135 results about "Adaptive environment" patented technology

A method and apparatus is provided for constructing a predictive model for a system based on a priori qualitative modeling of the system and on historical database collected from past activity of the system or past events in the system. An expert provides grouping of parameters and qualitative dependencies between parameters and attributes, wherein some of the attributes may be conceptual or virtual attributes. The present invention extends existing methods of `evolutionary algorithms` in order to build successive sets of quantitative predictive models for the system, wherein parts of each model are evolved by the evolutionary algorithm and parts of each model are derived using the historical database. According to the present invention a model constructed by this method can be incorporated as a predictive model into a diagnosis or control apparatus without the need for human inspection, as the model complies with the expert's knowledge about the system. The present invention also provides a method to update the constructed model when new data is delivered, thus adjusting the model to changes in the environment.

The invention discloses a damping crawler type unmanned mobile platform. The damping crawler type unmanned mobile platform comprises a vehicle body, a module carrying platform body and two sets of crawler wheel mechanisms. The module carrying platform body is arranged at the center of the top face of the vehicle body, a module mounting position is reserved on the module carrying platform body, and different modules can be carried according to the task requirements, such as manipulators and batteries. The two sets of crawler wheel mechanisms are symmetrically arranged on the left side and the right side of the vehicle body. Five damping mechanisms are arranged below the inner portion of each crawler wheel mechanism, vibration of crawler wheels can be reduced, and accordingly stress applied on the crawlers is even. According to the damping crawler type unmanned mobile platform, the crawler travelling mechanisms are designed by adopting an integrated mode that motors are buried in the crawler wheels, the structure is compact, the space use ratio is high, the damping crawler type unmanned mobile platform can passively adapt to the environment by additionally arranging the damping mechanisms, it is ensured that the unmanned mobile platform can still travel stably and not bump on the ground surface with barriers, and the structure is simple, practical and reliable.

The invention discloses a task unloading intelligent decision-making method based on an unmanned aerial vehicle group in an edge computing environment. The method comprises the following steps that (1) environment information is acquired; (2) meta-learning is carried out, and if it is found that the environment of an edge server or a cloud center changes, initial parameters of the model are modified; (3) a retrieval mechanism and reinforcement learning are carried out, the retrieval mechanism is responsible for retrieving whether similar tasks exist before or not, and if yes, a decision resultis directly output; and if not, reinforcement learning is carried out, the reinforcement learning is responsible for training and judging the whole reinforcement learning system, two used modules arenetwork freezing and experience playback, and the action with the maximum value function after judgment is taken as a decision result to be output. According to the scheme, the meta-learning model isadopted to quickly adapt to the environment, and when the environment of the decision-making system is changed, the scheme can be quickly adjusted and a reasonable result can be quickly given. For the similar tasks of the unmanned aerial vehicle group, a memory function is introduced in the scheme, and rapid decisions can be made for the similar tasks.

The invention discloses an ultra-dense heterogeneous network small station coding cooperative caching method based on value function approximation. The method includes the following steps: adopting areinforcement learning method based on value function approximation, expressing a value function as a function of state and action, taking the operation of maximizing the number of file requests directly served by the average cumulative small stations as an optimization objective, continuously interacting with the environment to adapt to the dynamic changes of the environment, mining a potential file request transfer mode, obtaining an approximate expression of the value function, and further obtaining a cooperative caching decision that matches the file request transfer mode; and encoding thecooperative caching decision by a macro base station, and transmitting encoded cooperative caching results to each small station. The scheme of the invention formulates the caching decision by meansof the transfer mode of the file requests in a real network mined by reinforcement learning, does not need any assumptions about the prior distribution of data, and is more suitable for an actual system; and moreover, through the real-time interaction with the environment, the time-varying file popularity can be tracked, the corresponding caching strategy is made, the process is simple and feasible, and the operation of solving NP-hard problems is not required.

The invention discloses an air cooling data center energy-saving method based on a deep Q learning decision network. The method comprises the following steps: S1, acquiring environment states of a sampling data center and a monitoring data center; and S2, inquiring environment states of the sampling data center and the monitoring data center periodically, performing training of a deep Q learning decision network, and generating the temperature of the energy-saving air conditioner set value. The method has the beneficial effects that self-learning of the controller is carried out by adopting a reinforcement learning mode, the controller can adapt to environmental changes by setting rewards and feedback, exploration and learning mechanisms, and the environment of the data center is ensured to be in a safe operation interval while learning and optimizing the air conditioner setting value of the air cooling unit; therefore, empirical data with a poor estimation effect can be trained in future learning; and a new evaluation item is added to the environment state by applying the decision network, and the sensitivity of the controller to the environment change is improved.

The invention discloses an energy-saving greenhouse which comprises a greenhouse body and a control device. A solar photovoltaic panel provides energy for the greenhouse, a carbon dioxide sensor, a photosensitive sensor, a temperature and humidity sensor and a soil humidity sensor are mounted to provide information for the control device, the control device processes inputted signals, and output devices include a solenoid valve, a water pump, a carbon dioxide generator, a flow valve, a fan, a supplementary lighting lamp, a heat preservation lamp, a deinsectization lamp and a volume curtain machine. The internal environment of the greenhouse is adjusted by controlling the output devices, and growth of plants in the adaptive environment is ensured.

The invention relates to a multi-source data fusion method for a low-speed small target detection system, and belongs to the technical field of radar information detection. The multi-source data fusion method includes six implementation steps, aiming at low, slow and small target track characteristics, calculates the single-source track quality data in real time, distributes a fusion weight in a self-adaptive manner according to the track quality data, is high in adaptability to complex and changing environment backgrounds, can utilize multi-radar cooperative detection to the maximum extent, can bring the advantages of multi-sensor networking detection into full play, can effectively improve the detection precision and the detection stability of low, slow and small target tracks, can savenetworking resources, and can improve the environment adaptability of the multi-source information fusion method. The multi-source data fusion method solves the problems that a weighted average fusionmethod is adopted in an existing multi-sensor multi-source information fusion method, and initially gives a weight coefficient, so that the method is difficult to adapt to the influence of environmental interference in low, slow and small target detection, and the fusion result precision is not high due to unreasonable setting of the weight coefficient, and networking resources are wasted.

The invention discloses an automatic monitoring method and device, terminal equipment and a computer storage medium. The method comprises the following steps: making a service provided by a monitoringsystem into a Docker mirror image package and an application template, and deploying the monitoring system into a Kubernetes platform based on the Docker mirror image package and the application template; monitoring a new application deployed on the Kubernetes platform, and when the new application is provided with a metrics interface, registering the new application into Prometheus of a monitoring system so as to start a monitoring service for the new application; acquiring monitoring data acquired by Prometheus, and inputting the monitoring data into the trained machine learning model to obtain an alarm judgment result. According to the invention, cloud native and Kubernetes technologies are utilized, and a machine learning model is combined to realize automatic discovery and registration of monitoring services and automatic learning of monitoring thresholds, so that the method can adapt to environment automatic alarm; the work of operation and maintenance personnel is reduced, andthe change of application deployment is better adapted.

The invention discloses a multi-model target state prediction method based on a deep neural network and reinforcement learning, effectively solves the problems of complex strong nonlinear environmentexpression of multi-model fusion and long-time prediction of a target state, facilitates the improvement of the prediction precision and prediction duration of the target state, and achieves the online parameter self-correction of multi-model fusion. The calibration method provided by the invention has the advantages: 1) adapting to environmental scene changes and self-adjusting multi-model targetstate prediction parameters; 2) improving the target state prediction precision; 3) improving the target state prediction time length; 4) performing online self-learning the multi-model prediction parameters and improving the target state prediction precision in a dynamic scene.

The invention discloses a micro-bionic hexapod robot based on a 3D printing technology. The micro-bionic hexapod robot comprises a robot body bearing member, six three-degree-of-freedom foot mechanisms and a robot-mounted control circuit board, wherein the robot-mounted control circuit board is assembled inside the robot body bearing member, and the three-degree-of-freedom foot mechanisms are respectively assembled at the six top points of the robot body bearing member; each of the three-degree-of-freedom foot mechanisms comprises a hip bone member connected with the robot body bearing member, a thigh member, a shank member and a double-link mechanism linked with the shank member; and the robot body bearing member, the hip bone member, the thigh member, the shank member and the double-link mechanism are manufactured by the 3D printing technology. According to the micro-bionic hexapod robot disclosed by the invention, the 3D printing technology is utilized, so that various kinds of complicated robot body members can be manufactured, the convenience for structural modification of the robot is improved, and the micro-bionic hexapod robot has the advantages of being small in volume, light in weight, centralized in quality load, high in motion control precision, low in price, high in environment adaptability, and the like; and after the topology between a control circuit and a robot body is further optimized, the micro-bionic hexapod robot can be widely applied to the fields of teaching and entertainment.

The invention provides a 10KV oil-immersed transformer. The 10KV oil-immersed transformer comprises an oil tank, an iron core, windings, insulating sleeves and an oil temperature measuring device. Transformer oil for cooling and insulating the windings is arranged in the oil tank. A high-permeability crystal orientation cold rolling silicon sheet is adopted in the iron core. The iron core is a core-type iron core. The windings are each of a multilayer cylinder type structure. The oil tank is provided with a conservator for storing and replenishing oil for the oil tank. The conservator is a capsule type conservator. The oil temperature measuring device is arranged at the top of the oil tank. The oil tank is provided with a pressure relief valve. The side wall of the oil tank is a corrugated wall. Dust is sprayed to the surface of the oil tank, and a film is painted to the surface of the oil tank so that an outer protective layer can be formed. The 10KV oil-immersed transformer is low in cost, easy to maintain, good in heat dissipation, high in overload capacity and wide in adaptive environment. It is guaranteed that oil does not need to be replaced when the oil-immersed transformer normally runs, the maintenance cost of the transformer is greatly reduced, and meanwhile the service life of the transformer is prolonged.

The present invention discloses a deformable crawler-type running mechanism. It is formed from three modules, between modules a connecting arm can be used for making connection, every module is formed from connecting arm, module tilt motor, crawler drive motor, crawler wheel, module deflection motor and deflection angular gear. Said invention also provides their connection mode, and according to the background requirements said invented running mechanism can be formed into three basic forms of linear type, triangular type and parallel type.

The invention discloses a method and system for automatically generating a lane by a traffic millimeter-wave radar, a medium and equipment. The method comprises the following steps of: 1) partitioningthe detection range of the radar along the normal direction of the radar; 2) counting the positions and the number of vehicles passing through each block in a period of time according to a directionperpendicular to a normal, and selecting a clustering center in the corresponding block; 3) screening the clustering centers in different blocks to obtain clustering target points of each single lane;and 4) fitting the screened clustering target points of the single lane to obtain a lane center line, and generating a corresponding lane. The method has the advantages of simple operation, automaticlane generation according to radar information, reduction of radar deployment difficulty, enhancement of radar adaptation environment and the like.

The invention relates to the technical field of landscaping, and provides a greening method of high-water-level coast dredger fill. The method comprises the following steps of collecting dredger filland conducting airing to remove water, improving the dredger fill, laying an isolation pad layer, laying a drainage and salt-discharging pipe, laying a water spraying layer, laying improvement soil and planting greening plants. The greening method has the advantages that the isolation pad layer and the drainage and salt-discharging pipe are arranged on an original horizontal plane, a greening areais utilized to isolate the stacker as a whole, a microcirculation system is regionally controlled, an adaptive environment for plants is built, the problems of a traditional greening technology thatthe salt discharging and drainage effect is poor and the greenery landscape is maintained for a short time after high-water-level dredger fill is utilized are solved, and the plant survival rate can reach 95%; meanwhile, mechanized work is facilitated, the greening level can be obviously increased, and the greening range is expanded.

[Problem] An application is adapted to an environment and operated with high performance.

[Solution] An off-load pattern creation section 115 configured to perform a code conversion according to a deployment destination environment and a verification environment performance measurement section 119 configured to compile a code-converted application, deploy the compiled application to a verification machine 14, and execute processing for measuring a performance of the application in the event of performing off-loading to the verification machine 14, are repeated, and an actual environment performance measurement test execution section 123 configured to, after an execution file has been deployed, perform an automatic execution of performance test items extracted by an actual environment performance measurement test extraction section 122, by using an operation device, and a control section 11 configured to perform an environment adaptation process that executes one or more of code conversion step, resource amounts setting step, a deployment place selection step, a performance measurement step, and a performance measurement test step, are included.

The invention discloses a method for realizing intellectualization of an optical fiber preform deposition process. Manual adjustment in the deposition process causes large fluctuation of quality key parameters of the optical fiber preform, resulting in high rejection rate of the preform, and in order to realize optimal control of the deposition process, the invention provides an intelligent methodbased on a model predictive control framework. Firstly, a decision table is formed based on historical production operation records, factors influencing the quality of the optical fiber preform are mined, a neural network online quality prediction model is established, secondly, a formula proportion is adjusted based on a prediction result to achieve rolling optimization, and then the formula proportion in the deposition process is fed back and corrected according to a PK test result to achieve the purpose of stably controlling the quality of the optical fiber preform. Finally, a field operation result proves the effectiveness of the method. The intelligent method provided by the invention is simple to operate and high in environmental change adaptability, the quality of the preform is accurately predicted, and the maximization of enterprise benefits is facilitated.