Summary of the invention
The object of the invention is to overcome the deficiencies in the prior art, a kind of warehousing system based on the cloud framework unified platform is provided, solves that traditional warehousing system storage space density is low, goods distribution can not self-adaptative adjustment, resource can not be shared, maintenance cost is high, inefficient problem.
The present invention solves its technical matters and takes following technical scheme to realize:
Based on a warehousing system for the cloud framework unified platform, comprise be arranged on warehouse district elasticity shelf, travel on the fork legged mobile robot fork truck that warehouse district transports goods and the warehouse control system based on the cloud framework unified platform that fork legged mobile robot fork truck is controlled;
Described fork legged mobile robot fork truck by mobile robot, be installed on the same side with mobile robot and can automatic lifting pallet fork, form for being connected mobile robot and pallet fork and supporting the pallet fork bracing frame that pallet fork moves up and down, in mobile robot, control device for moving robot be installed and be connected with the warehouse control system based on the cloud framework unified platform;
Described elasticity shelf are made up of shelf supports frame, the shelf tray guide supporting crossbeam being arranged on shelf supports frame both sides, shelf tray guide rail groove and pallet;
The described warehouse control system based on the cloud framework unified platform comprises execution level, service layer, cloud service layer and client layer, described execution level, service layer, is connected by wireless network between cloud service layer and client layer.
And described execution level comprises robot navigation's module, pallet fork lifting module, dynamic goods yard module, trans-regional multirobot ontology registration module and sensor data acquisition module for pitching legged mobile robot forklift-walking path; The lifting of robot navigation's module wherein, pallet fork module, trans-regional multirobot ontology registration module and sensor data acquisition module installation are in control device for moving robot;
Described robot navigation's module for pitching legged mobile robot forklift-walking path also comprises the guidance path in the ground running space of each shelf first floor;
Described pallet fork lifting module receives the instruction from service layer, automatically identify target goods layer, and automatic lifting is to this goods layer discharging of goods;
Described dynamic goods yard module is made up of goods, pallet and fork legged mobile robot fork truck; Fork legged mobile robot forklift-walking path comprises the guidance path in the ground running space of each shelf first floor, fork legged mobile robot fork truck receives the order from service layer, according to the needs of self-adaptative adjustment, same warehoused cargo is adjusted to the diverse location of storage according to different demand;
Described multirobot ontology registration module comprises the data of vehicle dimension, load carrying ability to server registration.
And described service layer comprises WEBSERVICE interface, map structuring and more new module, multirobot colony operational module, robot path planning's module, pallet fork elevating control module, self adaptation are stored in a warehouse goods yard adjusting module and robotary data module;
The map structuring of described service layer and more new module, multirobot colony operational module, robot path planning's module, pallet fork elevating control module, self adaptation are stored in a warehouse goods yard adjusting module, and the data of these modules all export from the cloud computing of cloud service layer.
And described cloud service layer comprises cloud interactive module, cloud computing module and cloud memory module.
Described cloud computing module carries out the degree of depth study of dispatching algorithm, and this dispatching algorithm comprises map structuring and more new algorithm, multirobot colony operative algorithm, Robot Path Planning Algorithm, pallet fork elevating control algorithm, self adaptation and to store in a warehouse goods yard adjustment algorithm
And described client layer comprises user operation terminal and emulates client.
And the width of described mobile robot is less than the width of shelf supports frame internal diameter, the length of mobile robot is less than the longitudinal length of elasticity shelf; The height height be less than between elasticity shelf ground floor and ground of mobile robot deducts the later height of fork height
Advantage of the present invention and good effect are:
The present invention has overturned traditional storage operation mode by fork legged mobile robot fork truck, elasticity shelf and the warehouse control system technology based on the cloud framework unified platform, realize raising efficiency, on-demand customization, be easy to safeguard three zones, for user brings brand-new logistic storage management to experience: fork legged mobile robot its pallet fork of fork truck and car body are positioned at the same side, bodywork length size saves 1/2nd, adds storage density; Elasticity shelf can the size of dynamic conditioning cargo storage as required, provides the foundation and guarantee for realizing the dynamic goods yard of storage; Store in a warehouse dynamic goods yard according to the change in season and order demand curve, the deposit position of self adaptation dynamic conditioning goods; Based on the warehouse control system of the cloud framework unified platform, present the new model of high in the clouds storage scheduling: move to high in the clouds by the complicated algorithms such as mass data process, analysis, map structuring and renewal, multirobot group collaboration, robot path planning, large data depth study, increase system efficiency and performance, lightweight robot body designs and reduces robot body Based Intelligent Control cost, simultaneously, sharing resources and United Dispatching, be convenient to information sharing.
Detailed description of the invention
Below in conjunction with accompanying drawing, the embodiment of the present invention is further described:
Based on a warehousing system for the cloud framework unified platform, comprise be arranged on warehouse district elasticity shelf, travel on the fork legged mobile robot fork truck that warehouse district transports goods and the warehouse control system based on the cloud framework unified platform that fork legged mobile robot fork truck is controlled.Below the various piece in system is described respectively.
As shown in Figure 1, described fork legged mobile robot fork truck by mobile robot 1-1, be installed on the same side with mobile robot and can automatic lifting pallet fork 1-2, form for being connected mobile robot and pallet fork and supporting the pallet fork bracing frame 1-4 that pallet fork moves up and down; Pallet fork bracing frame 1-4 is also provided with limit switch 1-4, and limit switch is used for the auxiliary lifting position controlling pallet fork.In mobile robot, control device for moving robot is installed and the controlling functions realized pitching legged mobile robot fork truck that is connected with the warehouse control system based on the cloud framework unified platform.The width of mobile robot is slightly less than the width of shelf supports frame internal diameter, and the length of mobile robot should be less than the longitudinal length of elasticity shelf; The height height that should be less than between elasticity shelf ground floor and ground of mobile robot deducts the later height of fork height.
As shown in Figures 2 and 3, described elasticity shelf are made up of shelf supports frame 2-1, the shelf tray guide supporting crossbeam 2-2, the shelf tray guide rail groove 2-3 and pallet 2-4 that are arranged on shelf supports frame both sides; Described shelf tray guide supporting crossbeam 2-2 is packed in the both sides of shelf supports frame 2-1 according to uniform intervals, described pallet 2-4 is erected on the shelf tray guide rail groove 2-3 of shelf both sides.The height of described elasticity shelf goods can carry out flexible adjustment according to actual needs, and not by the restriction of shelf every layer standing height, Fig. 2 shows because the height of goods is beyond the height of fixing goods layer, and by shelf the 2nd layer, the 3rd laminated and example used.
Pitch legged mobile robot fork truck picking process as shown in Figures 4 to 6, picking step 1: before picking, pallet fork receives the instruction of service layer, and automatic lifting is to the height of target goods layer, and it is ground floor that Fig. 4 shows target goods layer, and now pallet fork is down to the height of ground floor; Picking step 2, after pallet fork aims at the mark goods layer, pallet fork puts in shelf together with mobile robot's car body; Picking step 3, after pallet fork and mobile robot's car body put in shelf, the pallet of shelf the 1st layer and goods are slightly lifted under the instruction of service layer, the height that goods and pallet are lifted is no more than the height of the restriction shown in Fig. 6; Picking step 4, pulls out shelf by goods, so far, completes picking process after goods lifts by fork legged mobile robot fork truck.
As shown in Figure 7, the warehouse control system based on the cloud framework unified platform comprises execution level, service layer, cloud service layer, client layer, carries out intercommunication between execution level, service layer, cloud service layer, client layer by wireless network.
Described execution level comprises robot navigation's module, pallet fork lifting module, dynamic goods yard module, trans-regional multirobot ontology registration module, sensor data acquisition module for pitching legged mobile robot forklift-walking path.The lifting of robot navigation's module wherein, pallet fork module, trans-regional multirobot ontology registration module and sensor data acquisition module installation are in control device for moving robot;
Described robot navigation's module for pitching legged mobile robot forklift-walking path also comprises the guidance path in the ground running space of each shelf first floor;
Described pallet fork lifting module receives the instruction from service layer, automatically identify target goods layer, and automatic lifting is to this goods layer discharging of goods;
Described dynamic goods yard module is made up of goods, pallet and fork legged mobile robot fork truck; Fork legged mobile robot forklift-walking path comprises the guidance path in the ground running space of each shelf first floor, fork legged mobile robot fork truck is according to the order of service layer, according to the needs of self-adaptative adjustment, same warehoused cargo is adjusted to as required the diverse location of storage.
Described multirobot ontology registration module comprises the data of vehicle dimension, load carrying ability to server registration.
Described service layer comprises WEBSERVICE interface, map structuring and more new module, multirobot colony operational module, robot path planning's module, pallet fork elevating control module, self adaptation are stored in a warehouse goods yard adjusting module, robotary data module.
Reception and the sending function of data only do in described service layer, the map structuring of service layer and more new module, multirobot colony operational module, robot path planning's module, pallet fork elevating control module, self adaptation are stored in a warehouse goods yard adjusting module, and the data of these modules all export from the cloud computing of cloud service layer.
Described cloud computing module carries out the degree of depth study of dispatching algorithm, and this dispatching algorithm comprises map structuring and more new algorithm, multirobot colony operative algorithm, Robot Path Planning Algorithm, pallet fork elevating control algorithm, self adaptation and to store in a warehouse goods yard adjustment algorithm.
Described client layer comprises user operation terminal and emulates client operates controlling functions for realizing.
Dispatching algorithm in cloud computing module realizes degree of depth learning functionality, as shown in Fig. 8 to Figure 10, comprises the following steps:
(1) Study of Scheduling object is determined: Robot Path Planning Algorithm, multirobot colony operative algorithm, self adaptation storage goods yard adjustment algorithm;
(2) execution result value is determined: execution result embodies with the form of value, as path length short value, collision frequency value, self adaptation storage distribution law value;
(3) above-mentioned execution result is put into large database concept and as the input end of neural network;
(4) data of neural network to input are trained and export optimal value; The method that the data of neural network to input are trained is:
1. set up input layer, hidden layer, output layer, every one deck is made up of several nodes;
Set up the node of neural network input layer: the relation that should have restriction mutually between neural network input layer node.As shown in Figure 10, input end is that robot path planning's execution result, multirobot colony operation execution result, self adaptation storage goods yard adjustment execution result are as input end, their mutual restricting relations are: the number of times that path length then may collide is just many, self adaptation storage goods yard distribution law height then may occur that traffic cost is high and route is long, therefore, the relation will coordinating between them by neural network, by neural network training each time, continuous adjustment and corrected parameter, make result approach idealized;
Set up the hidden layer node of neural network: the effect of hidden layer adopts to regulate the method for weights extract from input end and store inherent rule, each hidden node has several weights, and each weights strengthen network mapping ability, a parameter of Approaching Results value, the number of plies of hidden layer and the number of every node layer will be determined by rule of thumb, too much all can bring drawback with how many, nodes very little, network each learning time is shorter, network ability of obtaining information from sample is poor, be not enough to summarize and embody sample rule, thus identify new samples difficulty, poor fault tolerance, node in hidden layer is too many, the learning time of network lengthens ...., the present embodiment designs the network of three layers, hidden layer only has 1 layer, can determine as required in practice.
Set up output layer: neural network output valve is end value, multirobot colony operating result value, the self adaptation storage goods yard distribution law end value of robot path planning;
2. using the input value of each execution result (execution result of dispatching algorithm) as neural network;
3. by constantly adjusting interconnective relation between the inner great deal of nodes of neural network thus exporting optimal value; The interconnective relation of described adjustment comprises the connection weights between knot modification;
4. neural network exports and optimizes later result and feed back to dispatching algorithm;
(5) the training optimal value that neural network exported of dispatching algorithm is as the initial value of dispatching algorithm next time, repeats above-mentioned steps until approach the expected value of dispatching algorithm.
In this step, dispatching algorithm will be updated to new dispatching algorithm according to the value of neural network feedback, and new dispatching algorithm is reentried new end value, and new end value, as the input value of neural metwork training next time, moves in circles, according to this until approach expected value.
Dispatching algorithm compares input value (when hidden layer only has 1 layer according to the end value that neural network exports, comparative figure is from input layer, when hidden layer is multilayer, comparative figure is the value of last layer hidden layer), if be enhance relatively, then adjust the parameter of neural network further, to obtain strengthening again next time, if is reduce relatively, then revise the parameter of neural network, amendment is compared later again, thus improves constantly and constantly adjust parameter.
It is emphasized that; embodiment of the present invention is illustrative; instead of it is determinate; therefore the present invention includes the embodiment be not limited to described in detailed description of the invention; every other embodiments drawn by those skilled in the art's technical scheme according to the present invention, belong to the scope of protection of the invention equally.