolley degree.
The more AGV carts arrive together, the smaller the node size, explain biThe larger the value, the greater the magnitude of suppression of tracking of the AGV.
The specific calculation method of the AGV trolley following degree comprises the following steps:
w3=αexp(-bi)w1
wherein, w3Degree of following of AGV Car, biFor the follow-up suppression amplitude, w, of this AGV Car1For the following demand of the staff, α is a hyperparameter, and α is made to be 0.8.
And S003, obtaining a final fusion path of the AGV according to the following demand degree, the distance relation and the AGV trolley following degree so as to enable the AGV trolley to travel along the final fusion path.
(1) A first fused path is obtained.
Specifically, when detecting personnel's transport goods, the AGV dolly arrives personnel department and carries out the transport of goods. When the person is not carrying goods, but has the possibility of carrying goods, which is expressed by the degree of the person following requirement, the task of the AGV shall be a combination of the following task and the own carrying task, specifically:
setting the AGV small parking space as Q1, the target position as Q2, the personnel position as Q3, the first task path from the AGV to the target position Q2 as G1, and the shortest path from the AGV to the personnel position as G2 according to the following demand degree w of the personnel1The weight as the shortest path G2 is fused with the first task path G1 to obtain a first fused path G4.
(2) A second fused path is obtained.
When w is2When the weight is larger than the preset weight threshold value, the weight is calculated by w2The weight as the shortest path G2 is fused with the first fused path G4 to obtain the second fused path G5 when w is2And when the weight is less than or equal to the preset weight threshold value, taking the first task path G1 as a second fusion path G5.
(3) And obtaining a final fusion path.
Obtaining K AGV trolleys with the largest following inhibition amplitude of the AGV trolleys, calculating the following degree of each AGV trolley in the K AGV trolleys, and aiming at the K AGV trolleys, calculating the following degree w3The weight as the first task path weight G1 is fused with the second fused path G5 to obtain the final fused path G6.
And step S004, when the fact that the personnel carry the goods is identified, selecting the AGV which reaches the position of the personnel.
The AGV trolley needs to travel along a final fusion path G6 in the period between the current moment and the next moment, the AGV trolley with the largest node size q follows the personnel along the shortest path G2 when the DNN network detects that the personnel have the action of carrying goods, and other AGV trolleys give up following to complete original tasks, so that the purpose of carrying goods by auxiliary personnel can be guaranteed, and the carrying tasks of other AGV trolleys are not influenced.
It should be noted that the DNN network of this embodiment is composed of an Opense network and a TCN network.
Preferably, the specific method for merging the shortest path G2 with the first task path G1 is as follows:
obtaining the shortest path g1 of the two paths, and cutting a path g2 with the same length as g1 to g1 and g1 on a longer path2 are the same length. And the same starting point is provided, the travel distance x of the AGV in unit time is obtained, one point is collected on the g1 every path with the length x, and finally, one z1 is obtained as { z1 ═1,...,z1k,...z1NThe sequence, similarly from g2, can also be obtained a sequence z2 ═ z21,…,z2k,...z2NN is the length of the sequence, the lengths of the z1 sequence and the z2 sequence are consistent, and the specific method for fusing the shortest path G2 with the first task path G1 is as follows:
let the k-th point in the sequence have a two-dimensional coordinate z1kAnd z2kObtaining a point zk,
Let zk=w1z2k+(1-w1z1k)
Wherein, w1The following demand degree of personnel.
Obtaining a position for each point in the sequence, and finally obtaining a sequence z ═ { z ═ z1,z2,...,zk,., connecting all points in sequence z, the path formed is w1As a result of weight fusion of the shortest path G2 with the first task path G1.
It should be noted that the path fusion method of the second fused path G5 and the path fusion method of the final fused path G6 obtained in the present embodiment are the same as the fusion method of the shortest path G2 and the first task path G1 described above.
In summary, in the embodiment, the internal information of the plant is collected first, and then the first fusion path is obtained by obtaining the degree required by the person; obtaining a second fusion path by obtaining the distance relationship between the second task path and the first task path; and obtaining a final fusion path by acquiring the AGV trolley following degree, and finally selecting the AGV trolley to be followed. The embodiment of the invention solves the problem that the AGV follows the personnel blindly, and improves the working efficiency of the AGV and the experience degree of the personnel.
Transformer station site selection system embodiment based on artificial intelligence and big data
The AGV automatic following system based on artificial intelligence of this embodiment includes a memory and a processor, and the processor executes the computer program stored in the memory to implement the AGV automatic following method based on artificial intelligence as described in the AGV automatic following method based on artificial intelligence.
Because the AGV automatic following method based on artificial intelligence has been described in the embodiments of the AGV automatic following method based on artificial intelligence, the description thereof is omitted here.
It should be noted that: the precedence order of the above embodiments of the present invention is only for description, and does not represent the merits of the embodiments. And specific embodiments thereof have been described above. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments.
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, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.