61 results about "Distributed model predictive control" patented technology

The invention belongs to the technical field of a distributed control system and relates, specifically, to a multi-agent cluster coordination method and a multi-UAV cluster coordination system. The multi-agent cluster coordination method of the invention allows agents to exchange information with neighbors within their respective interaction ranges and to form a cluster in a self-organized manner by solving for the control input through distributed mode predictive control. The method further allows the presence of a leader agent in the system to lead the overall motion trajectory of the agent cluster. The multi-UAV cluster coordination system provided by the invention enables multiple UAVs to autonomously achieve cluster flight through data exchange thereamong, enhances the efficiency of executing the tasks that need the coverage of a large area, such as aerial photography, monitoring, search and rescue, and the like, and has high adaptivity and redundancy.

The present invention relates to a cooperative control method of a multi-unmanned aerial vehicle formation based on model predictive control. The method comprises the following steps of I, initializing task requirements, related control parameters and the like; II, conducting preliminary route planning for a pilot unmanned aerial vehicle; III, detecting the condition of an environment of a flightarea in real time through a sensor, judging and selecting a proper flight formation, and calculating (updating) a virtual formation guidance point; IV, calculating a reference value according to the calculated virtual formation guidance point by using the guidance point as a cost, calculating a cost function, and conducting flight control by adopting a particle swarm optimization strategy based ondistributed model predictive control; and V, repeating the step II, the step III and the step IV, and controlling the multi-unmanned aerial vehicle formation to conduct cooperative flight until a target position is reached. The method effectively solves the problem of cooperative control over the formation in a complex environment, and enables a multi-unmanned aerial vehicle system to have stableformation retention capability and efficient formation transformation capability during flight.

The invention provides a multi-region cooperative joint frequency modulation control method based on a dual-layer model predictive structure. The method realizes the steady-state power optimized application of cross-regional AGC units under the multi-constraint condition on the basis of the economic model predictive control in the upper layer; and the dynamic frequency optimization control of themulti-region AGC units on the basis of the distributed model predictive control is realized in the lower layer. On the slow time scale (minutes), the steady-state power optimized allocation of the cross-regional AGC units considering the safety constraints of tie line sections is realized. On the short time scale (seconds), the dynamic frequency optimization control of the AGC units in each regionis realized.

The invention discloses a frequency control method and system for an interconnected power system containing energy storage resources. The control method comprises the following steps: establishing a frequency response model of a regional interconnected power system containing energy storage frequency modulation resources; performing frequency division processing on the ACE signal to obtain a powerreference value borne by a traditional unit and an energy storage resource; establishing a space state model and discretizing the space state model; designing a distributed model prediction controller, and establishing an optimization objective function and constraint conditions; and solving a quadratic programming problem formed by the optimal objective function and the constraint condition to determine an optimal control sequence, and enabling a first control variable in the control sequence to act on a traditional unit and energy storage resources to realize optimal control. The control strategy provided by the invention can reasonably allocate the frequency modulation responsibilities of a traditional unit and the energy storage resources, gives play to the advantages of quick outputof energy storage and large frequency modulation capacity of the unit, and can quickly stabilize the frequency fluctuation caused by load and new energy injection in a power system.

The invention relates to an optimized formation tracking control method based on distributed model predictive control and belongs to the field of motion control of underactuated multi-surface unmannedships. The method comprises the following steps: 1) establishing an underactuated unmanned ship motion model and a tracking error model; 2) establishing a performance index of a master unmanned ship,proposing a model predictive tracking control algorithm, and calculating an optimal input at the current time according to the performance index thereof; 3) based on the acquired neighbor node information, establishing the performance index of each unmanned ship, respectively, proposing a distributed model predictive control algorithm and calculating an optimized input of the current time according to the performance index; 4) updating the predictive information and continuously and iteratively optimizing the formation of the whole unmanned ship so that a certain formation is kept between theunmanned ships so as to track the target unmanned ship.

The invention discloses an optimized mowing type formation control method for an unmanned ship guided underwater vehicle group, which is used for solving a technical problem of poor practicability ofthe existing multi-ocean vehicle formation control method. The technical scheme is that a main vehicle is determined in underwater vehicles, the unmanned ship is enabled to communicate with the main vehicle through optical cable connection, and the main vehicle performs underwater acoustic communication with the other underwater vehicles according to the underwater acoustic communication distance.An error model of the unmanned ship and a reference target is constructed, and performance indexes are designed by using a model prediction control method so as to realize a problem of trajectory tracking optimization for the unmanned ship; and the underwater vehicles are divided into a main vehicle and slave vehicle, the main vehicle is based on state information of the unmanned ship and the slave vehicles, each slave vehicle is based on state information of the remaining underwater vehicles, and constrained formation optimization control is realized by using the distributed model predictioncontrol method. due to the adoption of the distributed model prediction control method, the problem of constrained formation optimization for the multiple ocean vehicles is solved, and the practicability is good.

The invention discloses a series-structure light-storing-type multi-micro-grid economical operation system and a method. The series-structure light-storing-type multi-micro-grid economical operation system comprises a primary system and a control system. The primary system is formed in the mode that two sub-micro-grids are connected in series through a grid-connection and off-grid switch K1, and the primary system comprises a photovoltaic power generation array, a storage battery and local loads. The control system is formed by sub-micro-grid controllers, photovoltaic controllers, energy storing controllers, load controllers and a communication network. According to the series-structure light-storing-type multi-micro-grid economical optimization method, a distributed model prediction control algorithm considering a peak-valley electricity price is adopted, and according to the source load characteristic and the operating characteristic in the multiple micro-grids, a micro-grid prediction model and an optimized objective function are established. The established series-structure light-storing-type multi-micro-grid economical operation system and the method have the important engineering application value in developing researching of planning, design, mode switching and optimizing operation of the light-storing-type multi-micro-grid system.

The invention discloses a comprehensive energy system multi-time scale optimization scheduling method and system, and a storage medium. The method comprises the following steps: building an energy storage side model, an energy conversion side model, an energy storage side model and a load side model according to a comprehensive energy system architecture; according to renewable energy and load day-ahead prediction information, gas purchase price and time-of-use electricity price, establishing a day-ahead optimization model by taking the lowest daily operation cost of the comprehensive energy system as a target; according to the intra-day renewable energy and the load day-ahead prediction information, establishing an intra-day optimal scheduling model with the minimum system operation costand the minimum start-stop punishment cost of all the units as targets; and according to the real-time prediction information, solving by adopting a distributed model prediction control algorithm withthe minimum total adjustment amount of system controllable equipment at the next moment of an intra-day scheduling plan as a target to obtain a final scheduling plan. According to the method, the distributed model prediction control algorithm is adopted to solve the scheduling model in the implementation stage, the complexity of the model can be reduced, the difficulty of online solving is reduced, and meanwhile the control performance of the system can be considered.

The invention discloses a wind power cluster integrated frequency control method based on hierarchical distributed model predictive control, which comprises the following steps of: establishing a hierarchical distributed model predictive control H-DMPC controller; The multi-time scale coordinated variable granularity frequency control framework of wind power cluster is established in the tertiaryfrequency modulation layer, the secondary frequency modulation layer and the primary frequency modulation layer. The rolling optimization models are established on the spatial scale in the tertiary frequency modulation layer, the secondary frequency modulation layer and the primary frequency modulation layer respectively. In the aspect of wind power prediction error correction, the feedback correction link is established to realize the closed-loop operation of integrated frequency control through the feedback correction of grid operation state and prediction data. The invention establishes a feedback correction link according to the real-time operation state of the system, realizes the closed-loop operation of the integrated frequency control strategy, and the proposed control strategy caneffectively realize the wind power cluster to participate in the system frequency modulation, and can consider both the economy and the safety objectives of the system.

The invention relates to a multi-ocean-robot cooperative circular scanning method based on distributed model predictive control, and belongs to the field of multi-ocean-robot control. The method is mainly used for multi-ocean-robot cooperative circular scanning, and is characterized by firstly achieving the following of real targets to respective trajectories by using the real targets to track moving virtual targets on the trajectories; and then in consideration of a relationship between the travel angles between virtual targets in a cooperative circular scanning process, introducing a cooperative performance index, and online optimizing an optimal control sequence by using a distributed model predictive control algorithm. The method, by using the distributed model predictive control algorithm, greatly reduces the communication pressure between systems, and solves two problems including path following and cooperative control by using one controller, thereby reducing the design difficulty of a control system.

The invention proposes a chaos grey wolf optimization-based unmanned aerial vehicle formation control method. A distributed model predictive control (MPC) frame of unmanned aerial vehicle formation control is built, each unmanned aerial vehicle only shares information with a neighbor, and the communication requirement and the computational complexity are reduced; a chaos optimization algorithm iscombined with a grey wolf optimization algorithm, and thus, the algorithm performance is improved; and the algorithm is combined with the distributed MPC, finite horizon optimal control problem (FHOCP) is solved, so that unmanned aerial vehicle formation control is achieved.

The invention provides a heavy-load locomotive adhesion control method based on distributed model predictive control, which comprises the following steps: acquiring operating parameters in the locomotive operating process, and establishing a locomotive kinetic model according to the operating parameters; constructing a creep rate calculation model according to the relationship among the creep rateof the locomotive, the angular speed of wheels and the speed of the locomotive; a distributed state space model is obtained by combining the locomotive dynamic model, the creep rate calculation modeland the vehicle rail model; acquiring real-time traction torque of a corresponding shaft through an MPC adhesion controller arranged at each shaft of the locomotive, and calculating to obtain a predicted traction torque value and a predicted creep rate value through a distributed state space model and a preset cost function according to the real-time traction torque; and the predicted creep ratevalue is compared with a preset creep rate value, and when the predicted creep rate value is equal to the preset creep rate value, the traction torque of each shaft is adjusted according to the predicted traction torque value.

The invention relates to a design method of a distributed model prediction controller of an automatic power generation control system. The design method comprises the following steps: building simulation models of the automatic power generation control system of a regional interconnection power system, wherein the simulation models comprise a wind power generation model and a photovoltaic power generation model; according to the simulation models of the automatic power generation control system, designing a distributed model prediction control algorithm and establishing and optimizing a targetfunction; determining a constraint condition according to the actual operation condition of the automatic power generation control system, and processing the constraint; and according to the target function and the constraint condition, optimizing by adopting a guided firework algorithm to obtain controller output. According to the design method, the dynamic performance of the interconnected power grid automatic power generation control system is effectively improved, and the safe and stable operation of a power system is ensured.

The invention discloses a hybrid vehicle queue distributed model prediction and control method. The method comprises the following steps of firstly, dividing a hybrid vehicle queue into a plurality of mutually associated local hybrid vehicle queues; then establishing a model of each local hybrid queue based on a car-following model of networked automatic driving vehicles and human-drive vehicles; secondly, based on the local hybrid queue model, according to a model prediction control algorithm, establishing an overall control problem of the hybrid vehicle queue; and finally, in combination with an alternating direction multiplier method, constructing a distributed model prediction controller for each local hybrid queue to realize the overall formation control of the hybrid vehicle queue. The effectiveness of the method provided by the invention is verified by performing an experiment on an MATLAB simulation platform. Experimental results show that the method can ensure that the networked automatic driving vehicles and the human-driven vehicles in the hybrid vehicle queue achieve a good formation effect, and the networked automatic driving vehicles can be driven in a formation manner at a smaller inter-vehicle distance, so that the traffic capacity is effectively improved.

The present invention discloses a distributed formation control system and method in a disturbance three-dimensional environment. The method combines model predictive control with extended Kalman filter, realizes the formation control over multiple controlled objects in the case that great uncertainty exists in the modeling of the controlled objects and the output measurement of the controlled objects has noise, integrates a swarm intelligence optimization algorithm library to realize the formation control over the multiple controlled objects in the case that the controlled objects have complex and coupled three-dimensional space dynamic characteristics and are limited by complex constraints. In addition, the method is achieved by distributed model predictive control so as to further reduce the computational cost of the control law solution and enhance the anti-strike capability of the formation control.

The invention discloses an unmanned aerial vehicle cluster collaborative flight path planning method considering communication time delay, and belongs to the technical field of flight path planning. Under the condition of considering communication time delay, a'global planning-local obstacle avoidance 'cooperative track planning framework is constructed, in the global planning stage, probability distribution of adjacent unmanned aerial vehicle positions caused by random communication time delay is considered, the probability distribution situation of the adjacent unmanned aerial vehicle positions caused by the communication time delay is combined into a sparse A * search algorithm, and the probability distribution situation of the adjacent unmanned aerial vehicle positions caused by the random communication time delay is combined into the sparse A * search algorithm. A priority decoupling mechanism is introduced to improve the algorithm search efficiency; in the local collision avoidance stage, the local collision risk still exists in the global flight path due to the uncertainty of communication time delay and the uncertainty of the position of an adjacent vehicle is considered, the optimal control problem of distributed model predictive control is solved, and flight path tracking and local adjustment of the unmanned aerial vehicle in the pipeline are achieved. Based on double-level cooperative flight path planning, the timeliness and reliability of unmanned aerial vehicle cluster cooperative flight path planning under communication time delay are ensured, and the cooperative capability of an unmanned aerial vehicle cluster in a real scene is improved.

The invention discloses a neutral buoyancy robot attitude and trajectory control method based on distributed model predictive control and belongs to the field of microgravity robot control. The methodmakes a study for the coupling problem and optimization problem of a plurality of neutral buoyancy robots having different sampling periods, and simulates a unified sampling period for the neutral buoyancy robots having different sampling periods through the Delta operator theory at the controller side; and when one or more neutral buoyancy robots need to update control quantity, states of otherneutral buoyancy robots can be obtained through a predictive mode, so that the states of other neutral buoyancy robots can be taken into account in constraints to solve their control quantity, and thewhole system is allowed to ensure that coupling probability constraint is met nonconservatively, and ensure that the system performance is good. The method combines the characteristics of computer control and converts the complex on-line optimization problem into quadratic programming, thereby greatly reducing calculation amount; and the method is suitable for engineering application.

The invention discloses a traditional Chinese medicine preparation process operation optimization method based on distributed model predictive control, and the method comprises the steps: building an operation control process model of each link of a traditional Chinese medicine preparation process, and the operation control process model comprises a bottom layer process control loop model and an operation index and bottom layer controller output model; constructing a controller of a bottom-layer process control ring; designing a suboptimal set value of a controller of a bottom process control loop in the pharmaceutical process by adopting a data driving method; wherein the upper-layer operation control ring is updated by using a set value based on Q-learning according to a set value of an operation index and is transmitted to the bottom-layer process control ring through a zero-order retainer, and a controller of the bottom-layer process control ring gives a controlled quantity to control a controlled object to track a set value by solving an optimization problem. According to the method, the set value is updated by using the data of the operation process of traditional Chinese medicine preparation, and the bottom layer uses the DMPC to track the set value, so that the ideal value of operation index tracking is realized.