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 relates to a multi-sensor low power consumption vehicle detection system and method based on an adaptive environment. A parking space node is installed in each parking space, when a vehicle enters or leaves the parking space, the geomagnetic field size of the parking space is changed obviously, a parking space node in a dormant state is woken, a parking space number and a parking space state are sent to a sink node, the sink node receives related data and read a Received Signal Strength Indication (RSSI) value, at the same time a microwave radar detector on the sink node of the area is started, the parking space which is calibrated in advance subjected to distance measurement again, and the state of the parking space at that time is determined when the vehicle enters or leaves. When the parking space information can not be detected by the RSSI value due to external factors such as a parking lot parking space environment or weather, a geomagnetic sensor is started as an auxiliary detection mode, and the reliability and accuracy of the method are improved.

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.

An interactive, educational system that provides an immersive and adaptive environment for a user to explore content through dimensional, time-sensitive, and emotion-sensitive inputs and outputs. The system is implemented on a networked device, such as a mobile tablet, and can contain a real-time, three-dimensional, content delivery engine that is capable of providing an augmented reality interaction model for the user. Therefore, a real world, physical movement that the user takes with the networked device will be reflected on the device. The system can use image recognition technology to recognize objects within the real world and display those objects on the interface. Further, the system can superimpose computer-generated images onto the virtual display of objects from the user's real world environment.

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.

A classifier is trained on a first set of examples, and the trained classifier is adapted to perform on a second set of examples. The classifier implements a parameterized labeling function. Initial training of the classifier optimizes the labeling function's parameters to minimize a cost function. The classifier and its parameters are provided to an environment in which it will operate, along with an approximation function that approximates the cost function using a compact representation of the first set of examples in place of the actual first set. A second set of examples is collected, and the parameters are modified to minimize a combined cost of labeling the first and second sets of examples. The part of the combined cost that represents the cost of the modified parameters applied to the first set is calculated using the approximation function.

The invention provides an automatic adjusting system for indoor air quality. The automatic adjusting system comprises a central control system, a temperature regulating system, a humidity regulating system, a self-adaptive environment regulation and control information acquisition system, a ventilation supply system and a sterilization system. The central control system is electrically connected with the temperature regulating system, the humidity regulating system, the self-adaptive environment regulation and control information acquisition system, the ventilation supply system and the sterilization system, and is used for controlling the temperature regulating system, the humidity regulating system, the self-adaptive environment regulation and control information acquisition system, theventilation supply system and the sterilization system to operate.

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.

A technique to provide well-balanced airflow in a computer data center that uses fan assisted tiles coupled to a host computer through a network. In an example embodiment, this is accomplished by disposing the fan tiles, including temperature and airflow sensors and a programmable switch to provide a Network address, in multiple locations in raised floor and ceiling of the computer data center. Further, the example embodiment includes coupling the fan tiles to the host computer through the network to control the fan tiles based on the feedback received from the temperature and airflow sensors to provide adaptive airflow balancing and thermal management in the computer data center.

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.

Disclosed is an epsilon-greed based online sequential perceiving and opportunity accessing method. In each time slot, users perceive channels sequentially and get access opportunistically to realize transmission. The method includes a step of initializing relative parameters and making access decisions based on online learning in each time slot, and has capabilities of learning environments actively and adapting to dynamic changes of environments. In addition, the method is an online decision implementing method, and the next decision is adjusted in real time according to each decision and feedback by systems, so that long-term accumulated throughput gain of the systems is maximized.

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.

Systems and methods provide adapted content to a visitor to a physical environment. An example method receives an image of a visitor to an environment. A visitor portion of the image is distinct from an environment portion of the image. The method detects one or more shapes in the visitor portion of the image using an automatic shape detection technique and defines an approximate boundary of the one or more shapes using a mask. The one or more shapes can be shapes of the visitor's clothing items. The method then calculates an attribute for an area of the image within the mask and identifies electronic content based on the attribute for the area of the image within the mask. The attribute can be a color attribute for the area such as a median color or a dominant color. The method provides the identified electronic content for display in the environment.

The invention discloses a target tracking and hunting method for an unmanned aerial vehicle group adaptive environment, and the method comprises the following steps: (1), building a multi-agent collaborative planning model through employing an MADDPG algorithm, and achieving the tracking and hunting of a target through the unmanned aerial vehicle group; and (2) when the unmanned aerial vehicle group approaches the threat area, performing autonomous adjustment and re-planning on the positions of the unmanned aerial vehicles by using a GA algorithm to avoid entering the threat area, wherein the survival rate of the unmanned aerial vehicles is improved, and meanwhile, a hunting task is completed. A layered hunting model is established and divided into two layers: a hunting layer and a multi-agent training layer. The unmanned aerial vehicle group interacts with the environment in real time, so that the current environment state can be obtained at any time. The surrounding layer determines whether to adjust the formation from the current state, and calculates a surrounding position distribution scheme. Aiming at the dynamic change of the environment and the task, the task execution success rate is improved in the relatively complex environment with the threat, meanwhile, the surrounding position is autonomously changed to avoid the threat area, and the risk of the unmanned aerial vehicle group is reduced.

A three-dimensional gait self-synchronizing and self-stabilizing method based on a control strategy presented by the invention mainly comprises a simple inverted pendulum model, defining virtual constraints, using a control strategy in periodic walking gait, stabilizing periodic walking gait through vertical oscillation, and extending the attributes to a real model of a humanoid robot. Firstly, virtual constraints are defined for the vertical position of a component object model in a simple inverted pendulum model; secondly, a control strategy for guiding self-synchronization is used in the constant-height periodic walking gait of the component object model; thirdly, the periodic walking gait is stabilized asymptotically through vertical oscillation of the component object model of the pendulum model; and finally, the attributes are extended to a real model of a humanoid robot. By using a self-synchronization control strategy, the walking speed, efficiency and adaptability to environment are improved. Moreover, the coordination is stronger, the stability is higher, and the method can be put into application more easily.

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 invention discloses an FOG feature and FAST algorithm-based eyes positioning method. The method comprises the following steps of: loading a trained SVM model file; obtaining the ith frame image to be positioned, and copying an image midImage; preprocessing the image midImage; carrying out facula detection by using a FAST algorithm, and obtaining a candidate area center point coordinate vector points; sequentially judging whether a candidate area is an eye area or not, and obtaining eye area center point vector pointsTru; screening points in the vector pointsTru so as to obtain an eye area center point vector, the final return vector of which is 2 for the most; if only one eye is detected, carrying out repair; and returning an eye area center point vector pointsFnl, traversing the vector pointsFnl, and intercepting a rectangular area which takes the pointsFnl as center to serve as an eye area. The method is capable of stably adapting to the environment change, and is relatively high in environment robustness and correctness and relatively low in false alarm rate.

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.

A participative kiosk configured to provide an interactive and adaptive environment is provided. The kiosk is comprised of a user interface, a user support, at least one biometric sensor, an audio sensor and an optical sensor. The kiosk includes a display and at least one audio speaker. The user support is positioned to provide a user with a view of the user interface. The biometric sensors are configured to measure real-time biometric data associated with the user. The auditory sensor receives vocal input commands from the user and the optical sensor detects movement of the user within the kiosk. The control system is configured to communicate with the user interface, the biometric sensors, the auditory sensor, and the optical sensor. The control system operable also is configured to control at least one environmental condition of the kiosk in response to an input from at least one of the biometric, auditory and optical sensors.

[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 relates to the field of WLAN (wireless local area network) wireless access, disclosing a device and a method for improving the frequency stability of a WIFI access point. The device for improving the frequency stability of the WIFI access point is composed of a temperature detection circuit, a voltage compensation circuit and a crystal oscillator, wherein the temperature detection circuit is connected with the voltage compensation circuit connected with the crystal oscillator. The defects of poor accuracy of the crystal temperature of the WLAN access point and weak fixed voltagecompensation adaptability is overcome. In the invention, the influence on the communication quality of the WLAN access point by environmental temperature is considered fully, communication quality byimproving access point frequency accuracy is improved and the orthogonality of OFDM (Orthogonal Frequency Division Multiplexing) subcarrier is ensured. Meanwhile, frequency accuracy is dynamically adjusted, so that equipment can adapt to environmental temperature variation and ensure the communication reliability of the WLAN access point.

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.

The invention provides an abnormal behavior detection and tracking method and device, a readable storage medium and equipment. The method comprises the steps of controlling a camera device to inspect a target region, and obtaining a current target picture shot by the camera device; inputting the current target picture into an abnormal behavior detection model for abnormal behavior detection, wherein the abnormal behavior detection model is updated according to environment change; and when the abnormal behavior detection model recognizes an abnormal behavior in the current target picture, tracking the abnormal behavior through a Kalman filtering algorithm. The abnormal behavior detection model can adapt to environment change through adaptive learning, so that the abnormal behavior can be effectively and accurately identified in various environments, and the abnormal behavior is ensured not to fail in tracking.

The invention relates to a flame retardant rubber windshield rubber part and a preparation method thereof; the rubber part consists of the following raw materials in parts by weight: 90-120 parts of natural rubber, 50-70 parts of butadiene rubber, 55-62 parts of recycled rubber, 32-45 parts of aromatic polyamide fibers, 15-20 parts of zinc oxide, 5-8 parts of paraffin, 35-43 parts of high wear resistant carbon black, 5-8 parts of a plasticizer, 6-9 parts of sulfur, 43-50 parts of a calcium powder, 3-6 parts of a vulcanizing agent, 1.5-3 parts of an accelerator CZ, 2-5 parts of an anti-aging agent, 3.5-6 parts of a composite flame retardant, and 2-4 parts of rubber oil. The rubber windshield rubber part has reasonable formula, and has good sound-insulation, flame-retardant and windproof performance; the aromatic polyamide fibers added in the raw materials increase the rubber windshield strength; a prepared rubber windshield for railway passenger and goods cars has the advantages of flame retardancy, safety, reliability, small wear, and tight contact, can prevent dust, rain and snow from entering the cars, and has the advantages of wide adaptive environment temperature and long service life.