CN111736465A - Wireless cloud control system scheduling method and system - Google Patents

Wireless cloud control system scheduling method and system Download PDF

Info

Publication number
CN111736465A
CN111736465A CN202010477398.3A CN202010477398A CN111736465A CN 111736465 A CN111736465 A CN 111736465A CN 202010477398 A CN202010477398 A CN 202010477398A CN 111736465 A CN111736465 A CN 111736465A
Authority
CN
China
Prior art keywords
sensor
control
controller
probability
model
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202010477398.3A
Other languages
Chinese (zh)
Other versions
CN111736465B (en
Inventor
康宇
李鹏飞
王涛
赵云波
陈绍冯
吕文君
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Science and Technology of China USTC
Original Assignee
University of Science and Technology of China USTC
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by University of Science and Technology of China USTC filed Critical University of Science and Technology of China USTC
Priority to CN202010477398.3A priority Critical patent/CN111736465B/en
Publication of CN111736465A publication Critical patent/CN111736465A/en
Application granted granted Critical
Publication of CN111736465B publication Critical patent/CN111736465B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
    • G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

The invention discloses a wireless cloud control system scheduling method and system, which comprises the steps of respectively establishing a system model according to each actual controlled object; configuring a cloud computing platform, and determining a prediction time domain of a model prediction control method, and requirements of successful transmission probability of a controller and a sensor; adjusting the sending probability of all sensors; determining a sending threshold according to the sending probability of each sensor; calculating an optimal control sequence according to whether respective sensor data is successfully received or not, and preparing to send the optimal control sequence to respective actuators through control channels; the central scheduler selects a controller to access the control channel; and each executor updates the content of the buffer according to whether the control sequence is successfully received or not, and selects a control quantity from the buffer for execution. The invention can make the sending probability of all sensors reach the high-efficiency Nash equilibrium point, save the sending power consumption of the sensors, avoid the collision of control signals among control channels and ensure the stability of all control systems.

Description

Wireless cloud control system scheduling method and system
Technical Field
The invention relates to the technical field of control, in particular to a wireless cloud control system scheduling method and system.
Background
The wireless cloud control system is a control system for transmitting sensing data and control data through a wireless communication network, and a local controller is transferred to the cloud. With the rapid development of cloud computing that "centralizes" computing resources and wireless communication technologies such as 5G capable of reliable data transmission with ultra-low latency, wireless cloud control systems (which may include a large number of independent control systems) have received increasing attention in recent years from both academic and industrial industries. Existing examples of wireless cloud control systems can be found in drone surveillance, underwater navigation, etc., and are expected to find further applications in the near future.
The wireless cloud control system has unique functions. First, a wireless cloud control system usually has a plurality of independent control systems; second, these control systems must competitively use the wireless communication channel to transmit data; finally, all controllers use shared cloud computing resources and compute control signals at the cloud.
These unique functions present unique challenges to the design and analysis of wireless cloud control systems, one of which is efficient channel access scheduling for all control systems. In practice, individual control systems may greedily access the communication channels, because more data transfers for each control system generally means better control system performance. However, the capacity of the wireless channel is always limited, which means that the above-mentioned demand cannot be satisfied without compensation. Therefore, an effective channel scheduling strategy of the control system is crucial to design the wireless cloud control system.
The traditional centralized or decentralized scheduling strategy is not ideal for the channel access scheduling strategy in the wireless cloud control system, because the global information required by the scheduler in the wireless cloud control system is neither perfect nor nonexistent. In fact, it may not be feasible to install a centralized channel access mechanism in a wireless cloud control system, while installing a decentralized mechanism may be too conservative. Therefore, a new scheduling strategy is needed to take these characteristics of the wireless cloud control system into account.
Disclosure of Invention
The scheduling method of the wireless cloud control system can solve the technical problem that the traditional centralized or distributed scheduling strategy cannot meet the requirement of the wireless cloud control system.
In order to achieve the purpose, the invention adopts the following technical scheme:
a wireless cloud control system scheduling method is based on a cloud computing platform and comprises the following steps:
s100, respectively establishing a system model according to each actual controlled object;
s200, configuring a cloud computing platform according to the established system model to obtain corresponding model prediction controllers and a central scheduler;
s300, determining a prediction time domain of the model prediction control method, and successful transmission probability requirements of a controller and a sensor according to a system model of a controlled object;
s400, adjusting the sending probability of all the sensors according to the successful transmission probability requirement set by each sensor and the successful sending probability of the sensor at the current moment;
s500, determining a sending threshold according to the sending probability of each sensor; when receiving that the current channel gain is higher than the respective threshold, each sensor sends data to the controller;
s600, each controller calculates an optimal control sequence according to whether the sensor data is successfully received or not, and prepares to send the optimal control sequence to each actuator through a control channel;
s700, the central scheduler selects a controller to access the control channel according to the successful transmission probability requirement of the controller determined in S200;
and S800, each executor updates the content of the buffer according to whether the control sequence is successfully received or not, and selects a control quantity from the buffer for execution.
Further, the S100 respectively establishes a system model according to each actual controlled object;
the method specifically comprises the following steps:
setting n controlled objects to respectively establish the following nonlinear systems:
xi(k+1)＝fi(xi(k),ui(k),wi(k))，i＝1,…,n (1)
wherein: k is the system operation time, and k is more than or equal to 0;
Figure BDA0002516280800000021
state vector at time k;
Figure BDA0002516280800000031
is an input vector;
Figure BDA0002516280800000032
is an external interference vector;
Figure BDA0002516280800000033
is a bounded set and contains an origin; f. ofiIs a continuous function; w is ai(k) Are independent and identically distributed random variables and satisfy
Figure BDA0002516280800000034
Further, the S200 performs configuration of a cloud computing platform according to the established system model to obtain corresponding model predictive controllers and central schedulers;
wherein the model predictive controller configuration process specifically includes:
first, a state estimator is provided for each system i (i ═ 1, …, n), respectively
Figure BDA0002516280800000035
Wherein
Figure BDA0002516280800000036
For the estimated state of system i at time k, ui(k-1) represents the true input of system i at time k-1;
secondly, a model predictive controller is set, specifically, the following constraint optimal control is carried out:
Figure BDA0002516280800000037
Figure BDA0002516280800000038
Figure BDA0002516280800000039
Figure BDA00025162808000000310
wherein u isi(k)＝{ui,0(k),…,ui,N-1(k) The decision variables for the optimization problem are,
Figure BDA00025162808000000311
for the estimated state of system i at time k, Li(x, u) and Fi(x) Respectively positive stage cost and terminal cost;
obtaining an optimal solution through optimization problem solving software
Figure BDA00025162808000000312
NiIs a prediction time domain;
and finally, designing a buffer to store the control sequence with successful transmission and provide real control input, wherein the specific process is as follows:
Figure BDA00025162808000000313
Figure BDA0002516280800000041
wherein b isi(k) For the buffer content of system i at time k, dc,i(k) Being a binary variable, 1 indicates success of transmission of the control sequence of system i and 0 indicates failure of transmission；
Figure BDA0002516280800000042
Further, the S200 performs configuration of a cloud computing platform according to the established system model to obtain corresponding model predictive controllers and central schedulers;
the central scheduler randomly selects a controller to access the control channel according to the following probability:
Figure BDA0002516280800000043
wherein p isciFor the selected probability of the central scheduler i,
Figure BDA0002516280800000044
is the successful transmission probability requirement of the sensor i to be designed.
Further, the S300 determines a prediction time domain of the model prediction control method, and successful transmission probability requirements of the controller and the sensor according to the system model of the controlled object;
the method specifically comprises the following steps:
if the successful transmission probability requirement of each sensor is determined
Figure BDA0002516280800000045
Successful transmission probability requirement of each controller
Figure BDA0002516280800000046
And the prediction time domain N of model predictive controli；
Then
Figure BDA0002516280800000047
And NiThe following inequalities need to be satisfied:
Figure BDA0002516280800000048
Figure BDA0002516280800000049
wherein, γiSatisfies Fi(fi(x,0,0))≤γiFi(x)，μi＝infx,uLi(x,u)/Fi(x)。
Further, in the step S400, the sending probabilities of all the sensors are adjusted according to the requirement of the successful transmission probability set by each sensor and the probability of successful sending of the sensor at the current time;
the method specifically comprises the following steps:
determining that each sensor i (i ═ 1, …, n) sends a status x to the central controller at time ki(k) The probability of (d); the transmission probability is adjusted according to the following formula:
Figure BDA0002516280800000051
wherein the initial value is set to
Figure BDA0002516280800000052
psi(k) β∈ [0,1) is a weighting factor,
Figure BDA0002516280800000053
the requirement of successful transmission probability of the sensor i needs to be set in advance, and q issi(psi(k),ps,-i(k) Is the probability of successful transmission, p, of sensor i at time ks,-i(k) The transmission probability of other sensors except the sensor i at the moment k is obtained; q. q.ssi(psi(k),ps,-i(k) Is estimated approximately by counting the ratio of the number of successful transmissions to the total number of transmissions.
Further, in step S500, a sending threshold is determined according to the sending probability of each sensor; when receiving that the current channel gain is higher than the respective threshold, each sensor sends data to the controller;
the method specifically comprises the following steps:
set threshold &siThen sensor i sends with probability 1 the channel gain that is and is only currently obtainedsi(k) Greater than &si；
And the threshold channel gain is obtained by the following equation:
Figure BDA0002516280800000054
wherein p issi(k) Obtained from the formula (3) osi(h) Is a channel gain probability density function of the measurement channel.
Further, the S700 hub scheduler selects a controller to access the control channel according to the successful transmission probability requirement of the controller determined in S200;
wherein the content of the first and second substances,
the selection rule is as follows: defining a random variable theta, wherein theta is i and represents that the ith controller is selected; then a random number e (k) between 0,1 is generated uniformly at time k and the controller is selected accordingly:
Figure BDA0002516280800000061
on the other hand, the invention also discloses a wireless cloud control system scheduling system, which comprises n system models, wherein n is a natural number greater than 1; wherein the content of the first and second substances,
each system model comprises a sensor, a model prediction controller, a central scheduler, a buffer and an actuator;
the controlled signal passes through the sensor, the model predictive controller, the central scheduler and the buffer to the actuator in sequence;
further comprising:
the sensor successful transmission probability determining unit is used for determining a prediction time domain of the model prediction control method, a controller and the successful transmission probability requirement of the sensor according to a system model of a controlled object;
a sensor transmission threshold determining unit for determining a transmission threshold according to the transmission probability of each sensor; when receiving that the current channel gain is higher than the respective threshold, each sensor sends data to the sensor;
wherein the content of the first and second substances,
the model predictive controller calculates an optimal control sequence according to whether respective sensor data is successfully received or not, and prepares to send the optimal control sequence to respective actuators through control channels;
the central scheduler selecting one controller to access the control channel according to the determined successful transmission probability requirement of the model predictive controller;
and the executor updates the content of the buffer according to whether the control sequence is successfully received or not, and selects a control quantity from the content of the buffer for execution.
Further, the n system models are represented as follows:
setting n controlled objects to respectively establish the following nonlinear systems:
xi(k+1)＝fi(xi(k),ui(k),wi(k))，i＝1,…,n (1)
wherein: k is the system operation time, and k is more than or equal to 0;
Figure BDA0002516280800000071
state vector at time k;
Figure BDA0002516280800000072
is an input vector;
Figure BDA0002516280800000073
is an external interference vector;
Figure BDA0002516280800000075
is a bounded set and contains an origin; f. ofiIs a continuous function; w is ai(k) Are independent and identically distributed random variables and satisfy
Figure BDA0002516280800000074
According to the technical scheme, the wireless cloud control system scheduling method and system provided by the invention provide a dual scheduling strategy under a prediction control framework based on a data packet; the scheduling strategies include a decentralized scheduling strategy of the sensor, a centralized scheduling strategy of the controller, and a model predictive controller based on data packets. The solution allows on the one hand to explicitly take into account status and control constraints and on the other hand to actively compensate for unsuccessfully transmitted information by packet-based transmission, thereby achieving stability of all systems.
The invention provides a dual scheduling strategy based on a data packet framework under a wireless cloud control system with a plurality of controlled objects, reasonably selects a proper successful transmission probability demand and a prediction time domain, can enable the transmission probability of all sensors to reach a high-efficiency Nash equilibrium point, saves the transmission power consumption of the sensors, avoids the collision of control signals among control channels, and ensures the stability of all control systems.
Drawings
FIG. 1 is a flow chart of a method of the present invention;
fig. 2 is a schematic diagram of the system of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention.
As shown in fig. 1, the method for scheduling a wireless cloud control system according to this embodiment, based on a dual scheduling policy under a packet prediction control framework, includes the following steps:
s100, respectively establishing corresponding system models according to actual controlled objects;
s200, configuring a cloud computing platform according to the established system model to obtain frames of corresponding model prediction controllers and a central scheduler;
s300, determining a prediction time domain of a model prediction control method and successful transmission probability requirements of a controller and a sensor according to a system model of a controlled object;
s400, adjusting the sending probability of all the sensors according to the successful transmission probability requirement set by each sensor and the successful sending probability of the sensor at the current moment;
s500, determining a sending threshold according to the sending probability of each sensor; when receiving that the current channel gain is higher than the respective threshold, each sensor sends data to the controller;
s600, the controllers of all the systems calculate optimal control sequences according to whether the respective sensor data are successfully received or not, and prepare to send the optimal control sequences to respective actuators through control channels;
s700, the central scheduler randomly selects a controller to access the control channel according to the successful transmission probability requirement of the controller determined in S200;
and S800, updating the content of the buffer by the executors of all the systems according to whether the control sequence is successfully received, and selecting a control quantity from the buffer for execution.
The above steps are explained in detail as follows:
step 1: respectively establishing the following nonlinear systems for the n controlled objects;
xi(k+1)＝fi(xi(k),ui(k),wi(k))，i＝1,…,n (1)
wherein: k is the system operation time, and k is more than or equal to 0;
Figure BDA0002516280800000081
state vector at time k;
Figure BDA0002516280800000082
is an input vector;
Figure BDA0002516280800000083
is an external interference vector.
Figure BDA0002516280800000084
Is a closed set and contains the origin. f. ofiIs a continuous function.
wi(k) Are independent and identically distributed random variables and satisfy
Figure BDA0002516280800000091
Step 2: a remote cloud computing platform is configured, including a package-based model predictive controller and a central scheduler.
(1) A packet-based model predictive controller;
first, a state estimator is provided for each system i (i ═ 1, …, n), respectively
Figure BDA0002516280800000092
Wherein
Figure BDA0002516280800000093
For the estimated state of system i at time k, ui(k-1) represents the true input of system i at time k-1.
Secondly, a model predictive controller is set, specifically, the following constraint optimal control problem is set:
Figure BDA0002516280800000094
Figure BDA0002516280800000095
Figure BDA0002516280800000096
Figure BDA0002516280800000097
wherein u isi(k)＝{ui,0(k),…,ui,N-1(k) The decision variables for the optimization problem are,
Figure BDA0002516280800000098
for the estimated state of system i at time k, Li(x, u) and Fi(x) Respectively positive stage cost and finalEnd cost. The optimal solution is obtained by commonly used optimization problem solving software (such as Matlab)
Figure BDA0002516280800000099
NiIs the prediction time domain.
Finally, a buffer is required to be designed to store the control sequence of successful transmission and provide the actual control input, and the specific process is as follows:
Figure BDA00025162808000000910
Figure BDA0002516280800000101
wherein b isi(k) For the buffer content of system i at time k, dc,i(k) Being a binary variable, 1 indicates that the control sequence of system i was successfully transmitted and 0 indicates that the transmission failed.
Figure BDA0002516280800000102
And step 3: determining a successful transmission probability requirement for each sensor
Figure BDA0002516280800000103
Successful transmission probability requirement of each controller
Figure BDA0002516280800000104
And the prediction time domain N of model predictive controli。
In fact, it is possible to use,
Figure BDA0002516280800000105
and NiThe following inequalities need to be satisfied:
Figure BDA0002516280800000106
Figure BDA0002516280800000107
wherein, γiSatisfies Fi(fi(x,0,0))≤γiFi(x)，μi＝infx,uLi(x,u)/Fi(x)。
And 4, step 4: determining that each sensor i (i ═ 1, …, n) sends a status x to the central controller at time ki(k) The probability of (c). The transmission probability is adjusted according to the following formula
Figure BDA0002516280800000108
Wherein the initial value is set to
Figure BDA0002516280800000109
psi(k) β∈ [0,1) is a weighting factor,
Figure BDA00025162808000001010
for the successful transmission probability requirement of sensor i, the setting, q, needs to be implementedsi(psi(k),ps,-i(k) Is the probability of successful transmission, p, of sensor i at time ks,-i(k) Is the transmission probability at time k for other sensors than sensor i. q. q.ssi(psi(k),ps,-i(k) May be approximated by counting the ratio of the number of successful transmissions to the total number of transmissions.
And 5: and determining whether the sensor i sends or not according to a threshold strategy.
Given threshold
Figure BDA0002516280800000111
Then sensor i sends a channel gain of if and only if it is currently obtained with probability 1si(k) Greater than &si. And the threshold channel gain may be obtained by the following equation:
Figure BDA0002516280800000112
wherein p issi(k) Obtained from the formula (3) osi(h) Is a channel gain probability density function of the measurement channel. The above equation can be solved approximately by bisection
Figure BDA0002516280800000113
The current channel gain needs to be measured before the sensor sends datasi(k) The measurement process can be easily implemented by a short frame pilot signal.
Step 6: all systems i (i ═ 1, …, n) respectively solve model predictive controllers, and the obtained
Figure BDA0002516280800000114
Ready to be sent to actuator i.
And 7: the central scheduler selects an appropriate control sequence to send to the actuators. The selection rule is as follows: defining a random variable θ, θ ═ i denotes selecting the ith controller. Then a random number e (k) between 0,1 is generated uniformly at time k and the controller is selected accordingly:
Figure BDA0002516280800000115
and 8: the actuators of the respective systems select appropriate control amounts and apply to the systems. The concrete way is by designing the buffer, and the updating of the content of the buffer is exactly the same as (2).
In the following, specific embodiments are discussed in connection with control examples of n mobile robots (n ═ 3):
if xi＝[xi,yi,θi]TTo represent the state vector of the i-th robot, ui＝[vi,wi]TFor corresponding control inputs, wdiFor the corresponding disturbance, the model of each robot is then a discrete-time nonlinear system as follows:
Figure BDA0002516280800000121
wherein (x)i,yi) Is the position of the ith robot, thetaiIs a direction, vi,wiLinear velocity and angular velocity, respectively, T ═ 0.6s is the sampling period, wdi(k)＝sin(7k)×[0.12 0.14 0.1]TIs an external disturbance.
The model predictive controller and the corresponding central scheduler in the cloud control platform are designed according to the step 2, and related parameters can be respectively determined as
Figure BDA0002516280800000122
And Ni＝12(i＝1,2,3)。
Updating the sending probability of the sensor, whether the sensor sends or not, calculating an optimal control sequence, randomly selecting a central scheduler and selecting a real control quantity can be respectively determined according to the steps 4 to 8.
On the other hand, the embodiment of the invention also discloses a wireless cloud control system scheduling system, which comprises n system models, wherein n is a natural number greater than 1; wherein the content of the first and second substances,
each system model comprises a sensor, a model prediction controller, a central scheduler, a buffer and an actuator;
the controlled signal passes through the sensor, the model predictive controller, the central scheduler and the buffer to the actuator in sequence;
further comprising:
the sensor successful transmission probability determining unit is used for determining a prediction time domain of the model prediction control method, a controller and the successful transmission probability requirement of the sensor according to a system model of a controlled object;
a sensor transmission threshold determining unit for determining a transmission threshold according to the transmission probability of each sensor; when receiving that the current channel gain is higher than the respective threshold, each sensor sends data to the controller;
wherein the content of the first and second substances,
the model predictive controller calculates an optimal control sequence according to whether respective sensor data is successfully received or not, and prepares to send the optimal control sequence to respective actuators through control channels;
the central scheduler selecting one controller to access the control channel according to the determined successful transmission probability requirement of the model predictive controller;
and the executor updates the content of the buffer according to whether the control sequence is successfully received or not, and selects a control quantity from the content of the buffer for execution.
Further, the n system models are represented as follows:
setting n controlled objects to respectively establish the following nonlinear systems:
xi(k+1)＝fi(xi(k),ui(k),wi(k))，i＝1,…,n (1)
wherein: k is the system operation time, and k is more than or equal to 0;
Figure BDA0002516280800000131
state vector at time k;
Figure BDA0002516280800000132
is an input vector;
Figure BDA0002516280800000133
is an external interference vector;
Figure BDA0002516280800000134
is a bounded set and contains an origin; f. ofiIs a continuous function; w is ai(k) Are independent and identically distributed random variables and satisfy
Figure BDA0002516280800000135
It is understood that the system provided by the embodiment of the present invention corresponds to the method provided by the embodiment of the present invention, and the explanation, the example and the beneficial effects of the related contents can refer to the corresponding parts in the method.
In summary, the invention provides a dual scheduling strategy based on a data packet framework under a wireless cloud control system with a plurality of controlled objects, reasonably selects a proper successful transmission probability requirement and a prediction time domain, can enable the transmission probability of all sensors to reach an efficient Nash equilibrium point, saves the transmission power consumption of the sensors, avoids the collision of control signals between control channels, and ensures the stability of all control systems.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims (10)

1. A wireless cloud control system scheduling method is based on a cloud computing platform and is characterized in that: the method comprises the following steps:
s100, respectively establishing a system model according to each actual controlled object;
s200, configuring a cloud computing platform according to the established system model to obtain corresponding model prediction controllers and a central scheduler;
s300, determining a prediction time domain of the model prediction control method, and successful transmission probability requirements of a controller and a sensor according to a system model of a controlled object;
s400, adjusting the sending probability of all the sensors according to the successful transmission probability requirement set by each sensor and the successful sending probability of the sensor at the current moment;
s500, determining a sending threshold according to the sending probability of each sensor; when receiving that the current channel gain is higher than the respective threshold, each sensor sends data to the controller;
s600, each controller calculates an optimal control sequence according to whether the sensor data is successfully received or not, and prepares to send the optimal control sequence to each actuator through a control channel;
s700, the central scheduler selects a controller to access the control channel according to the successful transmission probability requirement of the controller determined in S200;
and S800, each executor updates the content of the buffer according to whether the control sequence is successfully received or not, and selects a control quantity from the buffer for execution.
2. The wireless cloud control system scheduling method of claim 1, wherein: the S100 respectively establishes a system model according to each actual controlled object;
the method specifically comprises the following steps:
setting n controlled objects to respectively establish the following nonlinear systems:
xi(k+1)＝fi(xi(k)，ui(k)，wi(k))，i＝1，...，n (1)
wherein: k is the system operation time, and k is more than or equal to 0;
Figure FDA0002516280790000011
state vector at time k;
Figure FDA0002516280790000012
is an input vector;
Figure FDA0002516280790000013
is an external interference vector;
Figure FDA0002516280790000015
is a bounded set and contains an origin; f. ofiIs a continuous function; w is ai(k) Are independent and identically distributed random variables and satisfy
Figure FDA0002516280790000014
3. The wireless cloud control system scheduling method of claim 2, wherein: the S200 carries out configuration of the cloud computing platform according to the established system model to obtain corresponding model prediction controllers and a central scheduler;
wherein the model predictive controller configuration process specifically includes:
first, a state estimator is provided for each system i (i ═ 1.., n), respectively
Figure FDA0002516280790000021
Wherein
Figure FDA0002516280790000022
For the estimated state of system i at time k, ui(k-1) represents the true input of system i at time k-1;
secondly, a model predictive controller is set, specifically, the following constraint optimal control is carried out:
Figure FDA0002516280790000023
Figure FDA0002516280790000024
Figure FDA0002516280790000025
Figure FDA0002516280790000026
wherein u isi(k)＝{ui，0(k)，...，ui，N-1(k) The decision variables for the optimization problem are,
Figure FDA0002516280790000027
for the estimated state of system i at time k, Li(x, u) and Fi(x) Respectively positive stage cost and terminal cost;
obtaining an optimal solution through optimization problem solving software
Figure FDA0002516280790000028
NiIs a prediction time domain;
and finally, designing a buffer to store the control sequence with successful transmission and provide real control input, wherein the specific process is as follows:
Figure FDA0002516280790000029
Figure FDA00025162807900000210
wherein b isi(k) For the buffer content of system i at time k, dc，i(k) Being a binary variable, 1 represents that the control sequence of system i is successfully transmitted and 0 represents that the transmission is failed;
Figure FDA0002516280790000031
4. the wireless cloud control system scheduling method of claim 3, wherein: the S200 carries out configuration of the cloud computing platform according to the established system model to obtain corresponding model prediction controllers and a central scheduler;
the central scheduler randomly selects a controller to access the control channel according to the following probability:
Figure FDA0002516280790000032
wherein p isciFor the selected probability of the central scheduler i,
Figure FDA0002516280790000033
is the successful transmission probability requirement of the sensor i to be designed.
5. The wireless cloud control system scheduling method of claim 4, wherein: the S300 determines a prediction time domain of the model prediction control method, and successful transmission probability requirements of the controller and the sensor according to a system model of the controlled object;
the method specifically comprises the following steps:
if the successful transmission probability requirement of each sensor is determined
Figure FDA0002516280790000034
Successful transmission probability requirement of each controller
Figure FDA0002516280790000035
And the prediction time domain N of model predictive controli；
Then
Figure FDA0002516280790000036
And NiThe following inequalities need to be satisfied:
Figure FDA0002516280790000037
Figure FDA0002516280790000038
wherein, γiSatisfies Fi(fi(x，0,0))≤γiFi(x)，μi＝infx，uLi(x，u)/Fi(x)。
6. The wireless cloud control system scheduling method of claim 5, wherein: s400, adjusting the sending probability of all the sensors according to the successful transmission probability requirement set by each sensor and the successful sending probability of the sensor at the current moment;
the method specifically comprises the following steps:
determining that each sensor i (i 1.., n) sends a status x to the central controller at time ki(k) The probability of (d); the transmission probability is adjusted as followsA step of:
Figure FDA0002516280790000041
wherein the initial value is set to
Figure FDA0002516280790000042
psi(k) β∈ [0,1) is a weighting factor,
Figure FDA0002516280790000043
the requirement of successful transmission probability of the sensor i needs to be set in advance, and q issi(psi(k)，ps，-i(k) Is the probability of successful transmission, p, of sensor i at time ks，-i(k) The transmission probability of other sensors except the sensor i at the moment k is obtained; q. q.ssi(psi(k)，ps，-i(k) Is estimated approximately by counting the ratio of the number of successful transmissions to the total number of transmissions.
7. The wireless cloud control system scheduling method of claim 6, wherein: the S500 determines a sending threshold according to the sending probability of each sensor; when receiving that the current channel gain is higher than the respective threshold, each sensor sends data to the controller;
the method specifically comprises the following steps:
set threshold &siThen sensor i sends with probability 1 the channel gain that is and is only currently obtainedsi(k) Greater than &si；
And the threshold channel gain is obtained by the following equation:
Figure FDA0002516280790000044
wherein p issi(k) Obtained from the formula (3) osi(h) Is a channel gain probability density function of the measurement channel.
8. The wireless cloud control system scheduling method of claim 7, wherein: the S700 hub scheduler selecting a controller to access the control channel based on the successful transmission probability requirement of the controller determined in S200;
wherein the content of the first and second substances,
the selection rule is as follows: defining a random variable theta, wherein theta is i and represents that the ith controller is selected; then a random number e (k) between 0,1 is generated uniformly at time k and the controller is selected accordingly:
Figure FDA0002516280790000051
9. a wireless cloud control system scheduling system which characterized in that:
the method comprises n system models, wherein n is a natural number greater than 1; wherein the content of the first and second substances,
each system model comprises a sensor, a model prediction controller, a central scheduler, a buffer and an actuator;
the controlled signal passes through the sensor, the model predictive controller, the central scheduler and the buffer to the actuator in sequence;
further comprising:
the sensor successful transmission probability determining unit is used for determining a prediction time domain of the model prediction control method, a controller and the successful transmission probability requirement of the sensor according to a system model of a controlled object;
a sensor transmission threshold determining unit for determining a transmission threshold according to the transmission probability of each sensor; when receiving that the current channel gain is higher than the respective threshold, each sensor sends data to the controller;
wherein the content of the first and second substances,
the model predictive controller calculates an optimal control sequence according to whether respective sensor data is successfully received or not, and prepares to send the optimal control sequence to respective actuators through control channels;
the central scheduler selecting one controller to access the control channel according to the determined successful transmission probability requirement of the model predictive controller;
and the executor updates the content of the buffer according to whether the control sequence is successfully received or not, and selects a control quantity from the content of the buffer for execution.
10. The wireless cloud control system scheduling system of claim 9, wherein:
the n system models are represented as follows:
setting n controlled objects to respectively establish the following nonlinear systems:
xi(k+1)＝fi(xi(k)，ui(k)，wi(k))，i＝1，，...，n (1)
wherein: k is the system operation time, and k is more than or equal to 0;
Figure FDA0002516280790000052
state vector at time k;
Figure FDA0002516280790000053
is an input vector;
Figure FDA0002516280790000054
is an external interference vector;
Figure FDA0002516280790000055
is a bounded set and contains an origin; f. ofiIs a continuous function; w is ai(k) Are independent and identically distributed random variables and satisfy
Figure FDA0002516280790000061
CN202010477398.3A 2020-05-29 2020-05-29 Wireless cloud control system scheduling method and system Active CN111736465B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010477398.3A CN111736465B (en) 2020-05-29 2020-05-29 Wireless cloud control system scheduling method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010477398.3A CN111736465B (en) 2020-05-29 2020-05-29 Wireless cloud control system scheduling method and system

Publications (2)

Publication Number Publication Date
CN111736465A true CN111736465A (en) 2020-10-02
CN111736465B CN111736465B (en) 2021-12-14

Family

ID=72646510

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010477398.3A Active CN111736465B (en) 2020-05-29 2020-05-29 Wireless cloud control system scheduling method and system

Country Status (1)

Country Link
CN (1) CN111736465B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112363429A (en) * 2020-11-04 2021-02-12 北方工业大学 Remote intelligent one-key window breaking control method for large vehicle
CN112650251A (en) * 2020-12-24 2021-04-13 北京理工大学 Master-slave multi-agent prediction control system and method based on cloud computing

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102460393A (en) * 2009-05-01 2012-05-16 思杰系统有限公司 Systems and methods for establishing a cloud bridge between virtual storage resources
CN106597881A (en) * 2016-11-03 2017-04-26 深圳量旌科技有限公司 Cloud service robot based on distributed decision-making algorithm
CN107257565A (en) * 2017-06-02 2017-10-17 西安电子科技大学 Wireless sensor network reliability calculation method based on energy and transmission
CN110740456A (en) * 2019-10-15 2020-01-31 四川大学 data transmission method for underwater wireless sensor network
CN110905792A (en) * 2019-11-28 2020-03-24 浙江精工能源科技集团有限公司 Air compressor control system and method based on energy internet cloud computing
CN110930747A (en) * 2018-09-20 2020-03-27 南京锦和佳鑫信息科技有限公司 Intelligent internet traffic service system based on cloud computing technology

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102460393A (en) * 2009-05-01 2012-05-16 思杰系统有限公司 Systems and methods for establishing a cloud bridge between virtual storage resources
CN106597881A (en) * 2016-11-03 2017-04-26 深圳量旌科技有限公司 Cloud service robot based on distributed decision-making algorithm
CN107257565A (en) * 2017-06-02 2017-10-17 西安电子科技大学 Wireless sensor network reliability calculation method based on energy and transmission
CN110930747A (en) * 2018-09-20 2020-03-27 南京锦和佳鑫信息科技有限公司 Intelligent internet traffic service system based on cloud computing technology
CN110740456A (en) * 2019-10-15 2020-01-31 四川大学 data transmission method for underwater wireless sensor network
CN110905792A (en) * 2019-11-28 2020-03-24 浙江精工能源科技集团有限公司 Air compressor control system and method based on energy internet cloud computing

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
PENGFEI LI等: ""Packet-Based Model Predictive Control for Networked Control Systems"", 《2018 IEEE CONFERENCE ON DECISION AND CONTROL》 *
ZHENGKUN QIN等: ""Direct Assimilation of ABI Infrared Radiances"", 《IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING》 *
李恒武等: ""云计算技术下的光纤通信网络优化研究"", 《激光杂志》 *
杨立波等: ""基于云计算的电网预测模型与自适应调度研究"", 《电子器件》 *
疏官胜: ""移动云中基于计算迁移的应用性能优化研究"", 《中国博士学位论文全文数据库 信息科技辑》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112363429A (en) * 2020-11-04 2021-02-12 北方工业大学 Remote intelligent one-key window breaking control method for large vehicle
CN112650251A (en) * 2020-12-24 2021-04-13 北京理工大学 Master-slave multi-agent prediction control system and method based on cloud computing

Also Published As

Publication number Publication date
CN111736465B (en) 2021-12-14

Similar Documents

Publication Publication Date Title
Qi et al. Knowledge-driven service offloading decision for vehicular edge computing: A deep reinforcement learning approach
Zhou et al. Deep reinforcement learning for delay-oriented IoT task scheduling in SAGIN
CN113537514B (en) Digital twinning-based federal learning framework with high energy efficiency
EP3939302A1 (en) Low latency wireless communication system for teleoperated vehicle environments
US10764137B2 (en) Signal-flow architecture for cooperative control and resource allocation
CN111736465B (en) Wireless cloud control system scheduling method and system
CN111050330A (en) Mobile network self-optimization method, system, terminal and computer readable storage medium
CN113905347B (en) Cloud edge end cooperation method for air-ground integrated power Internet of things
CN112511614B (en) Intelligent networking vehicle task unloading method under vehicle-mounted edge environment based on reinforcement learning
CN114205353B (en) Calculation unloading method based on hybrid action space reinforcement learning algorithm
CN114866593A (en) Industrial network control architecture with cooperative end sides and resource joint allocation method
CN113781002A (en) Low-cost workflow application migration method based on agent model and multi-population optimization in cloud edge cooperative network
CN117580105B (en) Unmanned aerial vehicle task unloading optimization method for power grid inspection
CN117008995A (en) Industrial software component service function chain assembly integration method
Fu et al. Periodic asynchronous event-triggered control
CN115134370B (en) Multi-unmanned aerial vehicle assisted mobile edge computing and unloading method
CN114968402A (en) Edge calculation task processing method and device and electronic equipment
CN110855507B (en) Unmanned data network interaction method based on software definition
CN114025321A (en) Massive power internet of things terminal access control method based on rapid uplink authorization
Wu et al. Towards efficient dynamic uplink scheduling over multiple unknown channels
Miller et al. Sensing-constrained power control in digital health
JP6638313B2 (en) Control device, control method, and program
CN115633320B (en) Multi-unmanned aerial vehicle assisted data acquisition and return method, system, equipment and medium
CN113590229B (en) Industrial Internet of things graph task unloading method and system based on deep reinforcement learning
CN117880292B (en) Data center energy consumption optimization method based on electric power-computing power-communication cooperative scheduling

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant