Trade-off optimization method for information age and synchronous age of Internet of Things scene

Abstract

The invention discloses a compromise optimization method for information age and synchronization age of an Internet of Things scene, each device transmits in a random access network based on a time slot ALOHA algorithm, only when the device has an untransmitted data packet and the weighted gain of the information age and the synchronization age is greater than a threshold, the transmission is selected, otherwise, the transmission is kept silent. Any device updates the posterior joint probability distribution of the information age and the age gain of the network at the beginning of the time slot, estimates the weighted decline amount of the information age and the age gain of the network through the probability distribution, and then adaptively adjusts the transmission probability and the threshold value according to the weighted decline amount. And finally, the probability distribution is updated according to the channel state and the access parameter. According to the method, the information age and the age gain probability distribution of each device are updated at each time slot through the Bayesian updating rule, and the optimal transmission probability and threshold are calculated, so that the information age and the synchronization age of the random access network are effectively optimized.

Description

Technical Field

[0001] The present invention relates to the field of wireless network communication, and specifically to a method for compromising and optimizing the age of information and the age of synchronization in an Internet of Things (IoT) scenario. Background Art

[0002] In recent years, the number of connected devices in wireless networks has grown exponentially, and more and more latency-sensitive applications have emerged, shifting the focus of current communication metrics from traditional throughput and latency to more detailed performance metrics. AoI is such a metric that quantifies the freshness of information by measuring the time interval since the user last received an update. Synchronization is a key aspect of future wireless communication systems. Accurate synchronization is not only crucial for coordinating time-sensitive operations among different users but also an important metric for the IoT. IoT devices usually need to communicate with a central system or other devices. To ensure data consistency and timeliness, it is necessary to optimize both the age of information and the age of synchronization of IoT devices simultaneously, and make a trade-off between the optimization degrees of the two to achieve a compromise optimization effect. Summary of the Invention

[0003] The present invention aims to provide a method for compromising and optimizing the age of information and the age of synchronization in an IoT scenario, which can optimize the age of information and the age of synchronization of devices in a wireless network simultaneously to improve the timeliness and synchronization of information obtained by network devices. The technical solution for achieving the object of the present invention is: a method for compromising and optimizing the age of information and the age of synchronization in an IoT scenario, and the specific steps are as follows:

[0004] S1: Initialize relevant parameters, let t = 0, the number of devices in the network range be N, the weight parameter be α, the update generation probability be λ, let the initial age of information and the local age value of all devices be 0, and the probability distribution of the initial AoI and age gain of any device n in the network can be obtained as

[0005] S2: At the beginning of time slot t, device n calculates the probability distribution of the current time slot according to λ and the probability distribution at the end of the previous time slot Calculate the probability distribution of the current time slot

[0006] S3: Device n uses the probability distribution to estimate the weighted decrease of network AoI and AoS, and calculates the transmission probability p that maximizes the weighted decrease t And the threshold Γ t ;

[0007] S4: Device n calculates the weighted gain of AoI and AoS and determines whether it is greater than the threshold. If so, execute Scheme A; otherwise, execute Scheme B;

[0008] Scheme A: Device n transmits with probability p in time slot tt Transmit data packets;

[0009] Determine whether the transmission of device n is successful. If so, execute Scheme A1; otherwise, execute Scheme A2;

[0010] Scheme A1: The AoI of device n becomes the local age of device n in the previous time slot plus one, and execute S5;

[0011] Scheme A2: The AoI of device n increases by one compared to the previous time slot, and execute S5;

[0012] Scheme B: Device n remains silent in time slot t and execute A2;

[0013] S5: Device n calculates the probability distribution at the end of time slot t based on the feedback obtained channel state c t and p t and Γ t Calculate the probability distribution at the end of time slot t

[0014] S6: Let t = t + 1;

[0015] S7: Determine whether device n generates an update. If so, execute Scheme C; otherwise, execute Scheme D;

[0016] Scheme C: The local age of device n becomes 0, and execute S2;

[0017] Scheme D: The local age of device n increases by one compared to the previous time slot, and execute S2.

[0018] Furthermore, according to an age-of-information and synchronization-age trade-off optimization method for an Internet of Things scenario described in claim 1, it is characterized in that the network system consists of one AP and N independent devices; each device only accesses the network and sends data packets at the beginning of a time slot, and the time for sending data packets is one time slot; at the beginning of each time slot, each device independently generates a single-time-slot update with probability λ and does not generate an update at other time points; once a device generates a new data packet, the existing old data packet will be discarded; the weight of AoS optimization is α, and the weight of AoI is 1 - α.

[0019] Furthermore, according to an age-of-information and synchronization-age trade-off optimization method for an Internet of Things scenario described in claim 1, it is characterized in that for a known update generation probability λ and the probability distribution at the end of the previous time slot any device can calculate the probability distribution of the current time slot through the Bayesian update rule The specific calculation process is as follows:

[0020] Define the age of information and age gain distribution of the device in time slot t as The calculation formula is:

[0021]

[0022] Considering that each device independently generates an update at the beginning of each time slot with probability λ, it can be obtained that:

[0023]

[0024] For any Where:

[0025]

[0026] Furthermore, according to the method for compromising optimization of age of information and age of synchronization in an Internet of Things scenario described in claim 1, characterized in that, by using the distribution of the local age and age gain of the device, the weighted decrease of the network AoIAoS can be estimated, and the optimal transmission probability and threshold can be calculated by maximizing the weighted decrease. The specific process is as follows:

[0027] S31: First, consider a simple case, that is, under the condition of only considering AoI, it can be obtained that the AoI decrease of a specified device is g - 1 when the transmission is successful and -1 when the transmission fails; Let θ n,t be the probability of success when device n transmits in time slot t. The AoI decrease of device n in time slot t can be obtained as:

[0028]

[0029] Where is the expectation operator; I ψ is the indicator function, and when ψ is true, the function value is 1, and in other cases it is 0; u n,t represents the number of active devices excluding device n in time slot t, and the formula is:

[0030]

[0031] Define the AoI decrease of the network in time slot t as R t , which can be expressed as:

[0032]

[0033] S32. Define as the estimated number of active devices excluding the specified device in time slot t; use ξ t,u to represent , and according to the binomial distribution, the formula can be obtained:

[0034]

[0035] Where for any 0 ≤ u ≤ N - 1, there is:

[0036]

[0037] ρ t represents the probability that the specified device will become an active device;

[0038] S33. According to Equation (5), the successful transmission probability of any device in time slot t can be calculated as:

[0039]

[0040] Combining Equations (4) and (9), the network AoI decrease can be calculated as:

[0041]

[0042] S34. To consider the trade-off optimization between AoI and AoS, replace the AoI gain g in Equation (10) with the weighted gain g of AoI and AoS ; Define the AoI and AoS gains of the specified distribution t,w,g as and respectively, and obtain the formula for g as: t,w,g The formula for g

[0043]

[0044] There is a magnitude problem when calculating g t,w,g , so the device needs to normalize the AoI gain and AoS gain respectively before calculating g t,w,g ;

[0045] S35. Analyzing the definitions of AoI and AoS, it can be obtained that for the specified distribution of is g - 1 when the transmission is successful and 0 when the transmission fails; is 0 when the transmission fails. When the transmission is successful, the AoS of the network needs to be calculated. On the premise of using the distribution calculation, first, the AoI distribution of the network needs to be calculated:

[0046]

[0047] After calculating the AoI distribution, the average AoI value of the network can be calculated:

[0048]

[0049] Then, the AoS calculation formula using the distribution can be derived:

[0050]

[0051] S36. Subtract the AoS value of the network after successful device transmission in this distribution from the AoS value of the original network, and the AoS gain of this distribution can be calculated. After calculating the AoI gain and the AoS gain, the weighted gain g of AoI and AoS for the specified distribution can be calculated using Equation (11). of AoI and AoS t,w,g ; Use g t,w,g to replace g in Equation (10), and the change in AoS after successful device transmission can be considered and reflected in the function to obtain the weighted decrease in AoI-AoS:

[0052]

[0053] S37. Observing Equation (15), it can be seen that can be regarded as a function of p t , Γ t and can be expressed as Optimal transmission probability The solution of can be simplified to make the reciprocal of the active devices, that is:

[0054]

[0055] where Nρ t is the number of active devices; after obtaining the optimal transmission probability, becomes a univariate function of Γ t and the optimal threshold can be obtained by searching

[0056] Furthermore, according to an AoI and synchronization age trade-off optimization method for an Internet of Things scenario described in claim 1, characterized in that each device calculates its own AoI-AoS trade-off gain, and only the devices with the trade-off gain reaching the threshold can transmit with probability p t and otherwise will remain silent.

[0057] Furthermore, according to an AoI and synchronization age trade-off optimization method for an Internet of Things scenario described in claim 1, characterized in that a packet can be successfully transmitted only when there is a single device accessing and sending packets in the channel, otherwise a collision occurs and all devices' transmissions fail; after successfully receiving an update from a device, the AP will immediately send an acknowledgment signal to notify the device that there is no error or delay and transfer the channel state information to each device;

[0058] Furthermore, according to the method for compromising and optimizing the information age and synchronization age in an Internet of Things scenario described in claim 1, it is characterized in that after each device finishes transmitting and receives the feedback information of the channel, it will update the distribution based on the feedback information at the end of time slot t. The specific process is as follows:

[0059] Let w n,t+ and g n,t+ respectively represent the local age and age gain of any device n at the end of time slot t; at the end of time slot t, each device can use the Bayesian update rule to calculate t and access parameter p t , Γ t ) under the condition of all globally available information (including channel state c The specific formula is as follows:

[0060]

[0061] For any Where:

[0062]

[0063] Compared with the existing strategies for optimizing AoI or optimizing time synchronization, the present invention has significant advantages. Analyzing the optimization work of domestic and foreign scholars on AoI and AoS, it can be seen that a large amount of research work has been carried out on the optimization of AoI at present. However, in terms of the optimization of synchronization, the existing research is mainly limited to time synchronization, and there are few optimization methods defined for the synchronization age. This strategy simultaneously considers the optimization of AoI and AoS, and can freely choose whether to focus on improving the freshness of network information or on optimizing the synchronization of information by adjusting the weight of AoI-AoS optimization. At the same time, this paper adopts the idea of probability distribution to estimate the gains of AoI and AoS, and obtains the optimal transmission probability and threshold by solving to select the device with a large gain to compete for transmission in the channel. While reducing the device conflicts in the channel, it more effectively selects the device with a large gain for transmission. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 is the flowchart of the method design of the present invention.

[0065] Figure 2 is the simulation and comparison of the average AoI, average AoS, and the weighted sum of both with the number of devices under the optimal strategy, the static no-threshold strategy, and the strategy proposed by the present invention.

[0066] Figure 3 is the simulation and comparison of the average AoI, average AoS, and the weighted sum of both with the update generation probability under the optimal strategy, the static no-threshold strategy, and the strategy proposed by the present invention. Detailed Implementation Modes

[0067] The present invention will be further clarified below in conjunction with the accompanying drawings and specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and not to limit the scope of the present invention. After reading the present invention, those skilled in the art will fall within the scope defined by the appended claims of this application for various equivalent forms of modification of the present invention.

[0068] The present invention is applied to a random access network scenario based on the slotted ALOHA form. Assume that the network consists of an AP and N (N≥2) devices. Each device only accesses the network and sends data packets at the beginning of a time slot, and the time for sending data packets is one time slot. At the beginning of each time slot, each device independently generates a single-time-slot update with probability λ and does not generate updates at other time points. The weight ratio of AoS optimization is α, and the weight ratio of AoS optimization is 1-α. To maintain the freshness of information, once a device generates a new data packet, it will discard the existing old data packet. Devices interfere with each other, and successful transmission can occur if and only if only a single device accesses the channel; otherwise, a collision occurs and all transmitting devices fail to send. After successfully receiving the update of a device, the AP will immediately send an acknowledgment (ACK) signal to notify the device that there is no error or delay and inform the device of the channel state. Any device starts to maintain the posterior joint probability distribution of the local age and age gain of any device through the Bayesian update rule at the beginning of time slot t, and calculates the weighted decrease of AoI and AoS based on this probability distribution. Calculate the transmission probability and threshold when the weighted decrease is the largest as the optimal transmission probability and optimal threshold and use them as parameters to access the network, and design a transmission strategy to make the devices with AoI and AoS weighted gain greater than the threshold compete for transmission in the channel with the optimal probability, so as to achieve the purpose of compromising and optimizing AoI and AoS.

[0069] As Figure 1 shown, the specific content includes:

[0070] S1: Initialize relevant parameters, let t = 0, the number of devices within the network range is N, the weight parameter is α, the update generation probability is λ, let the initial information age and local age values of all devices be 0, and the probability distribution of the initial AoI and age gain of any device n in the network can be obtained as

[0071] S2: At the moment when time slot t starts, device n calculates the probability distribution of the current time slot according to λ and the probability distribution at the end of the previous time slot to obtain the probability distribution of the current time slot

[0072] S3: Device n uses the probability distribution to estimate the weighted decrease of network AoI and AoS, and calculates the transmission probability p t and threshold Γ t ;

[0073] S4: Device n calculates the weighted gain of AoI over AoS and determines whether it is greater than the threshold. If so, execute Plan A; otherwise, execute Plan B.

[0074] Plan A: Device n transmits the data packet with probability p at time slot t t and transmits the data packet.

[0075] Determine whether device n has successfully transmitted. If so, execute Plan A1; otherwise, execute Plan A2.

[0076] Plan A1: The AoI of device n becomes the local age of device n in the previous time slot plus one, and execute S5.

[0077] Plan A2: The AoI of device n increases by one compared with the previous time slot, and execute S5.

[0078] Plan B: Device n remains silent at time slot t and execute A2.

[0079] S5: Device n calculates the probability distribution at the end of time slot t according to the feedback channel state c t and p t and Γ t and calculates the probability distribution at the end of time slot t

[0080] S6: Let t = t + 1;

[0081] S7: Determine whether device n generates an update. If so, execute Plan C; otherwise, execute Plan D.

[0082] Plan C: The local age of device n becomes 0, and execute S2.

[0083] Plan D: The local age of device n increases by one compared with the previous time slot, and execute S2.

[0084] As a preferred solution, the network system consists of one AP and N independent devices; each device only accesses the network and sends data packets at the beginning of the time slot, and the time for sending data packets is one time slot; at the beginning of each time slot, each device independently generates a single time slot update with probability λ and does not generate updates at other time points; once a device generates a new data packet, the existing old data packet will be discarded; the weight of AoS optimization is α, and the weight of AoI is 1 - α.

[0085] As a preferred solution, given the update generation probability λ and the probability distribution at the end of the previous time slot any device can calculate the probability distribution of the current time slot through the Bayesian update rule

[0086] As a preferred solution, the device can estimate the weighted decrease of the network AoIAoS by using the distribution of the local age and age gain, and the optimal transmission probability and threshold can be calculated by maximizing the weighted decrease. The specific process is as follows: S31. The decrease of the network AoI considering only AoI is derived by using the local age and the actual value of AoI; S32. The probability that any device becomes an active device is derived by using the probability distribution; S33. The decrease of the network AoI is calculated by using the probability distribution; S34. The weighted gain of AoIAoS and how to calculate it are defined; S35. The network AoS is calculated by using the probability distribution; S36. The weighted decrease of the network AoIAoS is estimated by using the probability distribution; S37. The optimal transmission probability and the optimal threshold are calculated.

[0087] As a preferred solution, the device can calculate its own AoIAoS trade-off gain, and the device with the trade-off gain reaching the threshold can transmit with probability p t otherwise, it will remain silent.

[0088] As a preferred solution, the packet can be successfully sent only when there is a single device accessing and sending packets in the channel, otherwise a collision will occur and all devices will send failures; after successfully receiving the update of a certain device, the AP will immediately send an acknowledgment signal to notify the device that there is no error or delay and transfer the channel state information to each device;

[0089] As a preferred solution, after the transmission ends and the device receives the feedback information of the channel, it can update the distribution at the end of time slot t by using the channel state and access parameters through the Bayesian update rule.

[0090] Embodiment 1

[0091] The present invention implements the method by using MATLAB software, sets the number of devices N in the network to 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, the update generation probability λ to 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, the weight α to 0.2, the length of the data packet to 1, and the simulation time T to 100000 time slots; Figure 2 Under the condition that N = 50 and λ changes from 0.1 to 1 at an interval of 0.1, the simulation results of the AoIAoS trade-off optimization strategy and the optimal strategy proposed by the present invention and the comparison with the static threshold-free strategy are solved respectively; Figure 3Under the condition that λ = 0.5 and N changes from 10 to 100 at intervals of 010, the simulation results of the AoIAoS trade-off optimization strategy, the optimal strategy, and the static no-threshold strategy proposed by the present invention are solved respectively. It can be seen from the figure that the trends of the simulation results of the three strategies are the same. The AAoI, AAoS, and weighted value of the optimal strategy are the smallest, followed by the AoIAoS trade-off optimization strategy, and finally the static no-threshold strategy. The simulation of the optimal strategy calculates that the weighted sum of AAoI and AAoS of the network is the smallest, which means that under this transmission strategy, the network device has a smaller age of information and age of synchronization, that is, the device can obtain fresher information, and at the same time, the synchronization between devices is better. Therefore, the performance of this strategy is the best. Similarly, it can be obtained that the AoIAoS trade-off optimization strategy proposed by the present invention is the second best, and the static no-threshold strategy is the worst. Analyzing from the numerical values of the AoI and AoS trade-off optimization, the performance of the optimal strategy is about 5 times that of the static no-threshold strategy, and the performance of the AoIAoS trade-off optimization strategy is about 2.4 times that of the static no-threshold strategy. The strategy proposed by the present invention has obvious performance advantages compared with the static no-threshold strategy.

Claims

1. A method for compromising and optimizing the Age of Information (AoI) and Age of Synchronization (AoS) in the Internet of Things scenario, characterized in that the specific steps are as follows: S1: Initialize relevant parameters. Let \(t = 0\), the number of devices in the network be \(N\), the weight parameter be \(\alpha\), the update generation probability be \(\lambda\). Let the initial information age and local age value of all devices be 0. The probability distribution of the initial AoI and age gain of the network that any device \(n\) can obtain is S2: At the start of time slot t, device n calculates the probability distribution of the current time slot based on λ and the probability distribution at the end of the previous time slot ​ S3: Device n estimates the weighted AoIAoS degradation of the network using the probability distribution and calculates the transmission probability p that maximizes the weighted degradation t and the threshold Γ t ; S4: Device n calculates the weighted gain of AoI and AoS and determines whether it is greater than the threshold. If so, execute Scheme A; otherwise, execute Scheme B. Scheme A: Device n transmits data packets at time slot t with probability p t ; Determine whether the transmission of device n is successful. If so, execute Scheme A1; otherwise, execute Scheme A2. Scheme A1: The AoI of device n becomes the local age of device n in the previous time slot plus one, and execute S5. Scheme A2: The AoI of device n increases by one compared with the previous time slot, and execute S5. Scheme B: Device n remains silent in time slot t and execute A2. S5: Device n calculates the probability distribution at the end of time slot t based on the obtained channel state c t and p t and Γ t ​ S6: Let t = t + 1. S7: Determine whether device n generates an update. If so, execute Scheme C; otherwise, execute Scheme D. Scheme C: The local age of device n becomes 0, and execute S2. Scheme D: The local age of device n increases by one compared with the previous time slot, and execute S2.

2. The compromise optimization method for information age and synchronization age in an Internet of Things scenario according to claim 1, characterized in that The network system consists of a common access point (AP) and N independent devices; each device only accesses the network and sends data packets at the beginning of a time slot, and the time for sending data packets is one time slot; at the beginning of each time slot, each device independently generates a single-time slot update with probability λ and does not generate updates at other time points; once a device generates a new data packet, the existing old data packet will be discarded; the weight of AoS optimization is α, and the weight of AoI is 1 - α.

3. The compromise optimization method for information age and synchronization age in an Internet of Things scenario according to claim 1, characterized in that, For a known update generation probability λ and the probability distribution at the end of the previous time slot Any device can calculate the probability distribution of the current time slot through the Bayesian update rule The specific process is as follows: Define the device information age and age gain distribution in time slot t as The calculation formula is as follows: Considering that each device independently generates an update with probability λ at the beginning of each time slot, it can be obtained that: For any Wherein:

4. The method for compromising optimization of information age and synchronization age in an Internet of Things scenario according to claim 1, wherein Using the distribution of the device local age and age gain, the weighted decrease of network AoI and AoS can be estimated. By maximizing the weighted decrease, the optimal transmission probability and threshold can be calculated. The specific process is as follows: S31: First, consider a simple case. Under the condition of only considering AoI, the AoI decrease of the specified device can be obtained as g - 1 when the transmission is successful and -1 when the transmission fails. Let θ n,t be the probability of success when device n transmits at time slot t. The AoI decrease of device n at time slot t can be obtained as follows: Among them is the expectation operator; I ψ is the indicator function, whose function value is 1 when ψ is true and 0 otherwise; u n,t represents the number of active devices excluding device n in time slot t, and the formula is: Define the AoI decrease of the network in time slot t as R t , which can be expressed as: S32. Definition In time slot t, estimate the number of active devices excluding the specified device; use ξ t,u to represent The probability of can be obtained according to the binomial distribution, and the formula is: Where for any 0 ≤ u ≤ N - 1, there is: ρ t represents the probability that the specified device will become an active device; S33. According to Equation (5), the successful transmission probability of any device in time slot t can be calculated as: Combining Equations (4) and (9), the network AoI decrease can be calculated as: S34. To consider the trade-off optimization between AoI and AoS, the AoI gain g in Equation (10) is replaced with the weighted gain g of AoI and AoS ; Define the AoI and AoS gains of the specified distribution t,w,g as and respectively. The formula for obtaining g is: t,w,g ​ When calculating g t,w,g there is a magnitude problem, so the device needs to perform normalization processing on the AoI gain and the AoS gain respectively before calculating g t,w,g ; Analyzing the definitions of AoI and AoS yields a specified distribution of g - 1 when transmission is successful and 0 when transmission fails; 0 when transmission fails. When transmission is successful, calculating AoS of the network is required. On the premise of using distribution calculation, the AoI distribution of the network needs to be calculated first: After calculating the distribution of AoI, the average AoI value of the network can be calculated. After that, the AoS calculation formula using the distribution can be derived. Subtract the AoS value of the network after successful device transmission of this distribution from the AoS value of the original network, and the AoS gain of this distribution can be calculated. After calculating the AoI gain and the AoS gain, the AoI and AoS weighted gain g of the specified distribution can be calculated using Equation (11). of the AoI and AoS weighted gain g t,w,g ; Use g t,w,g Instead of g in Equation (10), the change in AoS upon successful device transmission can be considered and reflected in the function to obtain the AoI-AoS weighted decrease: Observing formula (15), it can be seen that can be regarded as p t , Γ t as a function of, which can be expressed as Optimal transmission probability The solution of can be simplified to let be the reciprocal of the active device, that is: wherein is the number of active devices; after obtaining the optimal transmission probability, becomes Γ t a univariate function of, and the optimal threshold can be obtained by searching 5. The compromise optimization method for information age and synchronization age in an Internet of Things scenario according to claim 1, characterized in that Let each device calculate its own AoI-AoS trade-off gain, and only the devices whose trade-off gain reaches the threshold can transmit with probability p t ; otherwise, it will remain silent.

6. The compromise optimization method for information age and synchronization age in an Internet of Things scenario according to claim 1, characterized in that, The transmission is successful only when there is only a single device accessing and sending data packets in the channel; otherwise, a collision occurs and all device transmissions fail; after successfully receiving the update of a device, the AP will immediately send an acknowledgment signal to notify the device that there is no error or delay and transfer the channel state information to each device.

7. An optimization method for the trade-off between the information age and the synchronization age in an Internet of Things scenario according to claim 1, characterized in that, After the transmission of each device ends and receives the feedback information of the channel, the distribution will be updated based on the feedback information at the end of time slot t. The specific process is as follows: Let w n,t+ and g n,t+ represent the local age and age gain of any device n at the end of time slot t, respectively; at the end of time slot t, each device can use the Bayesian update rule to calculate t and access parameter p t , Γ t ) given all globally available information (including channel state c The specific formula is as follows: For any Wherein:

Images