WO2022121985A1 - 一种静态和动态相结合的毫米波波束资源分配与优化方法 - Google Patents

一种静态和动态相结合的毫米波波束资源分配与优化方法 Download PDF

Info

Publication number
WO2022121985A1
WO2022121985A1 PCT/CN2021/136801 CN2021136801W WO2022121985A1 WO 2022121985 A1 WO2022121985 A1 WO 2022121985A1 CN 2021136801 W CN2021136801 W CN 2021136801W WO 2022121985 A1 WO2022121985 A1 WO 2022121985A1
Authority
WO
WIPO (PCT)
Prior art keywords
base station
vehicle
millimeter
wave
alliance
Prior art date
Application number
PCT/CN2021/136801
Other languages
English (en)
French (fr)
Inventor
张奇勋
冯志勇
马万明
尉志青
黄赛
张轶凡
Original Assignee
北京邮电大学
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 北京邮电大学 filed Critical 北京邮电大学
Publication of WO2022121985A1 publication Critical patent/WO2022121985A1/zh

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W72/00Local resource management
    • H04W72/04Wireless resource allocation
    • H04W72/044Wireless resource allocation based on the type of the allocated resource
    • H04W72/046Wireless resource allocation based on the type of the allocated resource the resource being in the space domain, e.g. beams
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W28/00Network traffic management; Network resource management
    • H04W28/02Traffic management, e.g. flow control or congestion control
    • H04W28/06Optimizing the usage of the radio link, e.g. header compression, information sizing, discarding information
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W28/00Network traffic management; Network resource management
    • H04W28/16Central resource management; Negotiation of resources or communication parameters, e.g. negotiating bandwidth or QoS [Quality of Service]
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W4/00Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/30Services specially adapted for particular environments, situations or purposes
    • H04W4/40Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P]
    • YGENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02DCLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00Reducing energy consumption in communication networks
    • Y02D30/70Reducing energy consumption in communication networks in wireless communication networks

Definitions

  • the present application relates to a millimeter-wave beam resource allocation technology in the field of wireless communication, in particular to a millimeter-wave beam resource allocation and optimization method combining static and dynamic.
  • existing resource allocation methods only consider beam interference within or between base stations, lacking a complete analysis of beam interference, and cannot be applied to hotspot scenarios where both millimeter-wave base stations and users deploy multiple beams.
  • the existing methods are only oriented to static resource allocation, and cannot achieve flexible allocation of millimeter-wave beam resources by dynamically sensing the impact of inter-beam interference on network performance, which seriously restricts the improvement of network performance.
  • mmWave frequency band provides the means for vehicle communication systems to achieve higher data rates.
  • High-speed data transmission can be used to exchange raw sensor data or entertainment information applications to improve on-board traffic safety and efficiency, as well as user experience, and has broad application prospects.
  • existing methods only consider resource allocation in simple mmWave vehicle-to-infrastructure (V2I) or mmWave vehicle-to-vehicle (V2V) transmission modes, and mmWave base stations have not been considered Joint optimization of communication links inside and outside the coverage area.
  • existing methods do not dynamically adjust mmWave beam resources based on the differentiated content requirements of vehicles. From the perspective of improving computational efficiency, the existing centralized resource allocation algorithms have the disadvantage of high computational complexity and cannot be applied to vehicle networks with highly dynamic changes in network topology.
  • the distributed resource allocation method is used to enable the millimeter-wave base station and users to have the ability to quickly and independently make decisions. a more efficient method.
  • the communication scenarios applied by the existing distributed resource allocation methods are relatively limited, and cannot be adapted to the static and dynamic millimeter-wave communication scenarios, nor can they flexibly eliminate the interference between millimeter-wave beams and satisfy the differentiated content of multiple users. need.
  • the impact of the differentiated backhaul capability among multiple millimeter-wave base stations on the traffic load balance between backhaul and access links is not fully considered, resulting in two situations: 1) When the backhaul capability of a millimeter-wave base station is much greater than that of the base station When the data rate of all current access links is the sum total, it will cause waste of backhaul spectrum resources and low resource utilization; 2) For millimeter-wave base stations with good access link quality, poor backhaul capability will not be able to satisfy current access users data rate requirements, causing traffic overload problems. Therefore, the number of failed access links increases, further resulting in a decrease in network spatial multiplexing gain and network throughput.
  • the beam interference within and between base stations has not been fully modeled and analyzed, especially when the millimeter-wave base stations and users are densely distributed, the inter-beam interference is more obvious.
  • the traditional centralized resource allocation algorithm has the problem of large amount of calculation and high complexity, which leads to low actual calculation efficiency.
  • the existing resource allocation methods are only for the ideal case where the backhaul link capacity is not limited, and do not solve the actual allocation problem under the limited backhaul of mmWave base stations, and do not consider multiple mmWave Differentiated backhaul capabilities between base stations.
  • the existing resource allocation methods only consider the beam interference within the millimeter-wave base station or between the base stations, and lack the ability to target millimeter-wave base stations. Complete analysis of wave multi-beam interference;
  • the existing mmWave resource allocation method only considers V2I (Vehicle to Infrastructure) or V2V (Vehicle to Vehicle) transmission alone, and has not yet achieved the end-to-end optimization of content distribution and transmission.
  • V2I Vehicle to Infrastructure
  • V2V Vehicle to Vehicle
  • this application proposes a static and dynamic millimeter wave beam resource allocation and optimization method for two practical scenarios, static hotspot area and high dynamic vehicle network, respectively.
  • the developed millimeter wave resource allocation and optimization method ensures the successful transmission of new high-bandwidth services such as virtual reality and augmented reality.
  • the specific steps of the millimeter wave beam resource allocation and optimization method are as follows:
  • Step 1 for a certain user i, determine whether the user's moving speed is less than the speed threshold ⁇ v , if so, go to step 2, classify it into static scene processing; otherwise, the user is located in the vehicle, enter step 9, classify to dynamic scene processing.
  • Step 2 constructing a downlink transmission scenario of a two-layer millimeter-wave heterogeneous cellular network
  • the scenario includes: M millimeter-wave cell base stations, using a set of represents, and each cell base station has the buffering capability.
  • Each millimeter-wave cell base station is connected to the macro base station 0 through the backhaul link, and is connected to I users through the access link, which is represented as a set of users
  • Step 3 Calculate the effective antenna gain, intra-base-station interference and inter-base-station interference of user i and base station m for the current iteration, and report it to base station m;
  • the main lobe gain of the base station transmit antenna is the main lobe gain of the base station transmit antenna, The main lobe gain of the receiving antenna for the user.
  • the intra-base interference gain is expressed as:
  • ⁇ t represents the main lobe beam width of the base station transmit antenna;
  • Indicates the transmit beam The offset angle relative to the reference direction m ⁇ i Indicates the transmit beam The offset angle relative to the reference direction m ⁇ i; is the side lobe gain of the base station transmit antenna.
  • Step 4 Calculate the transmission rate of the millimeter-wave cell base station m and the user i access link by using the effective antenna gain, intra-base station interference and inter-base station interference, and then construct respective matching utility functions U m (i) and U i ( m);
  • P m,i represents the transmit power from base station m to user i; h m,i is the small-scale shadow fading gain from base station m to user i; L m,i refers to the large-scale path loss gain from base station m to user i. Refers to the transmit beam to transmit beam The resulting interference gain within the base station. Refers to interfering transmit beams to the reference receive beam The resulting inter-base-station interference gain. h m',i refers to the small-scale shadow fading gain from the interfering transmitting base station m' to the reference receiving user i. L m',i refers to the large-scale path loss gain from the interfering transmitting base station m' to the reference receiving user i; is the reciprocal of L m',i .
  • P N refers to additive white Gaussian noise power.
  • the signal-to-interference-noise ratio ⁇ m,i is used to calculate the transmission rate of the access link from base station m to user i, which is expressed as:
  • R m,i B a log 2 (1+ ⁇ m,i );
  • B a represents the bandwidth of the access link.
  • is the weighting factor of the backhaul occupancy rate
  • ⁇ ⁇ 0 is the transmission rate from the macro base station 0 to the millimeter-wave cell base station m.
  • the utility function U m (i) of the base station m is as follows:
  • ⁇ 1 is the weight factor of the interference term
  • ⁇ 2 is the weight factor of the backhaul occupancy
  • Im ,i is the total interference caused by user i to other users
  • cm ,v( i) represents the current application content of user i.
  • Step 5 construct the matching utility function of base station m and each user, and form the matching preference list List m of base station m; construct the matching utility function of user i and each base station, and form the matching preference list List i of user i .
  • What is stored in the list List m are the serial numbers of each user arranged according to the utility function from large to small; the list List i is stored in the serial numbers of each base station arranged according to the utility function from large to small.
  • a matching preference list of each base station can be obtained for each base station, and a matching preference list of each user can also be obtained for each user.
  • Step 6 Match multiple base stations and multiple users according to the matching preference list of each base station and each user, obtain a base station-user coordination matrix X, and establish a downlink communication connection.
  • the specific matching process is as follows:
  • base station m adds all its requesting users to the set Judgment set Whether the number of users in is less than or equal to the quota Q m of base station m, if so, base station m accepts the set All user requests in ; otherwise, base station m only accepts users with the highest Q m function values and rejects other application users;
  • Step 7 Based on the base station-user connection state, optimize the transmission power of the millimeter wave beam of the base station based on the convex optimization theory, and generate the optimal solution of the transmission power of the current iteration.
  • x m,i is the base station-user collaborative variable, x m,i ⁇ X;
  • X refers to the base station-user collaborative optimization matrix, which represents whether the base station and the user establish a communication connection; the dimension of the matrix X is M*I.
  • P refers to the transmit power optimization matrix, which represents the transmit power from the transmit beam of the base station to the user; the dimension of the matrix P is M*I; the transmit power variable P m,i ⁇ P.
  • p f refers to the application probability of content f, and obeys the Zipf distribution, that is where ⁇ represents the popularity index of file f.
  • c m,f is a cache variable of the content f, and c m,f ⁇ ⁇ 1,0 ⁇ , representing whether the base station m pairs whether to cache the content f.
  • the constraint condition C1 ensures that the transmit power of each beam is non-negative, and the total transmit power of each base station does not exceed the maximum power value P max ;
  • Constraint C2 ensures that the backhaul capacity occupied by each base station to transmit unbuffered content does not exceed the backhaul link capability of the base station; Constraint C3 ensures that the access rate of user i is not less than the minimum rate requirement
  • the sum of the intra-base station interference and the inter-base station interference received; is the received power from base station m to user i and Sum;
  • Step 8 Calculate the network throughput based on the base station-user coordination matrix X and the transmit power of the current iteration, and determine whether the difference between the network throughput of the current iteration and the network throughput of the previous iteration is less than the convergence threshold, if so, That is, the network throughput converges, and the base station-user coordination matrix and transmit power of the current iteration are taken as the optimal solution, and the millimeter-wave base station performs downlink transmission until all data transmission is completed; otherwise, go back to step 3 to re-run the next iteration.
  • Step 9 Construct a content distribution and transmission scenario of the dynamic vehicle network.
  • the scene includes: N vehicles are randomly distributed on a two-lane lane of length L, and each vehicle is traveling in the same direction at different speeds. And the vehicle has cache capability and follows the FMM movement model.
  • the content distribution process can be divided into two stages: the V2I stage and the V2V stage.
  • the vehicle In the V2I phase, the vehicle is within the coverage of a millimeter-wave base station and receives partial segments of popular content ; each base station supports up to Q0 beams for simultaneous transmission.
  • the vehicle In the V2V stage, the vehicle is outside the coverage of the millimeter wave cell, and the vehicles share content through cooperation, so as to obtain as many remaining required content fragments as possible; each vehicle supports at most Q v beams to transmit simultaneously, or support a single millimeter wave wave receiving beam.
  • Popular content collections are Divide each content into unit content segments with a size of s bits, then the size of content c is expressed as sD c bits, where D c represents the number of unit content segments.
  • Step ten determine whether the vehicle is within the coverage of the base station, if yes, go to step eleven; otherwise, go to step thirteen.
  • Step 11 Calculate the channel state between the vehicle i and the millimeter-wave cell base station m, and report it to the millimeter-wave cell base station m;
  • Channel status including effective antenna gain and intra-base station interference
  • Step 12 In each scheduling time slot T t in the V2I stage, each base station selects its own optimal vehicle to transmit downlink data, and then proceeds to Step 17.
  • the specific process is:
  • the power received by the vehicle i from the base station m is calculated as:
  • Step 13 Under each scheduling time slot T t , determine the vehicle connected to each base station as the content sending vehicle, and use the unconnected vehicle as the receiving vehicle;
  • Step 14 Calculate the actual received content quantity of each link by using the links formed by each receiving vehicle and each content sending vehicle; the specific steps are:
  • Step 1401 for the link l i, j formed by the sending vehicle i and the receiving vehicle j, calculate the effective antenna gain G i,j ;
  • Step 1402 calculating that the current link l i,j is subject to interference I i,j from other links l i',j' that transmit content at the same time, expressed as:
  • V2V inter-link interference gain is the V2V inter-link interference gain, expressed as:
  • Step 1403 Calculate the signal-to-interference-noise ratio of the sending vehicle i and the receiving vehicle j by using the V2V inter-link interference, and further calculate the transmission rate R i,j of the access link l i ,j;
  • Step 1404 using the transmission rate R i,j of the access link li,j to calculate the actual number of transmittable content segments of the link li ,j
  • t s refers to the link-related communication time, that is, the time from the alignment to the misalignment of the vehicle sending and receiving beams; s is the bit size of each unit content segment.
  • Step 1405 calculate the number of hit content fragments of links li , j
  • Step 1406 the actual number of content segments that can be transmitted through the link l ij and the number of hit content fragments Calculate the actual received content quantity of link li ,j ;
  • Step 15 Calculate the utility function of alliance cooperation between vehicles and individual vehicles using the actual received content quantity of each link.
  • the utility function of alliance cooperation is the difference between the alliance revenue function and the cost function; the details are as follows:
  • the alliance revenue function is the amount of content actually received by all links l i, j in the alliance S, and the calculation formula is:
  • the cost function is proportional to the number of V2V links in the alliance, and the calculation formula is:
  • is a constant slightly greater than 1; Represents a collection of vehicles sending content; Represents a collection of vehicles that receive content.
  • Step 16 Implement an inter-vehicle alliance for each sending vehicle and each content receiving vehicle according to the alliance cooperation and the utility function of the individual vehicle, and establish a V2V link to realize content distribution.
  • the specific alliance process is as follows:
  • U(S m ⁇ i ⁇ ) refers to the alliance utility value obtained by alliance S m after vehicle i joins the new alliance S m ;
  • U(S k ⁇ i ⁇ ) refers to when vehicle i is still in the old alliance S k , the coalition utility value that the coalition Sk can obtain.
  • the update formula of the alliance combination is:
  • Step 18 Use the optimal results of base station-user collaboration and transmit power in static scenarios, or the optimal results of vehicle alliances in dynamic scenarios, to achieve a combination of static and dynamic millimeter wave beam resource allocation.
  • a static and dynamic millimeter-wave beam resource allocation and optimization method a distributed framework based on matching theory is constructed for the cooperation of multiple base stations and multiple users, which can alleviate the backhaul link while ensuring low complexity. stress and eliminate inter-beam interference, thereby increasing network throughput.
  • a static and dynamic millimeter-wave beam resource allocation and optimization method which constructs a bilateral matching utility function that can dynamically perceive the backhaul capability of millimeter-wave base stations and inter-beam interference, and achieves efficient allocation of millimeter-wave multi-beams.
  • a static and dynamic combination of millimeter wave beam resource allocation and optimization method, for dynamic vehicle network scenarios in the V2I stage, a low-complexity beam selection scheme is proposed to achieve preliminary caching of vehicle content within the coverage of the base station .
  • a cooperative method of sending and receiving vehicles based on alliance game is proposed, which can realize the dual guarantee of vehicle alliance and individual vehicle utility, and realize the efficient distribution of differentiated content.
  • FIG. 1 is a flowchart of a method for allocating and optimizing millimeter-wave beam resources combining static and dynamic functions provided by an embodiment of the present application;
  • FIG. 2 is a schematic diagram of a downlink transmission scenario of a two-layer millimeter-wave heterogeneous cellular network constructed in an embodiment of the present application;
  • FIG. 3 is a schematic diagram of reporting interference in a base station to a base station according to an embodiment of the present application
  • FIG. 3 is a schematic diagram of reporting inter-base station interference to a base station according to an embodiment of the present application
  • FIG. 4 is a schematic diagram of calculating a utility function of a user and a utility function of a base station by using the transmission rate of an access link according to an embodiment of the present application;
  • FIG. 5 is a schematic diagram of a content distribution and transmission scenario for constructing a dynamic vehicle network according to an embodiment of the present application
  • FIG. 6 is a schematic diagram of beam alignment (effective communication link) in a millimeter-wave V2V communication link according to an embodiment of the present application;
  • FIG. 6 is a schematic diagram of incomplete beam alignment (adjacent interference link) in a millimeter-wave V2V communication link according to an embodiment of the present application;
  • FIG. 7 is a schematic diagram of the relevant time of a millimeter-wave V2V communication link according to an embodiment of the present application.
  • FIG. 8 is a flowchart of a method for allocating and optimizing millimeter wave beam resources provided by an embodiment of the present application
  • FIG. 9 is a flowchart of another method for allocating and optimizing millimeter wave beam resources provided by an embodiment of the present application.
  • FIG. 10 is a flowchart of another millimeter wave beam resource allocation and optimization method provided by an embodiment of the present application.
  • the embodiment of the present application is a method for allocating and optimizing millimeter-wave beam resources combining static and dynamic.
  • a distributed multi-base station and multi-user collaboration framework based on matching theory is used to construct a backhaul capable of dynamically sensing millimeter-wave base stations.
  • the bilateral matching utility function of capability and intra-base station and inter-base-station beam interference enables efficient allocation of millimeter-wave multi-beams with low computational complexity.
  • the optimization strategy of millimeter-wave beam transmit power the theoretical convex upper and lower bounds of the optimal transmit power are derived to optimize the network throughput; for dynamic vehicle network scenarios, in the V2I stage, a low-complexity algorithm is proposed.
  • the beam selection scheme realizes preliminary caching of vehicle content within the coverage area of the base station.
  • a cooperative method of sending and receiving vehicles based on alliance game is proposed to achieve efficient distribution of differentiated content.
  • the embodiments of the present application optimize millimeter-wave beam resources of two different dimensions, including distributed multi-base station-multi-user collaborative decision-making based on matching theory, which is mainly responsible for beam spatial dimension allocation and transmit power optimization decision-making.
  • V2I Vehicle to Infrastructure
  • V2V Vehicle to Vehicle
  • the user in the embodiment of the present application refers to a terminal used by the user, where the terminal may be a device such as a mobile phone, a computer, or a tablet computer. As shown in Figure 1, the specific steps are as follows:
  • Step 1 for a certain user i, determine whether the user's moving speed is less than the speed threshold ⁇ v , if so, go to step 2, classify it into static scene processing; otherwise, the user is located in the vehicle, enter step 9, classify to dynamic scene processing. In other words, for user i, it is determined whether the user's moving speed is greater than the speed threshold ⁇ v . If yes, go to step 9 to build a dynamic car-connected network scenario. If not, go to step 2 to build a static hotspot scene.
  • Step 2 Construct a downlink transmission scenario of a two-layer millimeter-wave heterogeneous cellular network.
  • each millimeter-wave cell base station is connected to a macro base station (Macro Cell Base Station, MBS) through a backhaul link, and is connected to one user through an access link, which is represented as a set of users Among them, m is the serial number of the millimeter wave cell base station, i is the serial number of the user, and 0 represents the macro base station.
  • MBS Micro Cell Base Station
  • the millimeter-wave cell base station in this embodiment of the present application may also be referred to as a small cell base station (Small Cell Base Station, SBS).
  • SBS Small Cell Base Station
  • the three millimeter-wave base stations SBS1, SBS2, and SBS3 in the static scene are included, and the number of millimeter-wave base stations in the actual implementation is not limited to this.
  • the access link and the backhaul link operate in different millimeter-wave frequency bands (bandwidths Ba and Bb respectively) to avoid mutual interference between the access link and the backhaul link.
  • the macro base station 0 and the core network are connected by wired optical fibers. In each millimeter-wave cell base station, users are randomly and uniformly distributed.
  • users within the coverage of SBS1 in FIG. 2 include UE1, UE2, and indoor (Virtual Reality, VR)/Augmented Reality (AR) devices, where UE1 is the terminal of a pedestrian who is outdoors, UE2 and indoor VR /AR devices are indoor terminals.
  • Users within the coverage of SBS2 include UE3 and UE4, where UE3 is an outdoor vehicle, and UE4 is an outdoor VR/AR device.
  • the users within the coverage of SBS3 include UE4, UE5 and HD event live broadcast.
  • UE5 is an outdoor terminal
  • HD event live broadcast refers to the live broadcast equipment in the venue capable of high-definition event live broadcast.
  • one millimeter-wave base station may be determined from among M millimeter-wave base stations as a scheduling base station, and steps 3 to 8 are performed by scheduling the base station, thereby determining a user matched by each millimeter-wave base station.
  • Step 3 Calculate the effective antenna gain, intra-base-station interference and inter-base-station interference of user i and base station m for the current iteration, and report it to base station m.
  • the millimeter-wave communication model is analyzed.
  • directional antennas are used in millimeter-wave communication, and the sector antenna model is used. It is assumed that the antennas of each millimeter-wave cell base station have the same beam model Including the main lobe gain of the transmit antenna side lobe gain and the main lobe beamwidth ⁇ t .
  • Beam model representing each user including the main lobe gain of the receive antenna side lobe gain and the main lobe beamwidth ⁇ r . Therefore, the effective antenna gain from base station m to user i is expressed as:
  • the main lobe gain of the base station transmit antenna is the main lobe gain of the base station transmit antenna, The main lobe gain of the receiving antenna for the user.
  • the embodiments of the present application model two situations of intra-base station interference and inter-base station interference in the process of multi-beam parallel transmission.
  • the embodiment of the application performs a complete mathematical analysis on the interference of the millimeter wave beam, as shown in FIG. 3 .
  • FIG. 3 Referring to Fig. 3, (a) in Fig. 3 includes a millimeter-wave base station SBSm and two terminals UEi and UEi'. Since there is partial overlap between the transmit beams of SBSm for UEi and UEi', there is a gap between the two transmit beams. Interference, such interference between beams of the same base station is intra-base station interference. (b) in FIG.
  • 3 includes two millimeter wave base stations SBSm and SBSm', and two terminals UEi and UEi', and SBSm is connected to UEi', and SBSm' is connected to UEi and UEi' respectively.
  • UEi's receiving beams of SBSm and SBSm' partially overlap, resulting in interference between the two receiving beams.
  • Such interference between beams of different base stations is inter-base station interference.
  • the intra-base station interference gain is caused by the mutual overlap between the main lobes of different transmit beams, which is expressed as:
  • ⁇ t represents the main lobe beam width of the base station transmit antenna;
  • Indicates the transmit beam The offset angle relative to the reference direction m ⁇ i Indicates the transmit beam The offset angle relative to the reference direction m ⁇ i; is the side lobe gain of the base station transmit antenna. and respectively represent different transmit beams of the same millimeter-wave base station, Represents the transmit beam of a mmWave base station to transmit beam interference gain.
  • user i receives inter-base-station interference from other base stations, where the interfering transmit beam and the reference receive beam
  • the receive beam is the receive beam of user i for base station m, interferes with the transmit beam Main lobe or side lobe pointing towards user i, referenced to the receive beam Main or side lobes pointing to other base stations m'.
  • the millimeter-wave base station m can calculate the interference transmit beam by formula (2) to the reference receive beam interference, so the inter-base-station interference gain Expressed as:
  • Step 4 Calculate the transmission rate of the millimeter-wave cell base station m and the user i access link by using the effective antenna gain, intra-base station interference and inter-base station interference, and then construct respective matching utility functions U m (i) and U i ( m) .
  • the large-scale path fading gain L m,i from base station m to user i can be expressed as:
  • the gain of small-scale shadow fading from base station m to user i is denoted as h m,i , and h m,i is an exponentially distributed random variable with mean one.
  • P m,i represents the transmit power transmitted by base station m to user i
  • the noise N 0 is Gaussian white noise
  • the total bandwidth of the millimeter wave is B
  • the transmission rate R m,i of the access link of base station m and user i can be expressed as:
  • B a represents the bandwidth of the access link.
  • ⁇ m,i is the signal-to-interference-noise ratio from base station m to user i;
  • G m,i refers to the effective antenna gain from base station m to user i. is the inverse of the path loss gain L m,i .
  • P m,i' refers to the transmit power transmitted by base station m to user i'. Refers to the transmit beam to transmit beam The resulting interference gain within the base station.
  • P m',i' refers to the transmit power transmitted by base station m' to user i'. Refers to interfering transmit beams to the reference receive beam The resulting inter-base-station interference gain.
  • h m',i refers to the small-scale shadow fading gain from the interfering transmitting base station m' to the reference receiving user i.
  • L m',i refers to the large-scale path loss gain from the interfering transmitting base station m' to the reference receiving user i; is the reciprocal of L m',i .
  • P N refers to additive white Gaussian noise power.
  • the numerator of equation (4) represents the received power of user i; the three terms of the denominator are the intra-base station interference, inter-base station interference and additive white Gaussian noise (AWGN) power received by user i in turn.
  • AWGN additive white Gaussian noise
  • PN additive white Gaussian noise
  • the transmission rate R m from the macro base station 0 to the millimeter-wave cell base station m can be obtained.
  • the utility function U i (m) of user i and the utility function U m (i) of base station m are calculated using the transmission rate of the access link.
  • Figure 4 shows M millimeter-wave base stations and N users.
  • the compromise between the quality of the access link and the capability of the backhaul link needs to be considered when cooperating with the base station: That is, to cooperate with the base station with the highest transmission rate, or to cooperate with the base station with the largest backhaul capability.
  • the backhaul occupancy rate of the candidate access link m ⁇ i to the base station m is defined as Represents the backhaul pressure that the access link brings to the current base station. Therefore, the utility function U i (m) of user i is expressed as:
  • U i (m) represents the matching utility function of the terminal when the terminal i communicates with the base station m
  • R m,i represents the transmission rate of the downlink communication connection established between the base station m and the terminal i
  • R m represents the backhaul link of the base station m. transmission rate.
  • U m (i) For each base station m, it is necessary to comprehensively consider the access rate of the cooperative user i and the potential interference caused by the user i to other users in the network.
  • U m (i) optimizes the backhaul resource utilization by sensing the differentiated backhaul capabilities of the base stations. Therefore, the utility function U m (i) of base station m is as follows:
  • U m (i) represents the matching utility function of the base station when the terminal i communicates with the base station m .
  • Im,i is the total interference caused by user i to other users That is, I m,i represents the interference caused to the communication of other terminals when terminal i communicates with base station m;
  • cm ,v(i) represents the current application content segment of user i, that is , cm,v(i) represents the current Application shows the utility penalty for the access link m ⁇ i occupying the backhaul resources.
  • Step 5 construct the matching utility function of base station m and each user, and form the matching preference list List m of base station m; construct the matching utility function of user i and each base station, and form the matching preference list List i of user i .
  • What is stored in the list List m are the serial numbers of each user arranged according to the utility function from large to small; the list List i is stored in the serial numbers of each base station arranged according to the utility function from large to small. Similarly, a matching preference list of each base station can be obtained for each base station, and a matching preference list of each user can also be obtained for each user.
  • the matching preference relation can be expressed as Similarly, base station m with respect to the set of users
  • the matching preference relation of can be expressed as
  • Step 6 Match multiple base stations and multiple users according to the matching preference list of each base station and each user, obtain a base station-user coordination matrix X, and establish a downlink communication connection.
  • the specific matching process is as follows:
  • Each user can connect to a certain number of base stations, and an incompletely connected user refers to a terminal that can still connect to other base stations on the basis of the current connection.
  • each incompletely connected user selects the cell base station with the highest utility function value from the respective matching preference list List i to send the request operation.
  • base station m adds all its requesting users to the set Judgment set Whether the number of users in is less than or equal to the quota Q m of base station m, if so, base station m accepts the set All user requests in ; otherwise, base station m only accepts Q m users with the highest function value and rejects other application users.
  • the rejected users are still not fully connected, repeat the above process, and continue to the next round of iteration, that is, for each rejected user, return to the cell base station m with the highest utility function value in the matching preference list List i . ask until no user is rejected, the matching process ends, and the final optimized base station-user coordination matrix X is obtained.
  • the base station-user coordination matrix X calculated by the scheduling base station represents the matching relationship between each millimeter-wave base station and the user.
  • the scheduling base station can send the base station-user coordination matrix X to each millimeter-wave base station in the set m.
  • - User cooperation matrix X determine the user matched by itself, and establish a downlink communication connection with the determined user.
  • the scheduling base station selects the millimeter-wave base station with the highest utility function value in the matching preference list List i as a candidate matching base station for user i. Then, for each millimeter-wave base station m, the scheduling base station selects a preset number of users as the millimeter-wave base station according to the user's order in the matching preference list List m of the millimeter-wave base station m from the users who use it as a candidate matching base station. Users matched by the base station. The preset number is the maximum number of users that the millimeter-wave base station m can connect to, that is, the quota Q m of the millimeter-wave base station m .
  • Step 7 Based on the base station-user connection state and the convex optimization theory, the millimeter wave beam transmit power of the base station is optimized to generate an optimal solution of transmit power for the current iteration.
  • a cache model on the side of the millimeter-wave cell base station is established.
  • the embodiment of the present application introduces a cache model to effectively relieve the backhaul pressure of the millimeter-wave cell base station.
  • the cell base station m has a buffering capability S m (1 ⁇ S m ⁇ F).
  • S m 1 ⁇ S m ⁇ F.
  • c m,f ⁇ 1,0 ⁇ Represents the buffer variable of base station m for content f
  • the embodiment of the present application considers the worst case that the base station cannot a priori detect the popularity of the cached content, so a random cache strategy is adopted. That is, the millimeter-wave base station can randomly select and cache content from the content library.
  • the resource allocation problem can be mathematically expressed as:
  • x m,i is the base station-user collaborative variable, x m,i ⁇ X;
  • X refers to the base station-user collaborative optimization matrix, which represents whether the base station and the user establish a communication connection; the dimension of the matrix X is M*I.
  • P refers to the transmit power optimization matrix, which represents the transmit power from the transmit beam of the base station to the user; the dimension of the matrix P is M*I; the transmit power variable P m,i ⁇ P.
  • constraint C1 ensures that the transmit power of each beam is non-negative, and the total transmit power of each base station does not exceed the maximum power value P max ;
  • Constraint C2 ensures that the backhaul capacity occupied by each base station to transmit unbuffered content does not exceed this The backhaul link capability of the base station;
  • Constraint C3 ensures that the access rate of user i is not less than the minimum rate requirement
  • the scheduling base station can determine the transmit power of each transmit beam of each millimeter-wave base station based on the convex optimization theory.
  • the sum of the intra-base station interference and the inter-base station interference received; is the received power from base station m to user i and and and.
  • the optimal solution sequence ⁇ P (t) ⁇ of the transmit power is iteratively generated within the time t ⁇ 1,...,T ⁇ using the bounds.
  • the problem P1 can be further transformed into the following form:
  • P (t+1) is a feasible point that is better than P (t) . Since the solution sequence ⁇ P (t) ⁇ is bounded, according to Cauchy's theorem, there are subsequences converge to a finite point which is Therefore, for every time t there exists ⁇ such that:
  • Step 8 Calculate the network throughput based on the base station-user coordination matrix X and the transmit power of the current iteration, and determine whether the difference between the network throughput of the current iteration and the network throughput of the previous iteration is less than the convergence threshold, if so, That is, the network throughput converges, and the base station-user coordination matrix and transmit power of the current iteration are taken as the optimal solution, and the millimeter-wave base station performs downlink transmission until all data transmission is completed; otherwise, go back to step 3 to re-run the next iteration.
  • the optimal transmit power solution P * is obtained through a power optimization strategy; based on P * , the matching preference lists List m and List i on the base station side and the user side are updated.
  • the matching preference list List m is based on the latest access link rate
  • the utility values of all candidate users are sorted in descending order, and the same is true for matching the preference list List i . That is, the matching preference list List i ranks the utility values of all candidate base stations in descending order based on the latest access link rate. According to the updated matching preference list, the latest matching result is obtained. Finally iterate until convergence.
  • Step 9 Construct a content distribution and transmission scenario of the dynamic vehicle network.
  • the system model for dynamic vehicle network is first established, including vehicle movement model, millimeter wave channel model and content distribution calculation.
  • the dashed box on the right in Figure 5 is a schematic diagram of a dynamic moving scene, which includes: N vehicles are randomly distributed on a two-lane highway section of length L, and each vehicle moves towards the same direction at different speeds. direction driving, vehicle collection And the vehicle has cache capability and follows the freeway mobility model (FMM).
  • This scenario also includes a macro base station and a millimeter-wave base station.
  • the macro base station is communicatively connected to each millimeter-wave base station, and each millimeter-wave base station can also be communicatively connected to one or more vehicles.
  • the content distribution process can be divided into two stages: V2I stage and V2V stage.
  • V2I stage the vehicle is within the coverage of the millimeter-wave base station and receives some clips of popular content; considering the limitations of hardware equipment such as antenna arrays, each base station supports up to Q0 beams for simultaneous transmission.
  • the mmWave V2V communication link is shown in Figure 6, where (a) and (b) in Figure 6 represent and beam alignment (effective communication link) and beam incomplete alignment ( adjacent interfering links).
  • (a) in FIG. 6 includes a transmitting vehicle i and a receiving vehicle j, and the transmitting vehicle i and the transmitting beam are aligned with the receiving beam of the receiving vehicle j.
  • (b) in FIG. 6 includes the transmitting vehicle i and the receiving vehicle j, and the transmitting vehicle i and the transmitting beam are not perfectly aligned with the receiving beam of the receiving vehicle j.
  • the vehicle is outside the coverage of the millimeter wave cell, and the vehicles cooperate to share content, so as to obtain as many remaining required content segments as possible; each vehicle communicates in half-duplex mode, supporting at most Q v beams to transmit at the same time, or Supports a single mmWave receive beam. That is, each vehicle can communicate with at most Q v devices (including millimeter wave base stations and other vehicles) through the millimeter wave beam at the same time, and Q v is greater than or equal to 1.
  • the size of the content c can be expressed as sD c bits, where D c represents the size of the unit content segment. number.
  • the embodiment of the present application considers a freeway mobility model (freeway mobility model, FMM) with two lanes and no intersections. All vehicles are initially randomly distributed on the lanes and travel at an initial speed vi . For two adjacent vehicles on the same lane, the limit range of the distance between vehicles is [d min , d max ], where d min is the minimum safe distance, and d max is the maximum separation distance. for each car The limit speed range is [v min ,v max ]. The speed selection of each vehicle is independent, and acceleration or deceleration is randomly selected with probability p based on the acceleration a under each time slot. In order to ensure the safety of the vehicle, the embodiment of the present application does not consider the overtaking situation of the vehicle for the time being. Therefore, in order to prevent the occurrence of overtaking behavior, for vehicles i and j in the same lane, the behavior constraints of vehicle i are given as follows:
  • Step ten determine whether the vehicle is within the coverage of the base station, if yes, go to step eleven; otherwise, go to step thirteen.
  • each millimeter-wave base station can select a vehicle that matches itself through steps eleven and twelve, and use it as a transmitting vehicle to communicate with the transmitting vehicle.
  • the receiving vehicle can select its own vehicle alliance through steps 13 to 16, and communicate with the sending vehicle in the vehicle alliance.
  • Step eleven Calculate the channel state between the vehicle i and the millimeter-wave cell base station m, and report it to the millimeter-wave cell base station m.
  • Channel conditions include effective antenna gain and intra-base station interference.
  • Step 12 In each scheduling time slot T t in the V2I stage, each base station selects its own optimal vehicle to transmit downlink data, and then proceeds to Step 17.
  • the specific process is:
  • the power received by the vehicle i from the base station m is calculated as:
  • the received power of all vehicles in the coverage area received by the base station m is calculated, and the candidate vehicles are sorted according to the power from large to small, and the first Q 0 vehicles are selected for connection.
  • the selection of all optimal connected vehicles of the base station m is completed.
  • the millimeter-wave base station selects a vehicle, it uses the selected vehicle as a sending vehicle, establishes a V2I link with each sending vehicle, and transmits content fragments to the sending vehicle based on the V2I link.
  • Step 13 In each scheduling time slot T t , determine the vehicle connected to each base station as the content sending vehicle, and use the unconnected vehicle as the receiving vehicle.
  • the content sending vehicle in the embodiment of the present application may also be simply referred to as a sending vehicle or an originating vehicle.
  • Step 14 Calculate the actual received content quantity of each link by using the link formed by each receiving vehicle and each content sending vehicle.
  • the actual number of received content in this embodiment of the present application may be referred to as the number of received content segments.
  • the embodiment of the present application adopts a standard logarithmic distance path loss model.
  • the transmission path loss Li,j of the mmWave V2V link l i,j between the sending vehicle i and the receiving vehicle j can be expressed as:
  • Li ,j [dB] A+20log 10 (f c )+10 ⁇ i,j log(d i,j )
  • A represents the atmospheric attenuation value
  • f c represents the center carrier frequency of millimeter wave communication
  • ⁇ i,j represents the path loss index
  • d i,j represents the relative distance between vehicles i and j.
  • milliwatt (milliwatt, mW) and decibel (decibel, dB) are both power units
  • mW represents the linear value of power
  • dB represents the value of relative value.
  • the embodiment of the present application measures the content distribution efficiency from two perspectives: 1) V2V link transmission capability, that is, how many content fragments can actually be transmitted by the originating vehicle to the receiving vehicle; 2) The number of hit contents , that is, the number of application content cached in the originating vehicle.
  • the specific steps are:
  • Step 1401 for the link l i, j formed by the sending vehicle i and the receiving vehicle j, obtain the transmit antenna gain of the sending vehicle i and the receiving antenna gain on the receiving vehicle j side Calculate the effective antenna gain G i,j .
  • Step 1402 calculate that the current link l i,j is subject to interference I i,j from other links l i',j' that transmit content at the same time,
  • the link l i,j formed by the sending vehicle i and the receiving vehicle j is subject to the interference I i,j from the link l i',j ' formed by the sending vehicle i' and the receiving vehicle j' that simultaneously transmit content segments ,Expressed as:
  • V2V inter-link interference gain is the V2V inter-link interference gain, expressed as:
  • Step 1403 Calculate the Signal to Interference plus Noise Ratio (SINR) of the sending vehicle i and the receiving vehicle j by using the V2V inter-link interference, and further calculate the transmission rate R i,j of the access link l i ,j .
  • SINR Signal to Interference plus Noise Ratio
  • the SINR of the content distribution vehicle i can be obtained as Note that the interference Im, i received by vehicle i comes from the intra-base station interference caused by other simultaneous transmit beams of the millimeter-wave base station.
  • Step 1404 using the transmission rate R i,j of the access link l i,j to calculate the actual number of content segments that can be transmitted for the link l i,j
  • t s refers to the link-related communication time, that is, the time from the alignment to the misalignment of the sending and receiving beams of the vehicle, that is, the time from the alignment to the misalignment of the transmit beam of the transmitting vehicle i and the receive beam of the receiving vehicle j;
  • Each transmission time slot T t can be further divided into two phases: the beam alignment phase TA and the data transmission phase T d .
  • the beam alignment delay T A of the link l i,j can be expressed as in, and are the sector-level beamwidths of vehicles i and j, respectively, i.e. represents the sector-level beamwidth of the transmitting vehicle, represents the sector-level beamwidth of the receiving vehicle j, and the pilot represents the pilot transmission time T p .
  • the throughput of link l i,j can be calculated as Among them, B in Shannon's formula refers to the channel bandwidth, and the unit can be Hertz.
  • the link communication related time ts ie the time from alignment to misalignment of the vehicle's transmit and receive beams, is discussed next.
  • ⁇ 7 includes a transmitting vehicle i and a receiving vehicle j
  • B and D represent the position of the transmitting vehicle
  • A, C and E represent the location of the receiving vehicle j
  • represents the receiving beam direction of the receiving vehicle j and the speed of the receiving vehicle j
  • the angle between the directions, ⁇ represents the angle between the relative position between the sending vehicle i and the receiving vehicle j and the speed direction of the receiving vehicle j, represents half of the main lobe beamwidth of the transmit antenna transmitting vehicle i.
  • E is the virtual receiving vehicle mapping position assuming that the relative speed of vehicles i and j is 0, that is, assuming that the speed of receiving vehicle j is the same as that of sending vehicle i, when sending vehicle i moves from position B to position D, receiving vehicle j moves from position A to position A. Move to position E. then you can get Wherein, the length lCE can be obtained based on the following formula: Among them, the angle ⁇ is determined by the relative position and relative speed of the sending and receiving vehicles.
  • Step 1405 calculate the number of hit content fragments of links li , j
  • the efficiency of content distribution also depends on the diversity of content segments.
  • the number of hit content fragments of links l i, j can be defined as:
  • Step 1406 the actual number of content segments that can be transmitted through the link l i,j and the number of hit content fragments Calculate the actual amount of received content for link li ,j .
  • t s is the time from alignment to misalignment between the transmit beam of the transmitting vehicle i and the receive beam of the receiving vehicle j
  • Ri, j is the transmission rate of li,j
  • s is the size of each content segment
  • Step 15 Calculate the utility function of alliance cooperation between vehicles and individual vehicles by using the actual received content quantity of each link.
  • the revenue function of an alliance is defined as the amount of content actually received by all links l i, j .
  • the calculation formula is:
  • V(S) represents the total number of received content fragments corresponding to each V2V link in the vehicle alliance S, Indicates the number of received content segments corresponding to the V2V link established between the sending vehicle i and the receiving vehicle j in the vehicle alliance S.
  • is a constant slightly greater than 1; for the launching vehicle In terms of non-cooperation, the utility of utility in cooperation It means that more utility can be obtained in the alliance than acting alone; for the receiving vehicle, if there is no content to receive without cooperation, the utility obtained is 0, and after joining an alliance, the utility is given Therefore the receiving vehicle can gain more utility in the alliance than acting alone.
  • Step 16 Implement an inter-vehicle alliance for each sending vehicle and each receiving vehicle according to the utility function of alliance cooperation and individual vehicles, and establish a V2V link to realize content distribution.
  • the specific alliance process is as follows:
  • the initial vehicle alliance to which each receiving vehicle belongs is randomly determined, and each vehicle alliance includes a sending vehicle and one or more receiving vehicles.
  • the receiving vehicle For each receiving vehicle, it is judged whether the receiving vehicle prefers to select the new alliance S m compared to the old alliance Sk , that is, whether S m > i Sk , and if so, the receiving vehicle leaves the old alliance Sk and joins the new alliance S m , update the alliance combination at the same time; otherwise, the receiving vehicle continues to select other alliances in the current alliance combination for judgment, until the receiving vehicle finds a preferred new alliance or all other alliances in the current alliance are not preferred, and the alliance process of the receiving vehicle Terminate and proceed to the next receiving vehicle for alliance selection.
  • each receiving vehicle determines the vehicle alliance it is in through steps 13 to 16, and establishes a V2V link with the sending vehicle in its own vehicle alliance, and receives the transmission from the sending vehicle based on the V2V link. content fragment.
  • each millimeter-wave base station selects its own matching sending vehicle, and transmits content fragments to the sending vehicle through the V2I link until the data transmission is completed or the current scheduling period ends.
  • each receiving vehicle selects its own matching sending vehicle, and receives the content fragments sent by the sending vehicle through the V2V link until the data transmission is completed or the current scheduling period ends.
  • Step 18 Use the optimal results of base station-user coordination and transmit power in static scenarios, or the optimal results of vehicle alliances in dynamic scenarios, to achieve a combination of static and dynamic millimeter wave beam resource allocation.
  • V2I mode the base station has limited vehicles that can be connected to each time slot, and the vehicle cannot communicate with the base station when it leaves the coverage area of the base station, making it inefficient to distribute content to vehicles.
  • the vehicle acquires content from surrounding vehicles, which limits the content that the vehicle can acquire.
  • an embodiment of the present application provides a method for allocating and optimizing millimeter-wave beam resources.
  • the method is applied to a vehicle. As shown in FIG. 8 , the method includes:
  • the vehicle When the vehicle is within the coverage of the millimeter-wave base station, if the vehicle is connected to the millimeter-wave base station in the current time slot, the vehicle acts as a transmitting vehicle to receive the content segment sent by the millimeter-wave base station through the beam through the V2I link.
  • the content segment is obtained after segmenting the content in the popular content collection.
  • the current time slot refers to the current scheduled time slot.
  • the millimeter-wave base station may determine a sending vehicle that matches itself through the above steps 11 and 12, and send the first notification message to the sending vehicle.
  • the first notification message is used to notify that the vehicle is matched with the millimeter wave base station.
  • the vehicle receives the first notification message, it determines that it is the sending vehicle, establishes a V2I link with the millimeter-wave base station, and aligns the receiving beam with the millimeter-wave base station to receive the content sent by the millimeter-wave base station through the sending beam through the V2I link Fragment.
  • the vehicle can also broadcast second notification information to notify other vehicles that it is the sending vehicle.
  • the vehicle drives out of the coverage of the millimeter-wave base station, or the vehicle does not drive out of the coverage of the millimeter-wave base station, but the current time slot is not connected to the millimeter-wave base station, the vehicle joins the target vehicle alliance as a receiving vehicle, and passes The V2V link receives the content segment sent by the sending vehicle in the target vehicle alliance.
  • each vehicle alliance includes one sending vehicle and one or more receiving vehicles.
  • the receiving vehicle can obtain the content segment cached by the sending vehicle from the sending vehicle in the vehicle alliance it belongs to.
  • the vehicle can obtain content fragments from the base station when it is connected to the base station, and if it is not connected to the base station, it can obtain content fragments from the sending vehicle of the vehicle alliance to which it belongs, thereby Improve the efficiency of vehicle acquisition of content snippets. That is, in a dynamic scenario, the base station can communicate with the selected sending vehicle in each time slot, and transmit content fragments to the sending vehicle. At the same time, the receiving vehicle that is not connected to the base station can be obtained from the sending vehicle in the form of a vehicle alliance. Content Fragments. This not only ensures the distribution efficiency of the content segments, but also alleviates the limitation of the content acquired by the vehicle, and realizes the efficient distribution of differentiated content.
  • the above-mentioned S802 vehicle as a way for the receiving vehicle to join the target vehicle alliance, can be implemented as the following steps:
  • Step 1 Calculate the number of received content fragments that the receiving vehicle can receive from the V2V link if a V2V link is established between the receiving vehicle and the sending vehicle in the initial vehicle alliance.
  • the initial vehicle alliance is a randomly selected vehicle alliance.
  • Step 2 Calculate the cooperative utility function value of the initial vehicle alliance and the individual utility function value of the receiving vehicle if the receiving vehicle joins the initial vehicle alliance based on the number of received content fragments corresponding to the V2V link.
  • step 2 Calculate the cooperative utility function value of the initial vehicle alliance and the individual utility function value of the receiving vehicle if the receiving vehicle joins the initial vehicle alliance based on the number of received content fragments corresponding to the V2V link.
  • Step 3 Select a vehicle alliance, and calculate the cooperative utility function value of the selected vehicle alliance and the individual utility function value of the receiving vehicle if the receiving vehicle joins the selected vehicle alliance.
  • the method of calculating the cooperative utility function value and the individual utility function value in step 3 is the same as that in step 2, and the description of step 2 can be referred to.
  • Step 4 Determine whether the cooperative utility function value of the selected vehicle alliance is greater than the cooperative utility function value of the initial vehicle alliance, and whether the individual utility function value of the receiving vehicle joining the selected vehicle alliance is greater than the individual utility function value of the receiving vehicle joining the initial vehicle alliance . If both are yes, the selected vehicle alliance is determined as the target vehicle alliance. Otherwise, return to step 3 until the target vehicle alliance is determined.
  • step 4 For the specific implementation manner of step 4, reference may be made to the relevant description in the above-mentioned step sixteen.
  • Step 5 Establish a V2V link with the sending vehicle in the target vehicle alliance.
  • the cooperative utility function of the vehicle alliance is considered, that is, the number of content segments that the receiving vehicle can receive and the vehicle alliance overhead after the receiving vehicle joins the vehicle alliance, so as to minimize the cost of the vehicle alliance. Increase the number of content segments that the receiving vehicle can receive while reducing the overhead of vehicle alliances.
  • the individual utility function of the receiving vehicle is also considered, that is, the number of content fragments that can be received by the receiving vehicle is considered, which further ensures the number of content fragments that the receiving vehicle can receive.
  • calculating the number of received content fragments in the above step 1 includes the following steps:
  • Step 11 The product of the transmitting antenna gain of the transmitting vehicle of the V2V link and the receiving antenna gain of the receiving vehicle is taken as the effective antenna gain of the V2V link.
  • step 11 reference may be made to step 1401 above.
  • Step 12 Calculate the interference from other V2V links on the V2V link. For the specific implementation of step 12, reference may be made to step 1402 above.
  • Step 13 Calculate the transmission rate of the V2V link according to the effective antenna gain of the V2V link and the interference received by the V2V link from other links. For the specific implementation of step 13, reference may be made to step 1403 above.
  • Step 14 Calculate the number of received content segments of the V2V link according to the transmission rate of the V2V link.
  • the specific implementation of step 14 may refer to the above-mentioned steps 1404-1406.
  • the receiving vehicle can determine the transmission rate of the V2V link through the mutual interference of the V2V link between the vehicles, and then obtain the number of content fragments that the receiving vehicle can receive from the sending vehicle based on the V2V link. Since the greater the interference between the links, the slower the transmission rate of the link, and the fewer the content segments transmitted by the link. Therefore, the subsequent selection of a link that can transmit more content segments can reduce the inter-vehicle connection caused by the establishment of a vehicle alliance. In the case of communication interference, the transmission efficiency of content segments is improved.
  • an embodiment of the present application provides a method for allocating and optimizing millimeter-wave beam resources, which is applied to a millimeter-wave base station. As shown in FIG. 9 , the method includes the following steps:
  • S902 Calculate the channel state of each V2I link if a V2I link is established between the millimeter wave base station and each of the multiple vehicles.
  • the channel state includes the effective antenna gain from the millimeter-wave base station to the vehicle, the effective antenna gain is determined based on the intra-base station interference of the millimeter-wave base station, and the intra-base station interference includes the interference between different transmit beams of the millimeter-wave base station.
  • each V2I link select a preset number of vehicles from a plurality of vehicles as sending vehicles.
  • S904 establishing a V2I link with each sending vehicle, and sending content fragments to each sending vehicle through the V2I link, so that the sending vehicle sends content fragments to a receiving vehicle belonging to the same vehicle alliance as itself through the V2V link.
  • the receiving vehicle is a vehicle that is not connected to the millimeter-wave base station in the current time slot.
  • the selection is made based on the interference within the base station, that is, the interference between the selected transmitting vehicle and the V2I link established between the millimeter-wave base station is minimized, and the The communication speed between the millimeter wave base station and each vehicle improves the communication efficiency.
  • the above S903 selects a preset number of vehicles from a plurality of vehicles as sending vehicles according to the channel state of each V2I link, including:
  • the received power of each vehicle within the coverage area of the millimeter-wave base station to receive the millimeter-wave signal sent by the millimeter-wave base station is determined.
  • a preset number of vehicles are selected as sending vehicles.
  • the preset number is Q 0 .
  • the target transmitting vehicle is the transmitting vehicle with the largest receiving power. If so, replace the target sending vehicle with this vehicle. If not, it is determined that the vehicle cannot replace the target transmission vehicle.
  • the millimeter-wave base station can select a vehicle with higher receiving power as the transmitting vehicle, and replace the transmitting vehicle according to the network throughput, so as to obtain a transmitting vehicle matched by the millimeter-wave base station when the network throughput is the highest. That is, the embodiment of the present application can ensure the receiving power of the sending vehicle, avoid the problem of low transmission efficiency caused by too small receiving power, and the embodiment of the present application can improve the network throughput and improve the distribution efficiency of content segments.
  • an embodiment of the present application provides a method for allocating and optimizing millimeter-wave beam resources, which is applied to a millimeter-wave base station, wherein the millimeter-wave base station is located in a two-layer millimeter-wave heterogeneous cellular network, and the heterogeneous cellular network includes A macro base station and a set of millimeter wave base stations, and each millimeter wave base station in the set of millimeter wave base stations is respectively connected to the macro base station through a backhaul link.
  • the method includes the following steps:
  • S1001 Obtain, in the current time slot, the effective antenna gain of each millimeter-wave base station in the millimeter-wave base station set to each terminal within the signal coverage, the transmit power of each transmit beam of each millimeter-wave base station, the intra-base station interference and the inter-base station interference interference.
  • the intra-base station interference includes interference between different transmit beams of the same millimeter-wave base station, and the inter-base station interference includes interference between transmit beams of different millimeter-wave base stations.
  • the specific implementation of S1001 may refer to the above step 3.
  • the transmit power of each transmit beam of each millimeter-wave base station, intra-base station interference and inter-base station interference determine if each millimeter-wave base station is related to A downlink communication connection is established between each terminal within the signal coverage area, and the downlink communication connection corresponds to the matching utility function value of the millimeter-wave base station and the matching utility function value of the terminal.
  • the matching utility function of the terminal may refer to the above formula (5), and the matching utility function of the base station may refer to the above formula (6).
  • S1002 reference may be made to the foregoing step 4.
  • the matching preference list of the terminal for the millimeter-wave base stations includes the serial numbers of all the millimeter-wave base stations in the millimeter-wave base station set.
  • the sequence numbers of the millimeter-wave base stations in the matching preference list are arranged in descending order of the matching utility functions of the terminal, and each matching utility function of the terminal corresponds to a millimeter-wave base station.
  • the matching preference list of the millimeter-wave base station includes the serial numbers of all terminals within the coverage of the millimeter-wave base station.
  • the sequence number of the terminal in the matching preference list is arranged in descending order of matching utility function values of the millimeter-wave base station, and each matching utility function of the millimeter-wave base station corresponds to a terminal.
  • S1004 Create a base station-user coordination matrix according to the matching preference list of each terminal and the matching preference list of each millimeter-wave base station, so that each millimeter-wave base station establishes a downlink communication connection with its own matching terminal based on the base station-user coordination matrix .
  • the base station-user coordination matrix is used to represent the matching relationship between each millimeter-wave base station and the terminal.
  • the millimeter-wave base station when selecting a terminal matched by a millimeter-wave base station, not only considers the interference within the base station but also the interference between the base stations, that is, this method can minimize the communication between each millimeter-wave base station and the matching terminal.
  • Intra-base station interference and inter-base station interference at the same time improve the communication rate between the millimeter wave base station and the terminal, and improve the communication efficiency.
  • the method further includes the following steps:
  • Step I Calculate the connection status of each established downlink communication connection if each millimeter-wave base station establishes a downlink communication connection with a terminal that matches itself based on the base station-user coordination matrix.
  • the connection state includes the effective antenna gain from the millimeter-wave base station to the terminal, intra-base station interference, and inter-base station interference.
  • step 1 reference may be made to the above-mentioned step 7.
  • Step II According to the connection state of each downlink communication connection, the transmit power of each transmit beam of all the millimeter-wave base stations in the set of millimeter-wave base stations is updated based on the convex optimization theory. Wherein, for the specific implementation manner of step II, reference may be made to the above-mentioned step seven.
  • Step III Based on the transmit power of each transmit beam of each millimeter-wave base station, calculate the network throughput of the network formed by the set of millimeter-wave base stations and the terminal if a downlink communication connection is established between each millimeter-wave base station and the matched terminal, and determine Whether the difference between the currently calculated network throughput and the last calculated network throughput is less than a preset convergence threshold. If so, send the base station-user coordination matrix and the transmit power of the transmit beam of each millimeter-wave base station to other millimeter-wave base stations included in the set of millimeter-wave base stations, so that each millimeter-wave base station is based on the base station-user coordination matrix and its corresponding transmission. A terminal whose power matches itself establishes a downlink communication connection. If not, return to S1001. Wherein, for the specific implementation manner of step III, reference may be made to the above-mentioned step eight.
  • the millimeter-wave base station can optimize the transmit power of each transmit beam of all the millimeter-wave base stations in the millimeter-wave base station set with the goal of improving the network throughput of the network formed between the millimeter-wave base station and the terminal based on the convex optimization theory , thereby improving the transmission rate and transmission efficiency between the terminal and the millimeter-wave base station.
  • the above S1004 may be implemented as: for each terminal, the millimeter-wave base station with the highest ranking in the matching preference list of the terminal is selected as the candidate matching base station of the terminal; for each millimeter-wave base station, From the terminals that use the millimeter-wave base station as candidate matching base stations, according to the order of the terminals in the matching preference list of the millimeter-wave base station, a preset number of terminals are selected as the terminals matched by the millimeter-wave base station, and the base station-user collaboration is obtained. matrix.
  • the specific implementation of S1004 may refer to the above-mentioned step 6.
  • the millimeter-wave base station in FIG. 10 may be one of the base stations in the millimeter-wave base station set m, and the millimeter-wave base station is called a scheduling base station.
  • the scheduling base station can determine the matching terminal of each millimeter-wave base station in the set m of millimeter-wave base stations in the manner of FIG. 10, and then send a third notification message to other millimeter-wave base stations to notify the matching between each millimeter-wave base station and the terminal relation.
  • other millimeter-wave base stations After receiving the third notification message, other millimeter-wave base stations establish downlink communication connections with their own matching terminals, and communicate with the connected terminals according to their corresponding transmit power.
  • the scheduling base station can establish a downlink communication connection with its own matching terminal, and communicate with the connected terminal according to its corresponding transmit power.
  • the scheduling base station may select a terminal matched by each millimeter-wave base station in combination with inter-base station interference, intra-base station interference, and the pressure of the base station's backhaul link, so that the communication between the millimeter-wave base station and the terminal is facilitated.
  • the pressure on the backhaul link of the base station can be reduced as much as possible, and the interference between the downlink communication links can be reduced at the same time, and the network throughput can be improved.
  • the embodiments of the present application provide a vehicle, including:
  • the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the steps performed by the vehicle in the above method for allocating and optimizing millimeter wave beam resources.
  • a base station including:
  • the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the steps performed by the millimeter wave base station in the above method for allocating and optimizing millimeter wave beam resources.
  • the embodiments of the present application provide a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to execute the steps performed by the vehicle in the above method for allocating and optimizing millimeter-wave beam resources.
  • Embodiments of the present application provide a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to execute the steps performed by the millimeter-wave base station in the above-mentioned method for allocating and optimizing millimeter-wave beam resources.
  • An embodiment of the present application provides a computer program product, including a computer program, when the computer program is executed by a processor, the computer program implements the steps executed by the vehicle in the above method for allocating and optimizing millimeter wave beam resources.
  • An embodiment of the present application provides a computer program product, including a computer program, when the computer program is executed by a processor, the computer program implements the steps performed by the millimeter-wave base station in the foregoing method for millimeter-wave beam resource allocation and optimization.

Abstract

本申请公开了一种静态和动态相结合的毫米波波束资源分配与优化方法,属于无线通信领域;针对静态热点地区场景,运用基于匹配理论的分布式多基站多用户协同框架,构建了能够动态感知毫米波基站回程能力和基站内、基站间波束干扰的双边匹配效用函数,在计算复杂度低的前提下实现毫米波多波束的高效分配。利用毫米波波束发射功率的优化策略,推导得到最优发射功率的理论凸上界和凸下界,实现网络吞吐量的最优化;针对动态车辆网络场景,在V2I阶段,提出一种低复杂度的波束选择方案实现基站覆盖范围内的车辆内容初步缓存。在V2V阶段,提出一种基于联盟博弈的收发车辆合作方法,实现差异化内容的高效分发。

Description

一种静态和动态相结合的毫米波波束资源分配与优化方法
本申请要求于2020年12月10日提交中国专利局、申请号为202011454084.8,发明名称为“一种静态和动态相结合的毫米波波束资源分配与优化方法”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及无线通信领域中毫米波波束资源分配技术,具体是一种静态和动态相结合的毫米波波束资源分配与优化方法。
背景技术
近年来,智能终端数量的激增和新型业务的不断涌现使得无线通信面临多种挑战,但同时也带来了更大的机遇与更广阔的应用场景。根据国际电信联盟无线电通信部门(International Telecommunication Union-Radiocommunication Sector,ITU-R)项目联盟预测,到2030年,全球移动通信流量将达到每月5泽字节(zettabytes),个人数据速率将达到100每秒吉比特(Gigabits per second,Gbps)。然而,仅仅凭借6吉赫(Giga Hertz,GHz)以下频谱资源已经无法满足新型业务如虚拟现实、增强现实、触觉互联网等对大带宽和高速率的迫切需求。为了应对频谱资源短缺的严峻挑战,毫米波通信被认为是下一代无线通信中的关键技术之一,以60GHz频段为例,每个信道的可用信号带宽高达2GHz,能够支持上千兆的数据速率,在热点覆盖、沉浸交互式体验等方面,有着巨大的应用潜力。
在毫米波静态热点地区场景中,目前的大部分资源分配技术都是基于理想光纤的回程链路架构展开研究,不适用于历史建筑等特殊环境下的通信部署。在最新第三代伙伴项目(the Third Generation Partnership Project,3GPP)16版本中,提出了一种接入回程一体化(Integrated Access and Backhaul,IAB)的新架构,使接入链路和回程链路都能够基于同一网络基础设施进行毫米波通信。目前仅有少数资源分配方法考虑了接入回程一体化的通信架构,然而,这些方法只解决了回程链路容量不受限的理想情况,并没有考虑到实际中毫米波基站回程受限下的分配问题,且没有考虑到多毫米波基站间差异化的回程能力对网络流量负载均衡带来的影响,从而导致资源利用率的提升面临瓶颈。
另一方面,现有的资源分配方法只考虑基站内或基站间的波束干扰情况,缺乏对波束干扰的完备分析,无法适用于毫米波基站和用户均部署多波束的热点场景。与此同时,现有方法仅面向静态的资源分配,无法通过动态感知波束间干扰对网络性能的影响来实现毫米波波束资源的灵活分配,严重制约网络性能的提升。
进一步,从用户移动性角度出发,路面行驶的车辆与其他车辆、基站间的毫米波可靠通信也是当下迫切需要解决的问题之一。毫米波频段下的大带宽为车辆通信系统提供了获得更高数据速率的手段。高速的数据传输可用于交换原始的传感器数据或娱乐资讯应用,以改善车上的交通安全和效率,以及用户体验,具有广阔的应用前景。针对动态车辆网络,现有方法仅考虑简单的毫米波车辆到基础设施(Vehicle to Infrastructure,V2I)或毫米波车辆到车辆(Vehicle to Vehicle,V2V)传输模式下的资源分配,尚未考虑毫米波基站覆盖范围内外的通信链路联合优化。此外,现有方法未基于车辆差异化的内容需求做出毫米波波束资源的动态调整。从提高计算效率的角度出发,现有集中式资源分配算法存在计算复杂 度高的缺点,无法适用于网络拓扑结构高动态变化的车辆网络。
为了克服传统集中式方法引起的计算量过大的难题,实现毫米波多基站多用户场景下的高效波束资源分配,使用分布式资源分配方法、使毫米波基站和用户拥有快速独立决策的能力,是一种较为有效的方法。然而,现有的分布式资源分配方法所应用的通信场景较为受限,无法适配于静态和动态结合的毫米波通信场景,也无法灵活消除毫米波波束间干扰,满足多用户差异化的内容需求。
对于现有的毫米波资源分配技术,存在以下五个方面的提升空间:
第一,针对静态热点地区场景,目前大部分资源分配方案只考虑理想光纤回程部署,未考虑存在光纤部署的地理局限性。尚未考虑无线回程的通信架构,也无法解决毫米波基站在回程链路容量受限下的分配问题。
第二,未充分考虑多毫米波基站间差异化的回程能力对回程、接入链路之间的流量负载平衡的影响,导致2种情况:1)当毫米波基站的回程能力远大于该基站当前所有接入链路的数据速率的总和时,会造成回程频谱资源的浪费,资源利用率低;2)针对接入链路质量好的毫米波基站,回程能力差将无法满足当前接入用户的数据速率需求,造成流量超载问题。因此,失败的接入链路数量增加,进一步导致网络空间复用增益和网络吞吐量的下降。
第三,在毫米波多波束传输场景下,未对基站内及基站间的波束干扰进行完备建模与分析,尤其当毫米波基站和用户分布较为密集时,波束间干扰更为明显。
第四,针对动态车辆网络,现有方法仅优化毫米波V2I或V2V模式下的毫米波资源分配,尚未考虑毫米波基站覆盖范围内外的通信链路联合优化。此外,无法满足车辆差异化的内容需求。从计算效率角度出发,现有集中式资源分配算法存在计算复杂度高的缺点,无法适用于网络拓扑结构高动态变化的车辆网络。
第五,传统的集中式资源分配算法存在计算量大、复杂度高的问题,导致实际计算效率低下。
然而,在静态热点地区场景中,现有的资源分配方法仅针对回程链路容量不受限的理想情况,并没有解决实际中毫米波基站回程受限下的分配问题,且没有考虑多毫米波基站间差异化的回程能力。同时,当毫米波基站和用户分布较为密集时,多波束间干扰严重将导致网络性能增益的下降,但是现有的资源分配方法仅考虑毫米波基站内或基站间的波束干扰情况,缺乏针对毫米波多波束干扰的完备分析;
进一步,在动态车辆网络中,现有的毫米波资源分配方法仅单独考虑V2I(Vehicle to Infrastructure)或V2V(Vehicle to Vehicle)传输,尚未实现内容分发传输的端到端优化,同时也无法满足车辆差异化的内容需求;此外,传统的集中式资源分配算法存在计算复杂度高的缺点,不能够应对多波束多用户干扰消除带来的庞大计算量,实际应用扩展性不高。
发明内容
针对上述问题,本申请分别针对静态热点地区和高动态车辆网络两种实际场景,提出一种静态和动态相结合的毫米波波束资源分配与优化方法,是一种面向新型大带宽业务的系统级的毫米波资源分配与优化方法,保证虚拟现实、增强现实等大带宽新型业务的成功 传输。所述的毫米波波束资源分配与优化方法,具体步骤如下:
步骤一、针对某用户i,判断该用户的移动速度是否小于速度阈值η v,如果是,则转入步骤二,归类到静态场景处理;否则,用户位于车辆内,进入步骤九,归类到动态场景处理。
步骤二、构建两层毫米波异构蜂窝网络的下行传输场景;
场景中包括:M个毫米波小区基站,用集合
Figure PCTCN2021136801-appb-000001
表示,且每个小区基站具备缓存能力。各个毫米波小区基站通过回程链路连接到宏基站0,通过接入链路连接到I个用户,表示为用户集合
Figure PCTCN2021136801-appb-000002
步骤三、针对当前次迭代,计算用户i与基站m的有效天线增益、基站内干扰和基站间干扰,并上报给基站m;
基站m到用户i的有效天线增益,表示为:
Figure PCTCN2021136801-appb-000003
Figure PCTCN2021136801-appb-000004
为基站发射天线的主瓣增益,
Figure PCTCN2021136801-appb-000005
为用户接收天线的主瓣增益。
基站内干扰增益表示为:
Figure PCTCN2021136801-appb-000006
Figure PCTCN2021136801-appb-000007
表示发射波束
Figure PCTCN2021136801-appb-000008
Figure PCTCN2021136801-appb-000009
主瓣间的重叠角度,
Figure PCTCN2021136801-appb-000010
θ t表示基站发射天线的主瓣波束宽度;
Figure PCTCN2021136801-appb-000011
表示发射波束
Figure PCTCN2021136801-appb-000012
相对基准方向m→i的偏移角度,
Figure PCTCN2021136801-appb-000013
表示发射波束
Figure PCTCN2021136801-appb-000014
相对基准方向m→i的偏移角度;
Figure PCTCN2021136801-appb-000015
为基站发射天线的旁瓣增益。
基站间干扰增益
Figure PCTCN2021136801-appb-000016
表示为:
Figure PCTCN2021136801-appb-000017
Figure PCTCN2021136801-appb-000018
为干扰发射波束
Figure PCTCN2021136801-appb-000019
相对于基准方向m′→i的波束偏移角度;
Figure PCTCN2021136801-appb-000020
为参考接收波束
Figure PCTCN2021136801-appb-000021
相对于基准方向m′→i的波束偏移角度。
步骤四、利用有效天线增益、基站内干扰和基站间干扰,计算毫米波小区基站m和用户i接入链路的传输速率,进而分别构建各自的匹配效用函数U m(i) U i(m);
首先,利用基站内干扰和基站间干扰计算基站m到用户i的信干噪比γ m,i
Figure PCTCN2021136801-appb-000022
P m,i代表基站m发射给用户i的发射功率;h m,i为基站m到用户i的小尺度阴影衰落的增益;L m,i指基站m到用户i的大尺度路径损耗增益。
Figure PCTCN2021136801-appb-000023
指发射波束
Figure PCTCN2021136801-appb-000024
对发射波束
Figure PCTCN2021136801-appb-000025
造成的基站内干扰增益。
Figure PCTCN2021136801-appb-000026
指干扰发射波束
Figure PCTCN2021136801-appb-000027
对参考接收波束
Figure PCTCN2021136801-appb-000028
造成的基站间干扰增益。h m',i指干扰发射基站m'到参考接收用户i的小尺度阴影衰落增益。L m',i指干扰发射基站m'到参考接收用户i的大尺度路径损耗增益;
Figure PCTCN2021136801-appb-000029
是L m',i的倒数。P N指加性高斯白噪声功率。
然后,利用信干噪比γ m,i计算基站m到用户i的接入链路的传输速率,表示为:
R m,i=B alog 2(1+γ m,i);
B a表示接入链路的带宽。最后,利用接入链路的传输速率计算用户i的效用函数U i(m)和基站m的效用函数U m(i)
用户i的效用函数U i(m)表示为:
Figure PCTCN2021136801-appb-000030
其中,τ为回程占用率的权重因子,且τ≥0;R m为宏基站0到毫米波小区基站m的传输速率。基站m的效用函数U m(i)如下:
Figure PCTCN2021136801-appb-000031
ω 1为干扰项的权重因子,ω 2为回程占用率的权重因子,且ω 1≥0,ω 2≥0;I m,i为用户i对其他用户造成的总干扰;c m,v(i)表示用户i当前的申请内容。
步骤五、构建基站m和每个用户的匹配效用函数,形成基站m的匹配偏好列表List m;构建用户i和每个基站的匹配效用函数,形成用户i的匹配偏好列表List i
列表List m中存储的是按效用函数从大到小排列的各用户序号;列表List i中存储的是按效用函数从大到小排列的各基站序号。
同理,对每个基站都能得到各基站的匹配偏好列表,对每个用户也能得到各用户的匹配偏好列表。
步骤六、根据每个基站和每个用户的匹配偏好列表对多基站-多用户进行匹配,得到基站-用户协同矩阵X,建立下行通信连接。具体匹配过程如下:
首先,针对未完全连接的用户i,向匹配偏好列表List i中效用函数值最高的小区基站m发送接入请求
Figure PCTCN2021136801-appb-000032
并将该基站m的序号从列表List i中清除。同理,每个未完全连接的用户都从各自的匹配偏好列表List i中选择效用函数值最高的小区基站发送请求操作;
然后,基站m将它的所有请求用户加入集合
Figure PCTCN2021136801-appb-000033
判断集合
Figure PCTCN2021136801-appb-000034
中的用户数量是否小于等于基站m的配额Q m,如果是,基站m接受集合
Figure PCTCN2021136801-appb-000035
中所有的用户请求;否则,基站m只接受Q m个函数值最高的用户并且拒绝其他申请用户;
被拒绝的用户仍属于未完全连接,重复上述过程,继续进行下一轮迭代,直至最终不存在用户被拒绝时,匹配过程结束,得到最终优化的基站-用户协同矩阵X。
步骤七、基于基站-用户连接状态,基于凸优化理论优化基站的毫米波波束发射功率, 生成当前次迭代的发射功率最优解。
首先,构建基站的毫米波波束发射功率的资源分配模型:
Figure PCTCN2021136801-appb-000036
x m,i为基站-用户协同变量,x m,i∈X;X指基站-用户协同优化矩阵,表征基站和用户是否建立通信连接;矩阵X的维度为M*I。
P指发射功率优化矩阵,表征基站的发射波束到用户的发射功率大小;矩阵P的维度为M*I;发射功率变量P m,i∈P。
p f指内容f的申请概率,且服从Zipf分布,即
Figure PCTCN2021136801-appb-000037
其中δ代表文件f的流行度指数。c m,f为内容f的缓存变量,且c m,f∈{1,0},代表基站m对是否缓存内容f。
其中,约束条件C1保证每个波束的发射功率非负值,且每个基站的总发射功率不超过功率最大值P max
约束条件C2保证每个基站传输未缓存内容占用的回程容量不超过该基站的回程链路能力;约束条件C3保证用户i的接入速率不小于最低速率需求
Figure PCTCN2021136801-appb-000038
然后,利用连续凸逼近求解法,计算得到频谱效率r m,i的凸上界和凸下界。
具体过程如下:
a)、利用用户i与基站m的有效天线增益、基站内干扰和基站间干扰,计算中间函数:
Figure PCTCN2021136801-appb-000039
Figure PCTCN2021136801-appb-000040
是指发射波束
Figure PCTCN2021136801-appb-000041
受到的基站内干扰和基站间干扰的总和;
Figure PCTCN2021136801-appb-000042
是基站m到用户i的接收功率和
Figure PCTCN2021136801-appb-000043
的总和;
b)、然后利用两中间函数相减,计算频谱效率r m,i
Figure PCTCN2021136801-appb-000044
c)、分别利用泰勒展开公式和对数函数性质对两个函数进行凸逼近处理。
具体为:
对函数
Figure PCTCN2021136801-appb-000045
利用泰勒展开公式近似为一阶泰勒函数的形式:
Figure PCTCN2021136801-appb-000046
对函数
Figure PCTCN2021136801-appb-000047
利用对数函数性质
Figure PCTCN2021136801-appb-000048
近似为:
Figure PCTCN2021136801-appb-000049
d)、基于上述两个函数的凸逼近处理结果,得到频谱效率r m,i的凸上界
Figure PCTCN2021136801-appb-000050
表示为:
Figure PCTCN2021136801-appb-000051
相反地,将上述两个函数分别利用对数函数性质和泰勒展开的操作逼近,得到r m,i的凸下界,表示为:
Figure PCTCN2021136801-appb-000052
最后,利用边界在t∈{1,...,T}的时间内迭代生成发射功率的最优解序列{P (t)};
步骤八、基于当前次迭代的基站-用户协同矩阵X和发射功率计算网络吞吐量,并判断当前次迭代的网络吞吐量与前一次迭代的网络吞吐量的差值是否小于收敛阈值,如果是,即网络吞吐量收敛,当前次迭代的基站-用户协同矩阵和发射功率作为最优解,毫米波基站进行下行传输直到所有数据传输完毕;否则,返回步骤三重新运行下一次迭代。
步骤九、构建动态车辆网络的内容分发传输场景。
场景中包括:N个车辆随机分布在长度为L的双车道上,各个车辆以不同的速度朝同一方向行驶,车辆集合
Figure PCTCN2021136801-appb-000053
且车辆具备缓存能力,并遵从FMM移动模型。
内容分发过程可以分为两个阶段:V2I阶段和V2V阶段。
在V2I阶段,车辆处于毫米波基站的覆盖范围内,接收到流行内容的部分片段;每个基站最多支持Q 0个波束同时传输。
在V2V阶段,车辆在毫米波小区覆盖范围外,车辆间通过合作来进行内容共享,从而 获取尽可能多的剩余所需内容片段;每个车辆最多支持Q v个波束同时传输,或支持单个毫米波接收波束。
流行内容集合为
Figure PCTCN2021136801-appb-000054
将各个内容切分为大小为s比特的单位内容片段,则内容c的大小表示为sD c比特,其中D c表示单位内容片段的个数。
步骤十、判断车辆是否在基站覆盖范围内,如果是,转到步骤十一;否则,转到步骤十三。
步骤十一:计算车辆i与毫米波小区基站m的信道状态,并上报给毫米波小区基站m;
信道状态包括有效天线增益和基站内干扰;
步骤十二:在V2I阶段每个调度时隙T t下,每个基站挑选各自的最优车辆来传输下行数据,进入步骤十七。具体过程为:
首先,在V2I阶段,计算车辆i接收到基站m的功率为:
Figure PCTCN2021136801-appb-000055
其中,
Figure PCTCN2021136801-appb-000056
为基站m发射给车辆i的发射功率。
然后,同理计算基站m接收到覆盖范围内所有车辆的接收功率,并按功率从大到小对各候选车辆进行排序,选择与前Q 0个车辆进行连接;
最后,在当前调度时隙T t内,遍历所有已连接车辆,找到传输速率最高的已连接车辆i。判断当前未连接车辆i′替换掉车辆i后的网络吞吐量是否增加,如果是,将车辆i′替换掉车辆i,否则,将当前未连接车辆i′丢弃,继续进行下一个未连接车辆的替换工作。
依次将所有未连接车辆均替换完毕后,完成基站m的所有最优连接车辆选择。
步骤十三:每个调度时隙T t下,确定与各基站连接的车辆作为内容发送车辆,将未连接的车辆作为接收车辆;
步骤十四:利用每个接收车辆和每个内容发送车辆形成的链路,计算各链路的实际接收内容数量;具体步骤为:
步骤1401,针对发送车辆i和接收车辆j形成的链路l i,j,计算有效天线增益G i,j
Figure PCTCN2021136801-appb-000057
步骤1402,计算当前链路l i,j受到来自于同时传输内容的其他链路l i',j'的干扰I i,j,表示为:
Figure PCTCN2021136801-appb-000058
P t为车辆发射波束的功率;h i',j为干扰发送车辆i'到参考接收车辆j链路的小尺度阴影衰落;L i',j为干扰发送车辆i'到参考接收车辆j链路的大尺度路径损耗;
Figure PCTCN2021136801-appb-000059
为V2V链路间干扰增益,表示为:
Figure PCTCN2021136801-appb-000060
Figure PCTCN2021136801-appb-000061
为基站接收天线的旁瓣增益;
Figure PCTCN2021136801-appb-000062
为干扰发射波束
Figure PCTCN2021136801-appb-000063
相对于基准方向i′→j的波束偏移角度;
Figure PCTCN2021136801-appb-000064
为参考接收波束
Figure PCTCN2021136801-appb-000065
相对于基准方向i′→j的波束偏移角度。
步骤1403,利用V2V链路间干扰计算发送车辆i和接收车辆j的信干噪比,进一步计算接入链路l i,j的传输速率R i,j
步骤1404,利用接入链路l i,j的传输速率R i,j,计算链路l i,j的实际可传输内容片段数量
Figure PCTCN2021136801-appb-000066
Figure PCTCN2021136801-appb-000067
t s指链路相关通信时间,即车辆收发波束从对准到失去对准的时间;s为每个单位内容片段的比特大小。
步骤1405,计算链路l i,j的命中内容片段数量
Figure PCTCN2021136801-appb-000068
Figure PCTCN2021136801-appb-000069
Figure PCTCN2021136801-appb-000070
为发送车辆i的缓存内容片段集合,
Figure PCTCN2021136801-appb-000071
为接收车辆j的申请内容集合。
步骤1406,通过链路l ij的实际可传输内容片段数量
Figure PCTCN2021136801-appb-000072
和命中内容片段数量
Figure PCTCN2021136801-appb-000073
计算链路l i,j的实际接收内容数量;
Figure PCTCN2021136801-appb-000074
步骤十五:利用每个链路的实际接收内容数量计算车辆间联盟合作和车辆个体的效用函数。联盟合作的效用函数为联盟收益函数和开销函数的差值;具体如下:
联盟收益函数为联盟S中所有链路l i,j实际接收到的内容数量,计算公式为:
Figure PCTCN2021136801-appb-000075
开销函数与联盟中V2V链路数成正比,计算公式为:
Figure PCTCN2021136801-appb-000076
|S|-1为联盟S中的链路总数,其中β为常数。个体效用函数
Figure PCTCN2021136801-appb-000077
的计算公式如下:
Figure PCTCN2021136801-appb-000078
其中,δ是略大于1的常数;
Figure PCTCN2021136801-appb-000079
表示发送内容的车辆集合;
Figure PCTCN2021136801-appb-000080
表示接收内容的车辆集合。
步骤十六:将每个发送车辆和每个内容接收车辆,根据联盟合作和车辆个体的效用函数实现车辆间联盟,建立V2V链路实现内容分发。具体联盟过程如下:
首先,随机划分收发车辆,构建初始联盟组合
Figure PCTCN2021136801-appb-000081
并且初始化当前联盟组合
Figure PCTCN2021136801-appb-000082
和初始化迭代指示变量iter=0;
然后,针对发送车辆i,找到当前所属联盟
Figure PCTCN2021136801-appb-000083
以及随机选择当前联盟组合中的其他新联盟
Figure PCTCN2021136801-appb-000084
S m≠S k
判断车辆i相对于旧联盟S k,是否更偏好选择新联盟S m,即是否满足S miS k,如果是,即车辆i离开旧联盟S k并加入新联盟S m,同时更新联盟组合;否则,车辆i继续选择当前联盟组合中的其他联盟进行判断,直到车辆i找到偏好的新联盟或者当前联盟中的所有其他联盟都不被偏好,车辆i的联盟过程终止,继续下一个车辆进行联盟选择。
S miS k的计算公式为:
Figure PCTCN2021136801-appb-000085
Figure PCTCN2021136801-appb-000086
指车辆i在加入新联盟S m后可获得的个体效用值;
Figure PCTCN2021136801-appb-000087
指车辆i还在旧联盟S k中可以获得的个体效用。U(S m∪{i})指车辆i在加入新联盟S m后,联盟S m可获得的联盟效用值;U(S k\{i})指车辆i还在旧联盟S k中时,联盟S k可获得的联盟效用值。
联盟组合的更新公式为:
Figure PCTCN2021136801-appb-000088
最后,判断更新的联盟组合
Figure PCTCN2021136801-appb-000089
是否收敛于纳什均衡
Figure PCTCN2021136801-appb-000090
如果是,得到最优的车辆联盟结构;否则,选择下一个发送车辆重复上述过程,直至满足联盟组合的要求。
步骤十七:每个传输时隙T t执行V2I或V2V的毫米波波束资源联合优化,直到数据传输完毕或者达到调度周期T s=NT t
步骤十八、利用静态场景下基站-用户协同和发射功率的最优结果,或动态场景下车辆联盟的最优结果,实现静态和动态相结合的毫米波波束资源分配。本申请的优点在于:
1)、一种静态和动态相结合的毫米波波束资源分配与优化方法,与现有技术相比,重点解决了静态和动态场景下的毫米波波束资源分配效率低的瓶颈难题,保证新型业务的成功传输,因此有着巨大的应用潜力和实际意义。
2)、一种静态和动态相结合的毫米波波束资源分配与优化方法,针对静态热点地区场景,重点解决了毫米波基站差异化回程能力无法感知、波束干扰分析不完备、系统计算复杂度高这几个核心难题。
3)、一种静态和动态相结合的毫米波波束资源分配与优化方法,针对多基站多用户的协同构建了基于匹配理论的分布式框架,能够在保证低复杂度的前提下缓解回程链路压力、消除波束间干扰,从而提升网络吞吐量。
4)、一种静态和动态相结合的毫米波波束资源分配与优化方法,构建了能够动态感知 毫米波基站回程能力和波束间干扰的双边匹配效用函数,实现毫米波多波束的高效分配。
5)、一种静态和动态相结合的毫米波波束资源分配与优化方法,基于凸优化理论推导得到最优发射功率的理论凸上界和凸下界,从而实现网络总体性能的最优化。
6)、一种静态和动态相结合的毫米波波束资源分配与优化方法,针对动态车辆网络场景,在V2I阶段,提出一种低复杂度的波束选择方案实现基站覆盖范围内的车辆内容初步缓存。在V2V阶段,提出一种基于联盟博弈的收发车辆合作方法,能够实现车辆联盟和车辆个体效用的双重保证,实现差异化内容的高效分发。
附图说明
为了更清楚地说明本申请实施例和现有技术的技术方案,下面对实施例和现有技术中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图仅仅是本申请的一些实施例,本领域普通技术人员来讲还可以根据这些附图获得其他的附图。
图1为本申请实施例提供的一种静态和动态相结合的毫米波波束资源分配与优化方法流程图;
图2为本申请实施例构建的两层毫米波异构蜂窝网络的下行传输场景示意图;
图3中的(a)为本申请实施例基站内干扰上报给基站的示意图;
图3中的(b)为本申请实施例基站间干扰上报给基站的示意图;
图4为本申请实施例利用接入链路的传输速率计算用户的效用函数和基站的效用函数示意图;
图5为本申请实施例构建动态车辆网络的内容分发传输场景示意图;
图6中的(a)为本申请实施例毫米波V2V通信链路中波束对准(有效通信链路)示意图;
图6中的(b)为本申请实施例毫米波V2V通信链路中波束未完全对准(邻近干扰链路)示意图;
图7为本申请实施例毫米波V2V通信链路相关时间示意图;
图8为本申请实施例提供的一种毫米波波束资源分配与优化方法的流程图;
图9为本申请实施例提供的另一种毫米波波束资源分配与优化方法的流程图;
图10为本申请实施例提供的另一种毫米波波束资源分配与优化方法的流程图。
具体实施方式
为使本申请的目的、技术方案、及优点更加清楚明白,以下参照附图并举实施例,对本申请进一步详细说明。显然,所描述的实施例仅仅是本申请一部分实施例,而不是全部的实施例。本领域普通技术人员基于本申请中的实施例所获得的所有其他实施例,都属于本申请保护的范围。为了便于本领域普通技术人员理解和实施本申请,下面结合附图对本申请作进一步的详细和深入描述。
本申请实施例一种静态和动态相结合的毫米波波束资源分配与优化方法,针对静态热点地区场景,运用基于匹配理论的分布式多基站多用户协同框架,构建了能够动态感知毫米波基站回程能力和基站内、基站间波束干扰的双边匹配效用函数,在计算复杂度低的前提下实现毫米波多波束的高效分配。利用毫米波波束发射功率的优化策略,推导得到最优 发射功率的理论凸上界和凸下界,实现网络吞吐量的最优化;针对动态车辆网络场景,在V2I阶段,提出一种低复杂度的波束选择方案实现基站覆盖范围内的车辆内容初步缓存。在V2V阶段,提出一种基于联盟博弈的收发车辆合作方法,实现差异化内容的高效分发。
针对静态热点地区场景,本申请实施例优化两种不同维度的毫米波波束资源,包括基于匹配理论的分布式多基站-多用户协同决策,主要负责波束的空间维度分配,以及发射功率优化决策,主要负责波束的能量维度分配;针对动态车辆网络场景,在车辆到基础设施(Vehicle to Infrastructure,V2I)和车辆到车辆(Vehicle to Vehicle,V2V)进行毫米波波束分配优化,并分别提出一种V2I阶段的内容分发车辆选择算法和一种基于联盟博弈的收发车辆合作方法,实现毫米波波束的高效分发。本申请实施例中的用户指的是用户使用的终端,其中终端可以是手机、计算机或者平板电脑等设备。如图1所示,具体步骤如下:
步骤一、针对某用户i,判断该用户的移动速度是否小于速度阈值η v,如果是,则转入步骤二,归类到静态场景处理;否则,用户位于车辆内,进入步骤九,归类到动态场景处理。换言之,针对用户i,判断该用户的移动速度是否大于速度阈值η v。如果是,则进入步骤九,构建动态车联网络场景。如果否,则进入步骤二,构建静态热点地区场景。
步骤二、构建两层毫米波异构蜂窝网络的下行传输场景。
如图2所示,场景中密集部署M个毫米波小区基站,用集合
Figure PCTCN2021136801-appb-000091
表示,且每个小区基站具备缓存能力。各个毫米波小区基站通过回程链路连接到宏基站(Macro Cell Base Station,MBS)0,通过接入链路连接到I个用户,表示为用户集合
Figure PCTCN2021136801-appb-000092
其中,m为毫米波小区基站的编号,i为用户的编号,0表示宏基站。
其中,本申请实施例中的毫米波小区基站也可称为小小区基站(Small Cell Base Station,SBS),图2中左侧的虚线框内为静态场景示意图,图2中示例性地示出了静态场景中的SBS1、SBS2、SBS3这3个毫米波基站,实际实现中毫米波基站的数量不限于此。
特别地,接入链路和回程链路工作在不同毫米波频段(带宽分别为B a和B b)来避免接入链路和回程链路的相互干扰。宏基站0与核心网络之间通过有线光纤连接。在每个毫米波小区基站中,随机且均匀分布用户。
例如,图2中SBS1的覆盖范围内的用户包括UE1、UE2和室内(Virtual Reality,VR)/增强现实(Augmented Reality,AR)设备,其中,UE1为处于室外的行人的终端,UE2和室内VR/AR设备为室内的终端。SBS2的覆盖范围内的用户包括UE3和UE4,其中,UE3为处于室外的车辆,UE4为处于室外的VR/AR设备。SBS3的覆盖范围内的用户包括UE4、UE5和高清赛事直播,其中,UE5为处于室外的终端,高清赛事直播表示能够进行高清赛事直播的场地中的直播设备。
本申请实施例中可以从M个毫米波基站中确定一个毫米波基站作为调度基站,通过调度基站执行步骤三至步骤八,从而确定每个毫米波基站匹配的用户。
步骤三、针对当前次迭代,计算用户i与基站m的有效天线增益、基站内干扰和基站间干扰,并上报给基站m。
首先分析毫米波通信模型,为保证有效的数据传输,毫米波通信均采用定向天线,且采用扇形天线模型。设每个毫米波小区基站的天线都具有相同的波束模型
Figure PCTCN2021136801-appb-000093
包括发射天线的主瓣增益
Figure PCTCN2021136801-appb-000094
旁瓣增益
Figure PCTCN2021136801-appb-000095
和主瓣波束宽度θ t。同理,
Figure PCTCN2021136801-appb-000096
代表每个用户的波束模型,包括接收天线的主瓣增益
Figure PCTCN2021136801-appb-000097
旁瓣增益
Figure PCTCN2021136801-appb-000098
和主瓣波束宽度θ r。因此,基站m到用户i的有效天线增益,表示为:
Figure PCTCN2021136801-appb-000099
Figure PCTCN2021136801-appb-000100
为基站发射天线的主瓣增益,
Figure PCTCN2021136801-appb-000101
为用户接收天线的主瓣增益。
针对波束干扰问题,本申请实施例对多波束并行传输过程中的基站内干扰和基站间干扰两种情况进行建模,相较于现有大部分技术方法仅考虑其中一种干扰的情况,本申请实施例对毫米波波束干扰进行完备的数学分析,如图3所示。参见图3,图3中的(a)包括毫米波基站SBSm和两个终端UEi和UEi′,由于SBSm针对UEi和UEi′的发射波束之间存在部分重合,导致这两个发射波束之间存在干扰,这种同一基站的波束间的干扰为基站内干扰。图3中的(b)包括两个毫米波基站SBSm和SBSm′,以及两个终端UEi和UEi′,且SBSm与UEi′连接,SBSm′与UEi和UEi′分别连接。此时UEi对于SBSm和SBSm′的接收波束之间存在部分重合,导致这两个接收波束之间存在干扰,这种不同基站的波束间的干扰为基站间干扰。
一方面,基站内干扰增益由不同发射波束主瓣间的相互交叠引起,表示为:
Figure PCTCN2021136801-appb-000102
Figure PCTCN2021136801-appb-000103
表示发射波束
Figure PCTCN2021136801-appb-000104
Figure PCTCN2021136801-appb-000105
主瓣间的重叠角度,
Figure PCTCN2021136801-appb-000106
θ t表示基站发射天线的主瓣波束宽度;
Figure PCTCN2021136801-appb-000107
表示发射波束
Figure PCTCN2021136801-appb-000108
相对基准方向m→i的偏移角度,
Figure PCTCN2021136801-appb-000109
表示发射波束
Figure PCTCN2021136801-appb-000110
相对基准方向m→i的偏移角度;
Figure PCTCN2021136801-appb-000111
为基站发射天线的旁瓣增益。
Figure PCTCN2021136801-appb-000112
Figure PCTCN2021136801-appb-000113
分别表示同一个毫米波基站的不同的发射波束,
Figure PCTCN2021136801-appb-000114
表示毫米波基站的发射波束
Figure PCTCN2021136801-appb-000115
对发射波束
Figure PCTCN2021136801-appb-000116
的干扰增益。
另一方面,用户i接收来自其他基站的基站间干扰,其中干扰发射波束
Figure PCTCN2021136801-appb-000117
和参考接收波束
Figure PCTCN2021136801-appb-000118
分别指向对方的主瓣或旁瓣,即干扰发射波束
Figure PCTCN2021136801-appb-000119
为其他基站m′向其他用户i′的发射波束,参考接收波束
Figure PCTCN2021136801-appb-000120
为用户i针对基站m的接收波束,干扰发射波束
Figure PCTCN2021136801-appb-000121
指向用户i的主瓣或旁瓣,参考接收波束
Figure PCTCN2021136801-appb-000122
指向其他基站m′的主瓣或旁瓣。毫米波基站m可通过公式(2)计算干扰发射波束
Figure PCTCN2021136801-appb-000123
对参考接收波束
Figure PCTCN2021136801-appb-000124
的干扰,因此基站间干扰增益
Figure PCTCN2021136801-appb-000125
表示为:
Figure PCTCN2021136801-appb-000126
Figure PCTCN2021136801-appb-000127
为干扰发射波束
Figure PCTCN2021136801-appb-000128
相对于基准方向m′→i的波束偏移角度;
Figure PCTCN2021136801-appb-000129
为参考接收波束
Figure PCTCN2021136801-appb-000130
相对于基准方向m′→i的波束偏移角度。
步骤四、利用有效天线增益、基站内干扰和基站间干扰,计算毫米波小区基站m和用户i接入链路的传输速率,进而分别构建各自的匹配效用函数U m(i)和U i(m)
针对毫米波信道传输,基站m到用户i的大规模路径衰落增益L m,i可表示为:
Figure PCTCN2021136801-appb-000131
其中,
Figure PCTCN2021136801-appb-000132
为视距链路(line-of-sigh,LoS)的路损大小,
Figure PCTCN2021136801-appb-000133
为视距链路(line-of-sigh,LoS)的通信概率;非视距链路(non-line-of-sigh,LoS)同理,
Figure PCTCN2021136801-appb-000134
为非视距链路是路损大小,且非视距链路的通信概率
Figure PCTCN2021136801-appb-000135
基站m到用户i的小尺度阴影衰落的增益表示为h m,i,且h m,i是一个服从均值为一的指数分布的随机变量。P m,i代表基站m发射给用户i的发射功率,噪声N 0为高斯白噪声,毫米波总带宽为B,可得到噪声总功率为P N=BN 0。由香农公式,基站m和用户i的接入链路的传输速率R m,i可表示为:
R m,i=B alog 2(1+ γ m,i)      (3)
B a表示接入链路的带宽。γ m,i为基站m到用户i的信干噪比;
Figure PCTCN2021136801-appb-000136
G m,i指基站m到用户i的有效天线增益。
Figure PCTCN2021136801-appb-000137
是路径损耗增益L m,i的倒数。P m,i′指基站m发射给用户i′的发射功率。
Figure PCTCN2021136801-appb-000138
指发射波束
Figure PCTCN2021136801-appb-000139
对发射波束
Figure PCTCN2021136801-appb-000140
造成的基站内干扰增益。P m′,i′指基站m′发射给用户i′的发射功率。
Figure PCTCN2021136801-appb-000141
指干扰发射波束
Figure PCTCN2021136801-appb-000142
对参考接收波束
Figure PCTCN2021136801-appb-000143
造成的基站间干扰增益。h m',i指干扰发射基站m'到参考接收用户i的小尺度阴影衰落增益。L m',i指干扰发射基站m'到参考接收用户i的大尺度路径损耗增益;
Figure PCTCN2021136801-appb-000144
是L m',i的倒数。P N指加性高斯白噪声功率。
式(4)的分子代表用户i的接收功率;分母三项依次为用户i受到的基站内干扰、基站间干扰和加性高斯白噪声(AWGN)功率。
即,
Figure PCTCN2021136801-appb-000145
为用户i受到的基站内干扰,
Figure PCTCN2021136801-appb-000146
为基站间干扰,P N为加性高斯白噪声(AWGN)功率。
同理,将回程链路的带宽B b均匀分配给M个毫米波小区基站,则可以得到宏基站0 到毫米波小区基站m的传输速率R m
最后,利用接入链路的传输速率计算用户i的效用函数U i(m)和基站m的效用函数U m(i)。
如图4所示,图4中示出了M个毫米波基站和N个用户,对每个用户i,在与基站协同时需要考虑接入链路质量和回程链路能力的折中问题:即与最高传输速率的基站协同,还是与最大回程能力的基站协同。基于此,定义候选接入链路m→i对基站m的回程占用率为
Figure PCTCN2021136801-appb-000147
代表该接入链路对当前基站带来的回程压力。因此,用户i的效用函数U i(m)表示为:
Figure PCTCN2021136801-appb-000148
传输速率。即U i(m)表示终端i与基站m通信时终端的匹配效用函数,R m,i表示基站m和终端i之间建立的下行通信连接的传输速率,R m表示基站m的回程链路的传输速率。
对每个基站m,需要综合考虑协同用户i的接入速率以及用户i对网络中其他用户造成的潜在干扰。此外,对回程链路,U m(i)通过感知基站差异化的回程能力来优化回程资源利用率。因此,基站m的效用函数U m(i)如下:
Figure PCTCN2021136801-appb-000149
U m(i)表示终端i与基站m通信时基站的匹配效用函数,ω 1为干扰项的权重因子,ω 2为回程占用率的权重因子,且ω 1≥0,ω 2≥0。I m,i为用户i对其他用户造成的总干扰
Figure PCTCN2021136801-appb-000150
即I m,i表示终端i与基站m通信时对其他终端的通信造成的干扰;c m,v(i)表示用户i当前的申请内容片段,即c m,v(i)表示终端i当前申请
Figure PCTCN2021136801-appb-000151
示该接入链路m→i占用回程资源所受到的效用惩罚。
步骤五、构建基站m和每个用户的匹配效用函数,形成基站m的匹配偏好列表List m;构建用户i和每个基站的匹配效用函数,形成用户i的匹配偏好列表List i
列表List m中存储的是按效用函数从大到小排列的各用户序号;列表List i中存储的是按效用函数从大到小排列的各基站序号。同理,对每个基站都能得到各基站的匹配偏好列表,对每个用户也能得到各用户的匹配偏好列表。
如果用户i相较于小区基站m更偏好小区基站m′,则该匹配偏好关系可表示为
Figure PCTCN2021136801-appb-000152
相似地,基站m关于用户集合
Figure PCTCN2021136801-appb-000153
的匹配偏好关系可表示为
Figure PCTCN2021136801-appb-000154
步骤六、根据每个基站和每个用户的匹配偏好列表对多基站-多用户进行匹配,得到基站-用户协同矩阵X,建立下行通信连接。具体匹配过程如下:
首先,针对未完全连接的用户i,向匹配偏好列表List i中效用函数值最高的小区基站m发送接入请求
Figure PCTCN2021136801-appb-000155
并将该基站m的序号从列表List i中清除。每个用户可连接一定数量的基站,未完全连接的用户指的是在当前连接的基础上仍可以连接其他基站的终端。
同理,每个未完全连接的用户都从各自的匹配偏好列表List i中选择效用函数值最高的小区基站发送请求操作。
然后,基站m将它的所有请求用户加入集合
Figure PCTCN2021136801-appb-000156
判断集合
Figure PCTCN2021136801-appb-000157
中的用户数量是否小于等于基站m的配额Q m,如果是,基站m接受集合
Figure PCTCN2021136801-appb-000158
中所有的用户请求;否则,基站m只接受Q m个函数值最高的用户并且拒绝其他申请用户。被拒绝的用户仍属于未完全连接,重复上述过程,继续进行下一轮迭代,即针对每个被拒绝的用户,返回上述向匹配偏好列表List i中效用函数值最高的小区基站m发送接入请求
Figure PCTCN2021136801-appb-000159
的步骤,直至最终不存在用户被拒绝时,匹配过程结束,得到最终优化的基站-用户协同矩阵X。调度基站计算的基站-用户协同矩阵X表示各毫米波基站与用户之间的匹配关系,调度基站可向集合m中的各毫米波基站发送基站-用户协同矩阵X,每个毫米波基站根据基站-用户协同矩阵X,确定自身匹配的用户,并与确定的用户建立下行通信连接。
极端情况:用户一直申请、却一直被拒绝。考虑到用户i申请一个基站就删除List i中对应的该基站序号,直到List i变空,则用户i停止申请、不再参与循环,同时用户i未与任何基站连接。这种情况因为基站有配额限制Q m,发射波束有限,竞争失败的用户得不到与基站通信的机会。
即调度基站针对每个未完全连接的用户i,将匹配偏好列表List i中效用函数值最高的毫米波基站,作为用户i的候选匹配基站。然后调度基站针对每个毫米波基站m,从将其作为候选匹配基站的用户中,按照用户在毫米波基站m的匹配偏好列表List m中的排列顺序,选择预设数量的用户作为该毫米波基站匹配的用户。预设数量为毫米波基站m最多能够连接的用户数量,即毫米波基站m的配额Q m
步骤七、基于基站-用户连接状态,基于凸优化理论优化基站的毫米波波束发射功率,生成当前次迭代的发射功率最优解。
接下来建立毫米波小区基站侧的缓存模型,考虑到虚拟现实等新型业务对高速率的需求,本申请实施例引入缓存模型来有效缓解毫米波小区基站的回程压力。
假设用户从内容库
Figure PCTCN2021136801-appb-000160
中请求高清视频、文件等内容,小区基站m具有缓存能力S m(1<S m<F)。当用户i申请一个内容时,如果基站m已经缓存了该内容,则可以直接通过接入链路快速获取内容;否则,该内容从核心网通过回程链路传输给基站m,再由基站m通过接入链路传输给用户i。设每个内容f的申请概率p f服从齐普夫(Zipf)分布,即
Figure PCTCN2021136801-appb-000161
其中δ代表文件f的流行度指数。令c m,f∈{1,0},
Figure PCTCN2021136801-appb-000162
代表基站m对内容f的缓存变量,且c m,f∈{1,0},代表基站m对是否缓存内容f,即c m,f=1表示基站m缓存有内容f,c m,f=0表示基站m未缓存内容f;则基站m的缓存命中概率可表示为
Figure PCTCN2021136801-appb-000163
特别地,本申请实施例考虑基站无法先验探知缓存内容流行度的最差情况,因此采用随机缓存策略。即毫米波基站可随机地从内容库中选择并缓存内容。
针对高流量密度热点地区的容量提升瓶颈难题,有着巨大的应用潜力和实际意义。目标是通过联合优化多基站多用户协同策略和多波束发射功率来最大化网络吞吐量,资源分配问题可用数学表示为:
Figure PCTCN2021136801-appb-000164
x m,i为基站-用户协同变量,x m,i∈X;X指基站-用户协同优化矩阵,表征基站和用户是否建立通信连接;矩阵X的维度为M*I。P指发射功率优化矩阵,表征基站的发射波束到用户的发射功率大小;矩阵P的维度为M*I;发射功率变量P m,i∈P。
其中,约束条件C1保证每个波束的发射功率非负值,且每个基站的总发射功率不超过功率最大值P max;约束条件C2保证每个基站传输未缓存内容占用的回程容量不超过该基站的回程链路能力;约束条件C3保证用户i的接入速率不小于最低速率需求
Figure PCTCN2021136801-appb-000165
然后,由于接入带宽B a为常数,利用连续凸逼近求解法,对频谱效率r m,i=R m,i/B a进行讨论,得到频谱效率r m,i的凸上界和凸下界。通过连续凸逼近方法将复杂非凸的功率优化问题转化为凸问题,得到最优发射功率的理论凸上界和凸下界,并证明所求解理论可收敛。即调度基站可以基于凸优化理论确定各毫米波基站的各发射波束的发射功率。
具体过程如下:
a)、利用用户i与基站m的有效天线增益、基站内干扰和基站间干扰,计算中间函数:
Figure PCTCN2021136801-appb-000166
Figure PCTCN2021136801-appb-000167
是指发射波束
Figure PCTCN2021136801-appb-000168
受到的基站内干扰和基站间干扰的总和;
Figure PCTCN2021136801-appb-000169
是基站m到用户i的接收功率和
Figure PCTCN2021136801-appb-000170
的和。
b)、然后利用两中间函数简单的代数运算,计算频谱效率r m,i
Figure PCTCN2021136801-appb-000171
c)、分别利用泰勒展开公式和对数函数性质对两个函数进行凸逼近处理。具体为:
对函数
Figure PCTCN2021136801-appb-000172
考虑到当x>0时的对数函数log(x)为凸函数,利用泰勒展开公式近似为一阶泰勒函数的形式:
Figure PCTCN2021136801-appb-000173
对函数
Figure PCTCN2021136801-appb-000174
利用对数函数性质
Figure PCTCN2021136801-appb-000175
近似为:
Figure PCTCN2021136801-appb-000176
d)、基于上述两个函数的凸逼近处理结果,得到频谱效率r m,i的凸上界
Figure PCTCN2021136801-appb-000177
表示为:
Figure PCTCN2021136801-appb-000178
相反地,将上述两个函数分别利用对数函数性质和泰勒展开的操作逼近,得到r m,i的凸下界,表示为:
Figure PCTCN2021136801-appb-000179
最后,利用边界在t∈{1,...,T}的时间内迭代生成发射功率的最优解序列{P (t)}。
通过将凸边界
Figure PCTCN2021136801-appb-000180
Figure PCTCN2021136801-appb-000181
带入问题P1,则问题P1可进一步转化为以下形式:
Figure PCTCN2021136801-appb-000182
因此,非凸问题P1的求解问题可以转化为最大化凸问题P2的下界,得到以下结论:对问题P2的一个初始可行点{P (0)},在t∈{1,...,T}的时间内迭代生成问题P2的最优解序列{P (t)},则{P (t)}最终会收敛到卡罗需-库恩-塔克条件(Karush–Kuhn–Tucker conditions,KKT)点。证明:设P (t)和P (t+1)为问题P2的可行点,且有:
Figure PCTCN2021136801-appb-000183
即P (t+1)是优于P (t)的可行点。由于解序列{P (t)}是有界的,根据柯西定理,存在子序列
Figure PCTCN2021136801-appb-000184
收敛到有限点
Figure PCTCN2021136801-appb-000185
Figure PCTCN2021136801-appb-000186
因此,对每个时间t都存在ν使得以下条件成立:
Figure PCTCN2021136801-appb-000187
因此,可证明
Figure PCTCN2021136801-appb-000188
Figure PCTCN2021136801-appb-000189
为一个KKT点。
步骤八、基于当前次迭代的基站-用户协同矩阵X和发射功率计算网络吞吐量,并判断当前次迭代的网络吞吐量与前一次迭代的网络吞吐量的差值是否小于收敛阈值,如果是,即网络吞吐量收敛,当前次迭代的基站-用户协同矩阵和发射功率作为最优解,毫米波基站进行下行传输直到所有数据传输完毕;否则,返回步骤三重新运行下一次迭代。
首先,初始化毫米波小区基站侧的波束发射功率P m,i为基站最大发射功率的平均值P max/Q m,得到用户协同矩阵X (0)的初始值。进一步,通过功率优化策略得到最优发射功率解P *;基于P *更新基站侧和用户侧的匹配偏好列表List m和List i。其中,匹配偏好列表List m基于最新的接入链路速率
Figure PCTCN2021136801-appb-000190
对所有候选用户的效用值进行降序排列,匹配偏好列表List i同理。即,匹配偏好列表List i基于最新的接入链路速率对所有的候选基站的效用值进行降序排列。根据更新的匹配偏好列表,得到最新匹配结果。最后迭代至收敛为止。
步骤九、构建动态车辆网络的内容分发传输场景。
如图5所示,首先建立面向动态车辆网络的系统模型,包括车辆移动模型、毫米波信道模型和内容分发计算。可以看出,图5中右侧的虚线框中是动态移动场景的示意图,该场景中包括:N个车辆随机分布在长度为L的双车道高速公路路段上,各个车辆以不同的速度朝同一方向行驶,车辆集合
Figure PCTCN2021136801-appb-000191
且车辆具备缓存能力,并遵从高速公路移动模型(freeway mobility model,FMM)。该场景中还包括宏基站和毫米波基站,宏基站与每个毫米波基站通信连接,每个毫米波基站还可以与一个或多个车辆通信连接。
针对每个车辆,内容分发过程可以分为两个阶段:V2I阶段和V2V阶段。在V2I阶段,车辆处于毫米波基站的覆盖范围内,接收到流行内容的部分片段;考虑到天线阵列等硬件设备限制,每个基站最多支持Q 0个波束同时传输。在V2V阶段,毫米波V2V通信链路如图6所示,图6中的(a)和图6中的(b)分别表示和波束对准(有效通信链路)和波束未完全对准(邻近干扰链路)的情况。其中,图6中的(a)包括发送车辆i和接收车辆j,且发送车辆i和发送波束与接收车辆j的接收波束对准。图6中的(b)包括发送车辆i和 接收车辆j,且发送车辆i和发送波束与接收车辆j的接收波束未完全对准。
车辆在毫米波小区覆盖范围外,车辆间通过合作来进行内容共享,从而获取尽可能多的剩余所需内容片段;每个车辆采用半双工模式通信,最多支持Q v个波束同时传输,或支持单个毫米波接收波束。即每个车辆最多可以与Q v个设备(包括毫米波基站和其他车辆)同时通过毫米波波束通信,Q v大于等于1。
此外,假设流行内容集合为
Figure PCTCN2021136801-appb-000192
考虑到差异化的业务需求,不同内容的大小不同,若将各个内容切分为大小为s比特的单位内容片段,则内容c的大小可表示为sD c比特,其中D c表示单位内容片段的个数。
本申请实施例考虑双车道无交叉口的高速公路机动模型(freeway mobility model,FMM)。所有车辆初始随机分布在车道上,并以初始速度v i行驶。对于同一车道上的相邻两辆车,限制车辆间距离范围为[d min,d max],其中为d min最小安全距离,d max为最大间隔距离。对每辆车
Figure PCTCN2021136801-appb-000193
限制速度范围为[v min,v max]。每个车辆的速度选择是独立的,在每个时隙下基于加速度a以概率p随机选择加速或减速。为了保证车辆行驶的安全,本申请实施例暂不考虑车辆超车的情况。因此,为防止超车行为的发生,对于同一车道上的车辆i和j,给出车辆i的行为约束条件如下:
1)若d i,j≤d min,车辆i减速至v i(t+1)=v min
2)若d i,j≥d min,车辆i加速至v i(t+1)=v max
步骤十、判断车辆是否在基站覆盖范围内,如果是,转到步骤十一;否则,转到步骤十三。本申请实施例中,每个毫米波基站可通过步骤十一和十二选择自身匹配的车辆,并将其作为发送车辆,与发送车辆通信。接收车辆可通过步骤十三至十六选择自身所处的车辆联盟,并与车辆联盟中的发送车辆通信。
步骤十一:计算车辆i与毫米波小区基站m的信道状态,并上报给毫米波小区基站m。
信道状态包括有效天线增益和基站内干扰。
步骤十二:在V2I阶段每个调度时隙T t下,每个基站挑选各自的最优车辆来传输下行数据,进入步骤十七。具体过程为:
首先,在V2I阶段,计算车辆i接收到基站m的功率为:
Figure PCTCN2021136801-appb-000194
其中,
Figure PCTCN2021136801-appb-000195
为基站m发射给车辆i的发射功率,即
Figure PCTCN2021136801-appb-000196
表示毫米波基站m向车辆i发送信号的发射功率。
Figure PCTCN2021136801-appb-000197
表示车辆i接收毫米波基站m发送的毫米波信号的接收功率,G m,i表示毫米波基站m到车辆i的有效天线增益,h m,i表示毫米波基站m到车辆i的小尺度阴影衰落增益,
Figure PCTCN2021136801-appb-000198
为毫米波基站m到车辆i的路径损耗增益L m,i的倒数。
然后,同理计算基站m接收到覆盖范围内所有车辆的接收功率,并按功率从大到小对各候选车辆进行排序,选择与前Q 0个车辆进行连接。
最后,在当前调度时隙T t内,遍历所有已连接车辆,找到传输速率最高的已连接车辆i。判断当前未连接车辆i′替换掉车辆i后的网络吞吐量是否增加,如果是,将车辆i′替换掉车辆i,否则,将当前未连接车辆i′丢弃,继续进行下一个未连接车辆的替换工作。
依次将所有未连接车辆均替换完毕后,完成基站m的所有最优连接车辆选择。毫米波基站选择车辆后,将选择的车辆作为发送车辆,与每个发送车辆之间建立V2I链路,并基 于V2I链路向发送车辆传输内容片段。
步骤十三:每个调度时隙T t下,确定与各基站连接的车辆作为内容发送车辆,将未连接的车辆作为接收车辆。
其中,本申请实施例中的内容发送车辆也可简称为发送车辆或发端车辆。
步骤十四:利用每个接收车辆和每个内容发送车辆形成的链路,计算各链路的实际接收内容数量。其中,本申请实施例中的实际接收内容数量可以称为接收内容片段数量。
针对毫米波V2V通信链路,本申请实施例采用标准对数距离路径损耗模型。
因此,发送车辆i和接收车辆j间的毫米波V2V链路l i,j的传输路径损耗L i,j可表示为:
L i,j[dB]=A+20log 10(f c)+10δ i,jlog(d i,j)
其中,A代表大气衰减值,f c代表毫米波通信的中心载频,δ i,j代表路损指数,d i,j代表车辆i和j的相对距离。此外,有
Figure PCTCN2021136801-appb-000199
其中,毫瓦(milliwatt,mW)和分贝(decibel,dB)都是功率单位,mW表征功率的线性值,dB表征相对值的值。
为实现车辆通信中内容片段的高效分发,本申请实施例从两个角度衡量内容分发效率:1)V2V链路传输能力,即发端车辆实际能传输多少内容片段给接收车辆;2)命中内容数量,即发端车辆中已缓存有的申请内容数量。具体步骤为:
步骤1401,针对发送车辆i和接收车辆j形成的链路l i,j,得到发送车辆i的发射天线增益
Figure PCTCN2021136801-appb-000200
和接收车辆j侧的接收天线增益
Figure PCTCN2021136801-appb-000201
计算有效天线增益G i,j
Figure PCTCN2021136801-appb-000202
步骤1402,计算当前链路l i,j受到来自于同时传输内容的其他链路l i',j'的干扰I i,j
即发送车辆i与接收车辆j形成的链路l i,j受到的来自于同时传输内容片段的发送车辆i′与接收车辆j′形成的链路l i′,j′的干扰I i,j,表示为:
Figure PCTCN2021136801-appb-000203
P t为车辆发射波束的功率;h i',j为干扰发送车辆i′到参考接收车辆j链路的小尺度阴影衰落;L i',j为干扰发送车辆i'到参考接收车辆j链路的大尺度路径损耗;
Figure PCTCN2021136801-appb-000204
为V2V链路间干扰增益,表示为:
Figure PCTCN2021136801-appb-000205
Figure PCTCN2021136801-appb-000206
为基站接收天线的旁瓣增益;
Figure PCTCN2021136801-appb-000207
为干扰发射波束
Figure PCTCN2021136801-appb-000208
相对于基准方向i′→j的波 束偏移角度;
Figure PCTCN2021136801-appb-000209
为参考接收波束
Figure PCTCN2021136801-appb-000210
相对于基准方向i′→j的波束偏移角度。
步骤1403,利用V2V链路间干扰计算发送车辆i和接收车辆j的信干噪比(Signal to Interference plus Noise Ratio,SINR),进一步计算接入链路l i,j的传输速率R i,j
假设所有车辆发射波束的功率均为P t代表。因此,可得到接收车辆j的SINR:
Figure PCTCN2021136801-appb-000211
同理,针对毫米波基站m和车辆i间V2I链路l m,i,可以得到内容分发车辆i的SINR为
Figure PCTCN2021136801-appb-000212
注意到车辆i受到的干扰I m,i来自于毫米波基站的其他同时发射波束带来的基站内干扰。
步骤1404,利用接入链路l i,j的传输速率R i,j,计算链路l i,j的实际可传输内容片段数量
Figure PCTCN2021136801-appb-000213
Figure PCTCN2021136801-appb-000214
t s指链路相关通信时间,即车辆收发波束从对准到失去对准的时间,即发送车辆i的发送波束和接收车辆j的接收波束从对准到失去对准的时间;s为每个单位内容片段的比特大小。
假设在V2V阶段,每次调度周期时长为T s=NT t,即包含N个传输时隙T t。每个传输时隙T t可进一步分为两个阶段:波束对准阶段T A和数据传输阶段T d
对于波束对准阶段T A,假设每辆车已经完成扇区级对准来有效减少波束搜索时间。因此,链路l i,j的波束对准时延T A可表示为
Figure PCTCN2021136801-appb-000215
其中,
Figure PCTCN2021136801-appb-000216
Figure PCTCN2021136801-appb-000217
分别表示车辆i和j的扇区级波束宽度,即
Figure PCTCN2021136801-appb-000218
表示发送车辆的扇区级波束宽度,
Figure PCTCN2021136801-appb-000219
表示接收车辆j的扇区级波束宽度,导频表示导频传输时间T p。综上所述,可以根据香农公式计算出链路l i,j的吞吐量为
Figure PCTCN2021136801-appb-000220
其中,香农公式中的B指的是信道带宽,单位可以为赫兹。
对于数据传输阶段T d,考虑到车辆的高移动性,V2V链路传输能力不仅取决于链路吞 吐量,也取决于当前链路稳定性。因此,接下来讨论链路通信相关时间t s,即车辆收发波束从对准到失去对准的时间。图7中包括发送车辆i和接收车辆j,且B和D表示发送车辆的位置,A、C和E表示接收车辆j的位置,且α表示接收车辆j的接收波束方向与接收车辆j的速度方向之间的夹角,β表示发送车辆i与接收车辆j之间的相对位置与接收车辆j的速度方向之间的夹角,
Figure PCTCN2021136801-appb-000221
表示发送车辆i的发射天线的主瓣波束宽度的一半。以图7为例(x j>x i,y j>y i,v j>v i),假设接收车辆j从A移动至C,发送车辆i从B移动至D,A和B是收发车辆i和j的初始位置,即接收车辆j在位置A处的接收波束与发送车辆i在位置B的发送波束对准。C和D是收发波束失去对准的位置,即接收车辆j移动至位置C时的接收波束与发送车辆j移动至位置D时的发送波束失去对准。E是假设车辆i和j的相对速度为0的虚拟接收车辆映射位置,即假设接收车辆j与发送车辆i的速度相同,发送车辆i从位置B移动至位置D时,接收车辆j从位置A移动至位置E。则可以得到
Figure PCTCN2021136801-appb-000222
其中,长度l CE可基于下述公式得到:
Figure PCTCN2021136801-appb-000223
其中,角度β由收发车辆的相对位置和相对速度所决定。
步骤1405,计算链路l i,j的命中内容片段数量
Figure PCTCN2021136801-appb-000224
考虑到业务差异化需求,内容分发效率也取决于内容片段的多样性。内容片段多样性越高,命中内容片段数量也越大。设
Figure PCTCN2021136801-appb-000225
Figure PCTCN2021136801-appb-000226
分别为发射车辆i的缓存内容片段集合和接收车辆j的申请内容集合,则链路l i,j的命中内容片段数量可定义为:
Figure PCTCN2021136801-appb-000227
步骤1406,通过链路l i,j的实际可传输内容片段数量
Figure PCTCN2021136801-appb-000228
和命中内容片段数量
Figure PCTCN2021136801-appb-000229
计算链路l i,j的实际接收内容数量。
Figure PCTCN2021136801-appb-000230
其中,
Figure PCTCN2021136801-appb-000231
表示发送车辆i和接收车辆j之间形成的链路l i,j的接收内容片段数量,
Figure PCTCN2021136801-appb-000232
表示l i,j的可传输内容片段数量,
Figure PCTCN2021136801-appb-000233
t s表示发送车辆i的发送波束和接收车辆j的接收波束从对准到失去对准的时间,R i,j表示l i,j的传输速率,s表示每个内容片段的大小,
Figure PCTCN2021136801-appb-000234
表示l i,j的命中内容片段数量,
Figure PCTCN2021136801-appb-000235
C i表示发送车辆i的缓存内容片段集合,C j表示接收车辆j的申请内容片段集合,card表示集合的大小。
步骤十五:利用每个链路的实际接收内容数量计算车辆间联盟合作和车辆个体的效用函数。联盟合作同时存在收益和开销,因此联盟S的效用函数可表示为U(S)=V(S)-C(S)。具体如下:
一方面,为最优化网络整体收益,将一个联盟的收益函数定义为所有链路l i,j实际接收到的内容数量,计算公式为:
Figure PCTCN2021136801-appb-000236
S为车辆联盟的编号,V(S)表示车辆联盟S中各V2V链路对应的接收内容片段数量总和,
Figure PCTCN2021136801-appb-000237
表示车辆联盟S中的发送车辆i与接收车辆j之间建立的V2V链路对应的接收内容片段数量。
另一方面,尽管车辆间通过合作可以提升网络整体收益,但是这些收益可能被建立和维持联盟的开销所削减。具体地,为了建立一条新的链路,车辆会承受一定的能耗损失,同时为其他V2V通信链路带来干扰。因此,有必要建立开销函数,联盟内车辆开销的总和形成了联盟的开销。考虑到联盟S中的链路总数为|S|-1,其中|S|为车辆联盟S包括的车辆数量,β为常数,C(S)表示车辆联盟S中各V2V链路的开销,则开销函数可表示为:
Figure PCTCN2021136801-appb-000238
其次,设计联盟内每个成员车辆的个体效用函数。在车辆间的联盟博弈中,用u i表示车辆的个体效用。如果车辆能够通过联盟获得比独自行动更多的效用,即u i≥u({i}),
Figure PCTCN2021136801-appb-000239
这个博弈结果就是个体理性(individually rational)的。考虑到每个接收车辆希望尽可能接收到多的有效内容片段,定义车辆个体效用
Figure PCTCN2021136801-appb-000240
为:
Figure PCTCN2021136801-appb-000241
其中,δ是略大于1的常数;对于发射车辆
Figure PCTCN2021136801-appb-000242
来说,不合作时的效用是
Figure PCTCN2021136801-appb-000243
合作时的效用
Figure PCTCN2021136801-appb-000244
意味着在联盟中能够获得比单独行动更多的效用;对于接收车辆来说,不合作时无内容接收,则获得的效用为0,在加入一个联盟后,给予效用
Figure PCTCN2021136801-appb-000245
因此接收车辆在联盟中能够获得比单独行动更多的效用。
步骤十六:将每个发送车辆和每个接收车辆,根据联盟合作和车辆个体的效用函数实现车辆间联盟,建立V2V链路实现内容分发。具体联盟过程如下:
首先,随机划分收发车辆,构建初始联盟组合
Figure PCTCN2021136801-appb-000246
并且初始化当前联盟组合
Figure PCTCN2021136801-appb-000247
和初始化迭代指示变量iter=0。即随机地确定每个接收车辆所属的初始车辆联盟,每个车辆联盟包括一个发送车辆以及一个或多个接收车辆。
然后,针对接收车辆j,找到当前所属联盟
Figure PCTCN2021136801-appb-000248
以及随机选择当前联盟组合中的其他新联盟
Figure PCTCN2021136801-appb-000249
S m≠S k
针对每个接收车辆,判断接收车辆相对于旧联盟S k,是否更偏好选择新联盟S m,即是否满足S miS k,如果是,即接收车辆离开旧联盟S k并加入新联盟S m,同时更新联盟组合;否则,接收车辆继续选择当前联盟组合中的其他联盟进行判断,直到接收车辆找到偏好的新联盟或者当前联盟中的所有其他联盟都不被偏好,接收车辆的联盟过程终止,继续下一个接收车辆进行联盟选择。
S miS k的计算公式为:
Figure PCTCN2021136801-appb-000250
Figure PCTCN2021136801-appb-000251
指接收车辆i在加入新联盟S m后可获得的个体效用值;
Figure PCTCN2021136801-appb-000252
指接收车辆i还在旧联盟S k中可以获得的个体效用。U(S m∪{i})指接收车辆i在加入新联盟S m后,联盟S m可获得的联盟效用值;U( S k\{i})指接收车辆i还在旧联盟S k中时,联盟S k可获得的联盟效用值。联盟组合的更新公式为:
Figure PCTCN2021136801-appb-000253
最后,判断更新的联盟组合
Figure PCTCN2021136801-appb-000254
是否收敛于纳什均衡
Figure PCTCN2021136801-appb-000255
如果是,得到最优的车辆联盟结构;否则,选择下一个发送车辆重复上述过程,直至满足联盟组合的要求。在本申请实施例中,每个接收车辆通过步骤十三至步骤十六确定自身所处的车辆联盟,并与自身的车辆联盟中的发送车辆建立V2V链路,基于V2V链路接收发送车辆发送的内容片段。
步骤十七:每个传输时隙T t执行V2I或V2V的毫米波波束资源联合优化,直到数据传输完毕或者达到调度周期T s=NT t
即在动态场景下,在V2I阶段,每个毫米波基站选择自身匹配的发送车辆,通过V2I链路向发送车辆传输内容片段,直至数据传输完毕或者当前调度周期结束。以及V2V阶段,每个接收车辆选择自身匹配的发送车辆,通过V2V链路接收发送车辆发送的内容片段,直至数据传输完毕或者当前调度周期结束。
步骤十八、利用静态场景下基站-用户协同和发射功率的最优结果,或动态场景下车辆联盟的最优结果,实现静态和动态相结合的毫米波波束资源分配。
针对动态车辆网络场景,现有的方法仅能实现V2I模式下的毫米波资源分配,或者V2V模式下的毫米波资源分配。而在V2I模式下,基站在每个时隙能够连接的车辆有限,而且车辆驶离基站覆盖范围时无法与基站通信,使得向车辆分发内容的效率低。在V2V模型下,车辆从周围车辆中获取内容,使得车辆能够获取的内容存在局限性。
为了解决上述问题,本申请实施例提供了一种毫米波波束资源分配与优化方法,该方法应用于车辆,如图8所示,该方法包括:
S801、当车辆处于毫米波基站覆盖范围内时,若车辆在当前时隙与毫米波基站连接,则车辆作为发送车辆通过V2I链路接收毫米波基站通过波束发送的内容片段。其中内容片段为对流行内容集合中的内容进行切分后得到的。当前时隙指的是当前调度时隙。
一种实现方式中,毫米波基站可通过上述步骤十一和十二的方式确定与自身匹配的发送车辆,并向发送车辆发送第一通知消息。第一通知消息用于通知车辆与毫米波基站匹配。当车辆接收到第一通知消息时,确定自身为发送车辆,与毫米波基站建立V2I链路,并将接收波束对准毫米波基站,以通过V2I链路接收毫米波基站通过发送波束发送的内容片段。
可选的,车辆在确定自身为发送车辆后,还可以广播第二通知信息,以将自身为发送车辆的身份通知给其他车辆。
S802、当车辆行驶出毫米波基站的覆盖范围,或车辆未驶出毫米波基站的覆盖范围,但在当前时隙未与毫米波基站连接的情况下,车辆作为接收车辆加入目标车辆联盟,通过V2V链路接收目标车辆联盟中的发送车辆发送的内容片段。
一种实现方式中,车辆若在当前时隙未接收到基站发送的第一通知消息,且接收到其他车辆发送的第二通知消息时,确定自身为接收车辆,并确定发送第二通知消息的车辆为 发送车辆。其中,每个车辆联盟包括一个发送车辆和一个或多个接收车辆。接收车辆可从所在车辆联盟的发送车辆中获取发送车辆缓存的内容片段。
本申请实施例提供的毫米波波束资源分配与优化方法,车辆可以在与基站连接时,从基站处获取内容片段,如果未与基站连接,可以从所属车辆联盟的发送车辆处获取内容片段,从而提高车辆获取内容片段的效率。即在动态场景中,每个时隙下基站可与选择的发送车辆通信,并向发送车辆传输内容片段,同时,未与基站连接的接收车辆,可通过车辆联盟的形式,从发送车辆中获取内容片段。由此既保证了内容片段的分发效率,又缓解了车辆获取的内容的局限性,实现了差异化内容的高效分发。
在本申请的一个实施例中,上述S802车辆作为接收车辆加入目标车辆联盟的方式,可以实现为以下步骤:
步骤1、计算若接收车辆与初始车辆联盟中的发送车辆之间建立V2V链路,接收车辆能够从该V2V链路接收到的接收内容片段数量。其中,初始车辆联盟为随机选择的一个车辆联盟。计算接收车辆能够从该V2V链路接收到的接收内容片段数量的方式,可参考上述步骤十四中的描述。
步骤2、基于该V2V链路对应的接收内容片段数量,计算若接收车辆加入初始车辆联盟后,初始车辆联盟的合作效用函数值和接收车辆的个体效用函数值。具体计算过程可参考上述公式(20)-公式(22)。步骤2的具体实现方式可参考上述步骤十五中的相关描述。
步骤3、选择一个车辆联盟,并计算若接收车辆加入选择的车辆联盟后,选择的车辆联盟的合作效用函数值和接收车辆的个体效用函数值。其中,步骤3计算合作效用函数值和个体效用函数值的方式与步骤2相同,可参考步骤2的描述。
步骤4、判断选择的车辆联盟的合作效用函数值是否大于初始车辆联盟的合作效用函数值,以及接收车辆加入选择的车辆联盟的个体效用函数值是否大于接收车辆加入初始车辆联盟的个体效用函数值。若均为是,则确定选择的车辆联盟为目标车辆联盟。否则,则返回执行步骤3,直至确定出目标车辆联盟。其中,步骤4的具体实现方式可参考上述步骤十六中的相关描述。
步骤5、与目标车辆联盟中的发送车辆之间建立V2V链路。
采用上述方法,接收车辆在选择车辆联盟时,考虑到车辆联盟的合作效用函数,即考虑到接收车辆能够接收到的内容片段数量以及接收车辆加入车辆联盟后车辆联盟的开销,从而能够在尽量减少车辆联盟的开销的同时,提高接收车辆可接收的内容片段数量。而且接收车辆在选择车辆联盟时,还考虑了接收车辆的个体效用函数,即考虑到接收车辆可接收到的内容片段数量,进一步保证了接收车辆可接收到的内容片段数量。
在本申请实施例中,上述步骤1中计算接收内容片段数量,包括以下步骤:
步骤11、将该V2V链路的发送车辆的发射天线增益和接收车辆的接收天线增益的乘积,作为该V2V链路的有效天线增益。步骤11的具体实现方式可参考上述步骤1401。
步骤12、计算该V2V链路受到的来自于其他V2V链路的干扰。步骤12的具体实现方式可参考上述步骤1402。
步骤13、根据该V2V链路的有效天线增益和该V2V链路受到的来自于其他链路的干 扰,计算该V2V链路的传输速率。步骤13的具体实现方式可参考上述步骤1403。
步骤14、根据该V2V链路的传输速率,计算该V2V链路的接收内容片段数量。其中步骤14的具体实现方式可参考上述步骤1404-步骤1406。
采用上述方法,接收车辆可通过车辆间V2V链路的相互干扰情况,确定V2V链路的传输速率,进而得到接收车辆可基于V2V链路从发送车辆接收的内容片段数量。由于链路间干扰越大,链路的传输速率越慢,链路传输的内容片段数量也就越少,因此后续选择可传输更多内容片段的链路,能够减少建立车辆联盟导致的车辆间通信干扰的情况,提高内容片段的传输效率。
基于相同的发明构思,本申请实施例提供了一种毫米波波束资源分配与优化方法,应用于毫米波基站,如图9所示,该方法包括如下步骤:
S901、确定在当前时隙处于毫米波基站覆盖范围内的多个车辆。
S902、计算若毫米波基站与多个车辆中的每个车辆之间建立V2I链路,各V2I链路的信道状态。其中,信道状态包括毫米波基站到车辆的有效天线增益,有效天线增益基于毫米波基站的基站内干扰确定,基站内干扰包括毫米波基站的不同发射波束之间的干扰。S901和S902的具体实现方式可参考上述步骤十一的描述。
S903、根据各V2I链路的信道状态,从多个车辆中选择预设数量个车辆作为发送车辆。
S904、与每个发送车辆建立V2I链路,并通过V2I链路向每个发送车辆发送内容片段,以使得发送车辆通过V2V链路向与自身属于同一车辆联盟的接收车辆发送内容片段。其中,接收车辆为当前时隙未与毫米波基站连接的车辆。S903和S904的具体实现方式可参考上述步骤十二的描述。
采用上述方法,本申请实施例中毫米波基站在选择发送车辆时,基于基站内干扰进行选择,即尽量最小化选择的发送车辆与毫米波基站之间建立的V2I链路之间的干扰,提高毫米波基站与各车辆的通信速度,提高通信效率。
在本申请的一个实施例中,上述S903根据各V2I链路的信道状态,从多个车辆中选择预设数量个车辆作为发送车辆,包括:
根据每个V2I链路的信道状态,确定毫米波基站覆盖范围内的每个车辆接收毫米波基站发送的毫米波信号的接收功率。其中接收功率的计算方式可参考上述步骤十二中的相关描述。
按照接收功率从大到小的顺序,选择预设数量个车辆作为发送车辆。其中,预设数量为Q 0
针对每个未选择的车辆,判断若利用该车辆替换目标发送车辆,毫米波基站与各发送车辆之间建立的V2I链路所形成的网络的网络吞吐量是否增加。其中,目标发送车辆为接收功率最大的发送车辆。若是,则利用该车辆替换目标发送车辆。若否,则确定该车辆不能替换目标发送车辆。S903的具体实现方式可参考上述步骤十二。
本申请实施例中毫米波基站可以选择接收功率较高的车辆作为发送车辆,并依据网络吞吐量进行发送车辆的替换,从而获得在网络吞吐量最高的情况下毫米波基站匹配的发送车辆。即本申请实施例能够保证发送车辆的接收功率,避免接收功率过小而导致的传输效 率低的问题,而且本申请实施例能够提高网络吞吐量,提高内容片段的分发效率。
基于相同的发明构思,本申请实施例提供了一种毫米波波束资源分配与优化方法,应用于毫米波基站,其中,该毫米波基站位于两层毫米波异构蜂窝网络,异构蜂窝网络包括宏基站和毫米波基站集合,毫米波基站集合中的每个毫米波基站分别通过回程链路连接于宏基站。如图10所示,该方法包括如下步骤:
S1001、获取当前时隙内,毫米波基站集合中每个毫米波基站到信号覆盖范围内的各终端的有效天线增益、各毫米波基站的每个发射波束的发射功率、基站内干扰和基站间干扰。其中,基站内干扰包括同一个毫米波基站的不同发射波束之间的干扰,基站间干扰包括不同毫米波基站的发射波束之间的干扰。其中S1001的具体实现方式可参考上述步骤三。
S1002、根据每个毫米波基站到自身信号覆盖范围内的各终端的有效天线增益、各毫米波基站的每个发射波束的发射功率、基站内干扰和基站间干扰,确定若各毫米波基站与信号覆盖范围内的每个终端之间建立下行通信连接,该下行通信连接对应的毫米波基站的匹配效用函数值以及终端的匹配效用函数值。
其中,终端的匹配效用函数为可参考上述公式(5),基站的匹配效用函数可参考上述公式(6)。其中,S1002的具体实现方式可参考上述步骤四。
S1003、根据各下行通信连接对应的毫米波基站的匹配效用函数值以及终端的匹配效用函数值,确定每个终端针对毫米波基站的匹配偏好列表和每个毫米波基站针对终端的匹配偏好列表。S1003的具体实现方式可参考上述步骤五。其中,终端针对毫米波基站的匹配偏好列表包括毫米波基站集合中所有毫米波基站的序号。毫米波基站的序号在该匹配偏好列表中的排列顺序为终端的匹配效用函数从大到小的排列顺序,终端的每个匹配效用函数对应一个毫米波基站。毫米波基站的匹配偏好列表包括毫米波基站覆盖范围内所有终端的序号。终端的序号在该匹配偏好列表中的排列顺序为毫米波基站的匹配效用函数值从大到小的排列顺序,毫米波基站的每个匹配效用函数对应一个终端。
S1004、根据每个终端的匹配偏好列表和每个毫米波基站的匹配偏好列表,创建基站-用户协同矩阵,以使得每个毫米波基站基于基站-用户协同矩阵与自身匹配的终端建立下行通信连接。其中基站-用户协同矩阵用于表示每个毫米波基站与终端之间的匹配关系。S1004的具体实现方式可参考上述步骤六。
采用上述方法,在静态场景下,毫米波基站在选择毫米波基站匹配的终端时,不但考虑到基站内干扰还考虑到基站间干扰,即该方式能够最小化各毫米波基站与匹配的终端通信时的基站内干扰和基站间干扰,提高毫米波基站与终端之间的通信速率,提高通信效率。
在本申请的一个实施例中,在S1004之后,该方法还包括以下步骤:
步骤Ⅰ、计算若每个毫米波基站基于基站-用户协同矩阵与自身匹配的终端之间建立下行通信连接,所建立的每个下行通信连接的连接状态。其中,连接状态包括毫米波基站到终端的有效天线增益、基站内干扰和基站间干扰。步骤Ⅰ的实现方式可参考上述步骤七。
步骤Ⅱ、根据各下行通信连接的连接状态,基于凸优化理论更新毫米波基站集合中所有毫米波基站的各发射波束的发射功率。其中,步骤Ⅱ的具体实现方式可参考上述步骤七。
步骤Ⅲ、基于各毫米波基站的各发射波束的发射功率,计算若各毫米波基站与匹配的 终端之间建立下行通信连接,毫米波基站集合与终端所形成的网络的网络吞吐量,并确定当前计算的网络吞吐量与上一次计算的网络吞吐量之间的差值是否小于预设的收敛阈值。若是,则向毫米波基站集合包括的其他毫米波基站发送基站-用户协同矩阵和各毫米波基站的发射波束的发射功率,以使得每个毫米波基站基于基站-用户协同矩阵以及自身对应的发射功率与自身匹配的终端建立下行通信连接。若否,则返回S1001。其中,步骤Ⅲ的具体实现方式可参考上述步骤八。
采用上述方法,毫米波基站可以基于凸优化理论,以提高毫米波基站与终端之间的形成的网络的网络吞吐量为目标,优化毫米波基站集合中所有毫米波基站的各发射波束的发射功率,从而提高了终端与毫米波基站之间的传输速率和传输效率。
在本申请的一个实施例中,上述S1004可以实现为:针对每个终端,选择该终端的匹配偏好列表中排序最高的毫米波基站,作为该终端的候选匹配基站;针对每个毫米波基站,从将该毫米波基站作为候选匹配基站的终端中,按照终端在该毫米波基站的匹配偏好列表中的排列顺序,选择预设数量的终端作为该毫米波基站匹配的终端,得到基站-用户协同矩阵。其中,S1004的具体实现方式可参考上述步骤六。
本申请实施例中,图10中的毫米波基站可以是毫米波基站集合m的其中一个基站,将该毫米波基站称为调度基站。调度基站可通过图10的方式确定毫米波基站集合m中的每个毫米波基站匹配的终端,然后向其他毫米波基站发送第三通知消息,以通知每个毫米波基站与终端之间的匹配关系。使得其他毫米波基站接收到第三通知消息后,与自身匹配的终端建立下行通信连接,并按照自身对应的发射功率与连接的终端通信。同时调度基站可与自身匹配的终端建立下行通信连接,并按照自身对应的发射功率与连接的终端通信。
本申请实施例中,在静态场景下,调度基站可结合基站间干扰、基站内干扰以及基站回程链路的压力,选择每个毫米波基站匹配的终端,使得毫米波基站与终端之间的通信可以尽量减少基站回程链路的压力,同时减少各下行通信链路之间的干扰,提高网络吞吐量。
对应于上述方法实施例,本申请实施例提供了一种车辆,包括:
至少一个处理器;以及与至少一个处理器通信连接的存储器;其中,
存储器存储有可被至少一个处理器执行的指令,指令被至少一个处理器执行,以使至少一个处理器能够执行上述毫米波波束资源分配与优化方法中由车辆执行的步骤。
对应于上述方法实施例,本申请实施例提供了一种基站,包括:
至少一个处理器;以及与至少一个处理器通信连接的存储器;其中,
存储器存储有可被至少一个处理器执行的指令,指令被至少一个处理器执行,以使至少一个处理器能够执行上述毫米波波束资源分配与优化方法中由毫米波基站执行的步骤。
本申请实施例提供了一种存储有计算机指令的非瞬时计算机可读存储介质,其中,计算机指令用于使计算机执行上述毫米波波束资源分配与优化方法中车辆执行的步骤。
本申请实施例提供了一种存储有计算机指令的非瞬时计算机可读存储介质,其中,计算机指令用于使计算机执行上述毫米波波束资源分配与优化方法中由毫米波基站执行的步骤。
本申请实施例提供了一种计算机程序产品,包括计算机程序,计算机程序在被处理器 执行时实现上述毫米波波束资源分配与优化方法中由车辆执行的步骤。
本申请实施例提供了一种计算机程序产品,包括计算机程序,计算机程序在被处理器执行时实现上述毫米波波束资源分配与优化方法中由毫米波基站执行的步骤。
以上所述仅为本申请的较佳实施例,并不用以限制本申请,凡在本申请的精神和原则之内,所做的任何修改、等同替换、改进等,均应包含在本申请保护的范围之内。

Claims (21)

  1. 一种静态和动态相结合的毫米波波束资源分配与优化方法,其特征在于,具体步骤如下:
    步骤一、针对某用户i,判断该用户的移动速度是否小于速度阈值η v,如果是,则转入步骤二,归类到静态场景处理;否则,用户位于车辆内,进入步骤九,归类到动态场景处理;
    步骤二、构建两层毫米波异构蜂窝网络的下行传输场景;
    场景中包括:M个毫米波小区基站,用集合
    Figure PCTCN2021136801-appb-100001
    表示,且每个小区基站具备缓存能力;各个毫米波小区基站通过回程链路连接到宏基站0,通过接入链路连接到I个用户,表示为用户集合
    Figure PCTCN2021136801-appb-100002
    步骤三、针对当前次迭代,计算用户i与基站m的有效天线增益、基站内干扰和基站间干扰,并上报给基站m;
    基站m到用户i的有效天线增益,表示为:
    Figure PCTCN2021136801-appb-100003
    Figure PCTCN2021136801-appb-100004
    为基站发射天线的主瓣增益,
    Figure PCTCN2021136801-appb-100005
    为用户接收天线的主瓣增益;
    基站内干扰增益表示为:
    Figure PCTCN2021136801-appb-100006
    Figure PCTCN2021136801-appb-100007
    表示发射波束
    Figure PCTCN2021136801-appb-100008
    Figure PCTCN2021136801-appb-100009
    主瓣间的重叠角度,
    Figure PCTCN2021136801-appb-100010
    θ t表示基站发射天线的主瓣波束宽度;
    Figure PCTCN2021136801-appb-100011
    表示发射波束
    Figure PCTCN2021136801-appb-100012
    相对基准方向m→i的偏移角度,
    Figure PCTCN2021136801-appb-100013
    表示发射波束
    Figure PCTCN2021136801-appb-100014
    相对基准方向m→i的偏移角度;
    Figure PCTCN2021136801-appb-100015
    为基站发射天线的旁瓣增益;
    基站间干扰增益
    Figure PCTCN2021136801-appb-100016
    表示为:
    Figure PCTCN2021136801-appb-100017
    Figure PCTCN2021136801-appb-100018
    为干扰发射波束
    Figure PCTCN2021136801-appb-100019
    相对于基准方向m′→i的波束偏移角度;
    Figure PCTCN2021136801-appb-100020
    为参考接收波束
    Figure PCTCN2021136801-appb-100021
    相对于基准方向m′→i的波束偏移角度;
    步骤四、利用有效天线增益、基站内干扰和基站间干扰,计算毫米波小区基站m和用户i接入链路的传输速率,进而分别构建各自的匹配效用函数U m(i)和U i(m);
    首先,利用基站内干扰和基站间干扰计算基站m到用户i的信干噪比γ m,i
    Figure PCTCN2021136801-appb-100022
    P m,i代表基站m发射给用户i的发射功率;h m,i为基站m到用户i的小尺度阴影衰落的增益;L m,i指基站m到用户i的大尺度路径损耗增益;
    Figure PCTCN2021136801-appb-100023
    指发射波束
    Figure PCTCN2021136801-appb-100024
    对发射波束
    Figure PCTCN2021136801-appb-100025
    造成的基站内干扰增益;
    Figure PCTCN2021136801-appb-100026
    指干扰发射波束
    Figure PCTCN2021136801-appb-100027
    对参考接收波束
    Figure PCTCN2021136801-appb-100028
    造成的基站间干扰增益;h m',i指干扰发射基站m'到参考接收用户i的小尺度阴影衰落增益;L m',i指干扰发射基站m'到参考接收用户i的大尺度路径损耗增益;
    Figure PCTCN2021136801-appb-100029
    是L m',i的倒数;P N指加性高斯白噪声功率;
    然后,利用信干噪比γ m,i计算基站m到用户i的接入链路的传输速率,表示为:
    R m,i=B alog 2(1+γ m,i);
    B a表示接入链路的带宽;
    最后,利用接入链路的传输速率计算用户i的效用函数U i(m)和基站m的效用函数U m(i);用户i的效用函数U i(m)表示为:
    Figure PCTCN2021136801-appb-100030
    其中,τ为回程占用率的权重因子,且τ≥0;R m为宏基站0到毫米波小区基站m的传输速率;基站m的效用函数U m(i)如下:
    Figure PCTCN2021136801-appb-100031
    ω 1为干扰项的权重因子,ω 2为回程占用率的权重因子,且ω 1≥0,ω 2≥0;I m,i为用户i对其他用户造成的总干扰;c m,v(i)表示用户i当前的申请内容;
    步骤五、构建基站m和每个用户的匹配效用函数,形成基站m的匹配偏好列表List m;构建用户i和每个基站的匹配效用函数,形成用户i的匹配偏好列表List i
    步骤六、根据每个基站和每个用户的匹配偏好列表对多基站-多用户进行匹配,得到基站-用户协同矩阵X,建立下行通信连接;
    步骤七、基于基站-用户连接状态,基于凸优化理论优化基站的毫米波波束发射功率,生成当前次迭代的发射功率最优解;
    首先,构建基站的毫米波波束发射功率的资源分配模型:
    P2:
    Figure PCTCN2021136801-appb-100032
    s.t.C1:
    Figure PCTCN2021136801-appb-100033
    C2:
    Figure PCTCN2021136801-appb-100034
    C3:
    Figure PCTCN2021136801-appb-100035
    x m,i为基站-用户协同变量,x m,i∈X;X指基站-用户协同优化矩阵,表征基站和用户是否建立通信连接;矩阵X的维度为M*I;
    P指发射功率优化矩阵,表征基站的发射波束到用户的发射功率大小;矩阵P的维度为M*I;发射功率变量P m,i∈P;
    p f指内容f的申请概率,且服从Zipf分布,即
    Figure PCTCN2021136801-appb-100036
    其中δ代表文件f的流行度指数;c m,f为内容f的缓存变量,且c m,f∈{1,0},代表基站m对是否缓存内容f;
    其中,约束条件C1保证每个波束的发射功率非负值,且每个基站的总发射功率不超过功率最大值P max;约束条件C2保证每个基站传输未缓存内容占用的回程容量不超过该基站的回程链路能力;约束条件C3保证用户i的接入速率不小于最低速率需求
    Figure PCTCN2021136801-appb-100037
    然后,利用连续凸逼近求解法,计算得到频谱效率r m,i的凸上界和凸下界;
    最后,利用边界在t∈{1,...,T}的时间内迭代生成发射功率的最优解序列{P (t)};
    步骤八、基于当前次迭代的基站-用户协同矩阵X和发射功率计算网络吞吐量,并判断当前次迭代的网络吞吐量与前一次迭代的网络吞吐量的差值是否小于收敛阈值,如果是,即网络吞吐量收敛,当前次迭代的基站-用户协同矩阵和发射功率作为最优解,毫米波基站进行下行传输直到所有数据传输完毕;否则,返回步骤三重新运行下一次迭代;
    步骤九、构建动态车辆网络的内容分发传输场景;
    场景中包括:N个车辆随机分布在长度为L的双车道上,各个车辆以不同的速度朝同一方向行驶,车辆集合
    Figure PCTCN2021136801-appb-100038
    且车辆具备缓存能力,并遵从FMM移动模型;
    内容分发过程可以分为两个阶段:V2I阶段和V2V阶段;在V2I阶段,车辆处于毫米波基站的覆盖范围内,接收到流行内容的部分片段;每个基站最多支持Q 0个波束同时传输;
    在V2V阶段,车辆在毫米波小区覆盖范围外,车辆间通过合作来进行内容共享,从而获取尽可能多的剩余所需内容片段;每个车辆最多支持Q v个波束同时传输,或支持单个毫米波接收波束;
    流行内容集合为
    Figure PCTCN2021136801-appb-100039
    将各个内容切分为大小为s比特的单位内容片段,则内容c的大小表示为sD c比特,其中D c表示单位内容片段的个数;
    步骤十、判断车辆是否在基站覆盖范围内,如果是,转到步骤十一;否则,转到步骤十三;
    步骤十一:计算车辆i与毫米波小区基站m的信道状态,并上报给毫米波小区基站m;
    信道状态包括有效天线增益和基站内干扰;
    步骤十二:在V2I阶段每个调度时隙T t下,每个基站挑选各自的最优车辆来传输下行数据,进入步骤十七;
    步骤十三:每个调度时隙T t下,确定与各基站连接的车辆作为内容发送车辆,将未连接的车辆作为接收车辆;
    步骤十四:利用每个接收车辆和每个内容发送车辆形成的链路,计算各链路的实际接收内容数量;具体步骤为:
    步骤1401,针对发送车辆i和接收车辆j形成的链路l i,j,计算有效天线增益G i,j
    Figure PCTCN2021136801-appb-100040
    步骤1402,计算当前链路l i,j受到来自于同时传输内容的其他链路l i',j'的干扰I i,j,表示为:
    Figure PCTCN2021136801-appb-100041
    P t为车辆发射波束的功率;h i',j为干扰发送车辆i'到参考接收车辆j链路的小尺度阴影衰落;L i',j为干扰发送车辆i'到参考接收车辆j链路的大尺度路径损耗;
    Figure PCTCN2021136801-appb-100042
    为V2V链路间干扰增益,表示为:
    Figure PCTCN2021136801-appb-100043
    Figure PCTCN2021136801-appb-100044
    为基站接收天线的旁瓣增益;
    Figure PCTCN2021136801-appb-100045
    为干扰发射波束
    Figure PCTCN2021136801-appb-100046
    相对于基准方向i′→j的波束偏移角度;
    Figure PCTCN2021136801-appb-100047
    为参考接收波束
    Figure PCTCN2021136801-appb-100048
    相对于基准方向i′→j的波束偏移角度;
    步骤1403,利用V2V链路间干扰计算发送车辆i和接收车辆j的信干噪比,进一步计算接入链路l i,j的传输速率R i,j
    步骤1404,利用接入链路l i,j的传输速率R i,j,计算链路l i,j的实际可传输内容片段数量
    Figure PCTCN2021136801-appb-100049
    Figure PCTCN2021136801-appb-100050
    t s指链路相关通信时间,即车辆收发波束从对准到失去对准的时间;s为每个单位内容片段的比特大小;
    步骤1405,计算链路l i,j的命中内容片段数量
    Figure PCTCN2021136801-appb-100051
    Figure PCTCN2021136801-appb-100052
    Figure PCTCN2021136801-appb-100053
    为发送车辆i的缓存内容片段集合,
    Figure PCTCN2021136801-appb-100054
    为接收车辆j的申请内容集合;
    步骤1406,通过链路l i,j的实际可传输内容片段数量
    Figure PCTCN2021136801-appb-100055
    和命中内容片段数量
    Figure PCTCN2021136801-appb-100056
    计算链路l i,j的实际接收内容数量;
    Figure PCTCN2021136801-appb-100057
    步骤十五:利用每个链路的实际接收内容数量计算车辆间联盟合作和车辆个体的效用 函数;联盟合作的效用函数为联盟收益函数和开销函数的差值;具体如下:
    联盟收益函数为联盟S中所有链路l i,j实际接收到的内容数量,计算公式为:
    Figure PCTCN2021136801-appb-100058
    开销函数与联盟中V2V链路数成正比,计算公式为:
    Figure PCTCN2021136801-appb-100059
    |S|-1为联盟S中的链路总数,其中β为常数;个体效用函数
    Figure PCTCN2021136801-appb-100060
    的计算公式如下:
    Figure PCTCN2021136801-appb-100061
    其中,δ是略大于1的常数;
    Figure PCTCN2021136801-appb-100062
    表示发送内容的车辆集合;
    Figure PCTCN2021136801-appb-100063
    表示接收内容的车辆集合;
    步骤十六:将每个发送车辆和每个内容接收车辆,根据联盟合作和车辆个体的效用函数实现车辆间联盟,建立V2V链路实现内容分发;
    步骤十七:每个传输时隙T t执行V2I或V2V的毫米波波束资源联合优化,直到数据传输完毕或者达到调度周期T s=NT t
    步骤十八、利用静态场景下基站-用户协同和发射功率的最优结果,或动态场景下车辆联盟的最优结果,实现静态和动态相结合的毫米波波束资源分配。
  2. 如权利要求1所述的一种静态和动态相结合的毫米波波束资源分配与优化方法,其特征在于,所述步骤五中列表List m中存储的是按效用函数从大到小排列的各用户序号;列表List i中存储的是按效用函数从大到小排列的各基站序号;同理,对每个基站都能得到各基站的匹配偏好列表,对每个用户也能得到各用户的匹配偏好列表。
  3. 如权利要求1所述的一种静态和动态相结合的毫米波波束资源分配与优化方法,其特征在于,所述步骤六具体匹配过程如下:
    首先,针对未完全连接的用户i,向匹配偏好列表List i中效用函数值最高的小区基站m发送接入请求
    Figure PCTCN2021136801-appb-100064
    并将该基站m的序号从列表List i中清除;同理,每个未完全连接的用户都从各自的匹配偏好列表List i中选择效用函数值最高的小区基站发送请求操作;
    然后,基站m将它的所有请求用户加入集合
    Figure PCTCN2021136801-appb-100065
    判断集合
    Figure PCTCN2021136801-appb-100066
    中的用户数量是否小于等于基站m的配额Q m,如果是,基站m接受集合
    Figure PCTCN2021136801-appb-100067
    中所有的用户请求;否则,基站m只接受Q m个函数值最高的用户并且拒绝其他申请用户;
    被拒绝的用户仍属于未完全连接,重复上述过程,继续进行下一轮迭代,直至最终不存在用户被拒绝时,匹配过程结束,得到最终优化的基站-用户协同矩阵X。
  4. 如权利要求1所述的一种静态和动态相结合的毫米波波束资源分配与优化方法,其特征在于,所述步骤七中连续凸逼近的具体过程如下:
    a)、利用用户i与基站m的有效天线增益、基站内干扰和基站间干扰,计算中间函数:
    Figure PCTCN2021136801-appb-100068
    Figure PCTCN2021136801-appb-100069
    是指发射波束
    Figure PCTCN2021136801-appb-100070
    受到的基站内干扰和基站间干扰的总和;
    Figure PCTCN2021136801-appb-100071
    是基站m到用户i的接收功率和
    Figure PCTCN2021136801-appb-100072
    的总和;
    b)、然后利用两中间函数相减,计算频谱效率r m,i
    Figure PCTCN2021136801-appb-100073
    c)、分别利用泰勒展开公式和对数函数性质对两个函数进行凸逼近处理;具体为:
    对函数
    Figure PCTCN2021136801-appb-100074
    利用泰勒展开公式近似为一阶泰勒函数的形式:
    Figure PCTCN2021136801-appb-100075
    对函数
    Figure PCTCN2021136801-appb-100076
    利用对数函数性质
    Figure PCTCN2021136801-appb-100077
    近似为:
    Figure PCTCN2021136801-appb-100078
    d)、基于上述两个函数的凸逼近处理结果,得到频谱效率r m,i的凸上界
    Figure PCTCN2021136801-appb-100079
    表示为:
    Figure PCTCN2021136801-appb-100080
    相反地,将上述两个函数分别利用对数函数性质和泰勒展开的操作逼近,得到r m,i的凸下界,表示为:
    Figure PCTCN2021136801-appb-100081
  5. 如权利要求1所述的一种静态和动态相结合的毫米波波束资源分配与优化方法,其特征在于,所述的步骤十二具体为:
    首先,在V2I阶段,计算车辆i接收到基站m的功率为:
    Figure PCTCN2021136801-appb-100082
    其中,
    Figure PCTCN2021136801-appb-100083
    为基站m发射给车辆i的发射功率;
    然后,同理计算基站m接收到覆盖范围内所有车辆的接收功率,并按功率从大到小对各候选车辆进行排序,选择与前Q 0个车辆进行连接;
    最后,在当前调度时隙T t内,遍历所有已连接车辆,找到传输速率最高的已连接车辆i,判断当前未连接车辆i′替换掉车辆i后的网络吞吐量是否增加,如果是,将车辆i′替换掉车辆i,否则,将当前未连接车辆i′丢弃,继续进行下一个未连接车辆的替换工作;
    依次将所有未连接车辆均替换完毕后,完成基站m的所有最优连接车辆选择。
  6. 如权利要求1所述的一种静态和动态相结合的毫米波波束资源分配与优化方法,其特征在于,所述的步骤十六具体为:
    首先,随机划分收发车辆,构建初始联盟组合
    Figure PCTCN2021136801-appb-100084
    并且初始化当前联盟组合
    Figure PCTCN2021136801-appb-100085
    和初始化迭代指示变量iter=0;
    然后,针对发送车辆i,找到当前所属联盟
    Figure PCTCN2021136801-appb-100086
    以及随机选择当前联盟组合中的其他新联盟
    Figure PCTCN2021136801-appb-100087
    判断车辆i相对于旧联盟S k,是否更偏好选择新联盟S m,即是否满足S miS k,如果是,即车辆i离开旧联盟S k并加入新联盟S m,同时更新联盟组合;否则,车辆i继续选择当前联盟组合中的其他联盟进行判断,直到车辆i找到偏好的新联盟或者当前联盟中的所有其他联盟都不被偏好,车辆i的联盟过程终止,继续下一个车辆进行联盟选择;
    S miS k的计算公式为:
    Figure PCTCN2021136801-appb-100088
    Figure PCTCN2021136801-appb-100089
    指车辆i在加入新联盟S m后可获得的个体效用值;
    Figure PCTCN2021136801-appb-100090
    指车辆i还在旧联盟S k中可以获得的个体效用;U(S m∪{i})指车辆i在加入新联盟S m后,联盟S m可获得的联盟效用值;U(S k\{i})指车辆i还在旧联盟S k中时,联盟S k可获得的联盟效用值;
    联盟组合的更新公式为:
    Figure PCTCN2021136801-appb-100091
    最后,判断更新的联盟组合
    Figure PCTCN2021136801-appb-100092
    是否收敛于纳什均衡
    Figure PCTCN2021136801-appb-100093
    如果是,得到最优的车辆联盟结构;否则,选择下一个发送车辆重复上述过程,直至满足联盟组合的要求。
  7. 一种毫米波波束资源分配与优化方法,其特征在于,所述方法应用于车辆,所述方法包括:
    当所述车辆处于毫米波基站覆盖范围内时,若所述车辆在当前时隙与毫米波基站连接,则所述车辆作为发送车辆通过V2I链路接收毫米波基站通过波束发送的内容片段;所述内容片段为对流行内容集合中的内容进行切分后得到的;
    当所述车辆行驶出所述毫米波基站的覆盖范围,或所述车辆未驶出所述毫米波基站的覆盖范围,但在当前时隙未与毫米波基站连接的情况下,所述车辆作为接收车辆加入目标车辆联盟,通过V2V链路接收所述目标车辆联盟中的发送车辆发送的内容片段。
  8. 根据权利要求7所述的方法,其特征在于,所述车辆作为接收车辆加入目标车辆联盟,包括:
    计算若所述接收车辆与所述初始车辆联盟中的发送车辆之间建立V2V链路,所述接收车辆能够从该V2V链路接收到的接收内容片段数量,所述初始车辆联盟为随机选择的一个车辆联盟;
    基于该V2V链路对应的接收内容片段数量,计算若所述接收车辆加入所述初始车辆联盟后,所述初始车辆联盟的合作效用函数值和所述接收车辆的个体效用函数值;
    选择一个车辆联盟,并计算若所述接收车辆加入选择的车辆联盟后,选择的车辆联盟的合作效用函数值和所述接收车辆的个体效用函数值;
    判断选择的车辆联盟的合作效用函数值是否大于所述初始车辆联盟的合作效用函数值,以及所述接收车辆加入选择的车辆联盟的个体效用函数值是否大于所述接收车辆加入所述初始车辆联盟的个体效用函数值;
    若均为是,则确定选择的车辆联盟为目标车辆联盟;否则,返回所述选择一个车辆联盟的步骤,直至确定出目标车辆联盟;
    与所述目标车辆联盟中的发送车辆之间建立V2V链路。
  9. 根据权利要求8所述的方法,其特征在于,
    车辆联盟的合作效用函数为:U(S)=V(S)-C(S);
    其中,
    Figure PCTCN2021136801-appb-100094
    Figure PCTCN2021136801-appb-100095
    接收车辆的个体效用函数为:
    Figure PCTCN2021136801-appb-100096
    S为车辆联盟的编号,V(S)表示车辆联盟S中各V2V链路对应的接收内容片段数量总和,C(S)表示车辆联盟S中各V2V链路的开销,
    Figure PCTCN2021136801-appb-100097
    表示车辆联盟S中的发送车辆i与接收车辆j之间建立的V2V链路对应的接收内容片段数量,|S|表示车辆联盟S包括的车辆数量,β和δ为预设的常数,
    Figure PCTCN2021136801-appb-100098
    表示接收车辆j在车辆联盟S中的个体效用函数值。
  10. 根据权利要求8或9所述的方法,其特征在于,计算若所述接收车辆与所述初始车辆联盟中的发送车辆之间建立V2V链路,所述接收车辆能够从该V2V链路接收到的接收内容片段数量,包括:
    将该V2V链路的发送车辆的发射天线增益和所述接收车辆的接收天线增益的乘积,作为该V2V链路的有效天线增益;
    计算该V2V链路受到的来自于其他V2V链路的干扰;
    根据该V2V链路的有效天线增益和该V2V链路受到的来自于其他链路的干扰,计算该V2V链路的传输速率;
    根据该V2V链路的传输速率,计算该V2V链路的接收内容片段数量;
    其中,该V2V链路的接收内容片段数量为:
    Figure PCTCN2021136801-appb-100099
    其中,
    Figure PCTCN2021136801-appb-100100
    表示发送车辆i和接收车辆j之间建立的V2V链路l i,j的接收内容片段数量,
    Figure PCTCN2021136801-appb-100101
    表示l i,j的可传输内容片段数量,
    Figure PCTCN2021136801-appb-100102
    t s表示发送车辆i的发送波束和接收车辆j的接收波束从对准到失去对准的时间,R i,j表示l i,j的传输速率,s表示每个内容片段的大小,
    Figure PCTCN2021136801-appb-100103
    表示l i,j的命中内容片段数量,
    Figure PCTCN2021136801-appb-100104
    C i表示发送车辆i的缓存内容片段集合,C j表示接收车辆j的申请内容片段集合,card表示集合的大小。
  11. 一种毫米波波束资源分配与优化方法,其特征在于,应用于毫米波基站,所述方法包括:
    确定在当前时隙处于所述毫米波基站覆盖范围内的多个车辆;
    计算若所述毫米波基站与所述多个车辆中的每个车辆之间建立V2I链路,各V2I链路的信道状态,所述信道状态包括毫米波基站到车辆的有效天线增益,所述有效天线增益基于所述毫米波基站的基站内干扰确定,所述基站内干扰包括所述毫米波基站的不同发射波束之间的干扰;
    根据各V2I链路的信道状态,从所述多个车辆中选择预设数量个车辆作为发送车辆;
    与每个发送车辆建立V2I链路,并通过V2I链路向每个发送车辆发送内容片段,以使得发送车辆通过V2V链路向与自身属于同一车辆联盟的接收车辆发送内容片段,所述接收车辆为当前时隙未与所述毫米波基站连接的车辆。
  12. 根据权利要求11所述的方法,其特征在于,所述根据各V2I链路的信道状态,从所述多个车辆中选择预设数量个车辆作为发送车辆,包括:
    根据每个V2I链路的信道状态,确定所述毫米波基站覆盖范围内的每个车辆接收所述毫米波基站发送的毫米波信号的接收功率;
    按照接收功率从大到小的顺序,选择预设数量个车辆作为发送车辆;
    针对每个未选择的车辆,判断若利用该车辆替换目标发送车辆,所述毫米波基站与各发送车辆之间建立的V2I链路所形成的网络的网络吞吐量是否增加;所述目标发送车辆为接收功率最大的发送车辆;
    若是,则利用该车辆替换所述目标发送车辆。
  13. 根据权利要求11所述的方法,其特征在于,所述毫米波基站覆盖范围内的每个车辆的接收功率为:
    Figure PCTCN2021136801-appb-100105
    其中,
    Figure PCTCN2021136801-appb-100106
    表示车辆i接收毫米波基站m发送的毫米波信号的接收功率,
    Figure PCTCN2021136801-appb-100107
    表示毫米波基站m向车辆i发送信号的发射功率,G m,i表示毫米波基站m到车辆i的有效天线增益,h m,i表示毫米波基站m到车辆i的小尺度阴影衰落增益,
    Figure PCTCN2021136801-appb-100108
    为毫米波基站m到车辆i的路径损耗增益L m,i的倒数。
  14. 一种毫米波波束资源分配与优化方法,其特征在于,应用于毫米波基站,所述毫米波基站位于两层毫米波异构蜂窝网络,所述异构蜂窝网络包括宏基站和毫米波基站集合,所述毫米波基站集合中的每个毫米波基站分别通过回程链路连接于所述宏基站,所述方法包括:
    获取当前时隙内,所述毫米波基站集合中每个毫米波基站到信号覆盖范围内的各终端 的有效天线增益、各毫米波基站的每个发射波束的发射功率、基站内干扰和基站间干扰,所述基站内干扰包括同一个毫米波基站的不同发射波束之间的干扰,所述基站间干扰包括不同毫米波基站的发射波束之间的干扰;
    根据每个毫米波基站到自身信号覆盖范围内的各终端的有效天线增益、各毫米波基站的每个发射波束的发射功率、基站内干扰和基站间干扰,确定若各毫米波基站与信号覆盖范围内的每个终端之间建立下行通信连接,该下行通信连接对应的毫米波基站的匹配效用函数值以及终端的匹配效用函数值;
    根据各下行通信连接对应的毫米波基站的匹配效用函数值以及终端的匹配效用函数值,确定每个终端针对毫米波基站的匹配偏好列表和每个毫米波基站针对终端的匹配偏好列表;
    根据每个终端的匹配偏好列表和每个毫米波基站的匹配偏好列表,创建基站-用户协同矩阵,所述基站-用户协同矩阵用于表示每个毫米波基站与终端之间的匹配关系,以使得每个毫米波基站基于所述基站-用户协同矩阵与自身匹配的终端建立下行通信连接。
  15. 根据权利要求14所述的方法,其特征在于,在所述根据每个终端的匹配偏好列表和每个毫米波基站的匹配偏好列表,创建基站-用户协同矩阵之后,所述方法还包括:
    计算若每个毫米波基站基于所述基站-用户协同矩阵与自身匹配的终端之间建立下行通信连接,所建立的每个下行通信连接的连接状态,其中,所述连接状态包括毫米波基站到终端的有效天线增益、基站内干扰和基站间干扰;
    根据各下行通信连接的连接状态,基于凸优化理论更新所述毫米波基站集合中所有毫米波基站的各发射波束的发射功率;
    基于各毫米波基站的发射波束的发射功率,计算若各毫米波基站与匹配的终端之间建立下行通信连接,所述毫米波基站集合与终端所形成的网络的网络吞吐量,并确定当前计算的网络吞吐量与上一次计算的网络吞吐量之间的差值是否小于预设的收敛阈值;
    若是,则向所述毫米波基站集合包括的其他毫米波基站发送所述基站-用户协同矩阵和各毫米波基站的发射波束的发射功率,以使得每个毫米波基站基于所述基站-用户协同矩阵以及自身对应的发射功率与自身匹配的终端建立下行通信连接;
    若否,则返回所述获取当前时隙内,所述毫米波基站集合中每个毫米波基站到信号覆盖范围内的各终端的有效天线增益、各毫米波基站的每个发射波束的发射功率、基站内干扰和基站间干扰的步骤。
  16. 根据权利要求14所述的方法,其特征在于,所述根据每个终端的匹配偏好列表和每个毫米波基站的匹配偏好列表,创建基站-用户协同矩阵,包括:
    针对每个终端,选择该终端的匹配偏好列表中排序最高的毫米波基站,作为该终端的候选匹配基站;
    针对每个毫米波基站,从将该毫米波基站作为候选匹配基站的终端中,按照终端在该毫米波基站的匹配偏好列表中的排列顺序,选择预设数量的终端作为该毫米波基站匹配的终端,得到所述基站-用户协同矩阵。
  17. 根据权利要求14-16任一项所述的方法,其特征在于,终端的匹配效用函数为:
    Figure PCTCN2021136801-appb-100109
    基站的匹配效用函数为:
    Figure PCTCN2021136801-appb-100110
    其中,U i(m)表示终端i与基站m通信时终端的匹配效用函数值,R m,i表示基站m和终端i之间建立的下行通信连接的传输速率,R m表示基站m的回程链路的传输速率,τ为预设的权重因子,U m(i)表示终端i与基站m通信时基站的匹配效用函数值,ω 1和ω 2为预设的权重因子,I m,i表示终端i与基站m通信时对其他终端的通信造成的干扰,c m,v(i)表示终端i当前申请获取的内容片段。
  18. 一种车辆,包括:至少一个处理器;以及与所述至少一个处理器通信连接的存储器;其中,所述存储器存储有可被所述至少一个处理器执行的指令,所述指令被所述至少一个处理器执行,以使所述至少一个处理器能够执行权利要求7-10中任一项所述的方法。
  19. 一种基站,包括:至少一个处理器;以及与所述至少一个处理器通信连接的存储器;其中,所述存储器存储有可被所述至少一个处理器执行的指令,所述指令被所述至少一个处理器执行,以使所述至少一个处理器能够执行权利要求11-13或14-17中任一项所述的方法。
  20. 一种存储有计算机指令的非瞬时计算机可读存储介质,其中,所述计算机指令用于使所述计算机执行根据权利要求7-10中任一项所述的方法。
  21. 一种存储有计算机指令的非瞬时计算机可读存储介质,其中,所述计算机指令用于使所述计算机执行根据权利要求11-13或14-17中任一项所述的方法。
PCT/CN2021/136801 2020-12-10 2021-12-09 一种静态和动态相结合的毫米波波束资源分配与优化方法 WO2022121985A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202011454084.8 2020-12-10
CN202011454084.8A CN112616189B (zh) 2020-12-10 2020-12-10 一种静态和动态相结合的毫米波波束资源分配与优化方法

Publications (1)

Publication Number Publication Date
WO2022121985A1 true WO2022121985A1 (zh) 2022-06-16

Family

ID=75234093

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/136801 WO2022121985A1 (zh) 2020-12-10 2021-12-09 一种静态和动态相结合的毫米波波束资源分配与优化方法

Country Status (2)

Country Link
CN (1) CN112616189B (zh)
WO (1) WO2022121985A1 (zh)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115515100A (zh) * 2022-09-19 2022-12-23 西北工业大学 一种基于ris-uav的无源波束位置优化方法、系统、介质及设备
CN115580879A (zh) * 2022-09-07 2023-01-06 重庆邮电大学 一种基于联邦强化学习的毫米波网络波束管理方法
CN116437451A (zh) * 2023-06-07 2023-07-14 天地信息网络研究院(安徽)有限公司 一种基于时隙序列的定向自组网动态功率分配方法
CN116582842A (zh) * 2023-07-13 2023-08-11 中国科学技术大学 一种irs辅助的车载毫米波广播方法、系统及储存介质
CN117255334A (zh) * 2023-11-17 2023-12-19 国网浙江省电力有限公司信息通信分公司 一种面向应急卫星通信的多级协同调度方法及系统

Families Citing this family (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112616189B (zh) * 2020-12-10 2022-09-09 北京邮电大学 一种静态和动态相结合的毫米波波束资源分配与优化方法
CN113242579B (zh) * 2021-05-06 2022-12-09 西北工业大学 一种双连接小区切换参数的测量方法
CN113225707B (zh) * 2021-05-12 2022-08-19 江苏亨鑫众联通信技术有限公司 一种无线网络对车道潮汐现象的接入系统及方法
CN113473245B (zh) * 2021-06-11 2022-04-08 吉林大学 基于可再生能源und视频流等待时间优化方法
CN113825238B (zh) * 2021-09-03 2024-02-20 天地信息网络研究院(安徽)有限公司 一种卫星通信中柔性信令波束多场景应用方法
CN113852972B (zh) * 2021-09-13 2023-10-10 金华航大北斗应用技术有限公司 一种基于波束共享的高速移动终端波束调度方法
CN114071432B (zh) * 2021-11-18 2024-04-26 中国矿业大学 一种用于地下空间灾后应急场景的无人机中继选择方法
CN114172849A (zh) * 2021-11-26 2022-03-11 华中科技大学 一种基于博弈论的确定性流量整形方法
CN114553284B (zh) * 2022-04-27 2022-07-05 四川太赫兹通信有限公司 一种波束对准方法、装置、基站及计算机可读存储介质
CN115276932B (zh) * 2022-06-02 2023-06-23 敦煌研究院 一种毫米波接入与回传一体化网络中的子信道分配方法
CN114759955A (zh) * 2022-06-15 2022-07-15 华南理工大学 一种面向毫米波通信的基于可调波束的资源优化分配方法
CN115484682A (zh) * 2022-08-30 2022-12-16 北京邮电大学 一种实现通信感知一体化的无线基带处理方法及装置
CN116582163B (zh) * 2023-07-07 2023-12-05 北京邮电大学 一种无线感知协同的多用户波束对准方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110769514A (zh) * 2019-11-08 2020-02-07 山东师范大学 一种异构蜂窝网络d2d通信资源分配方法及系统
CN110856260A (zh) * 2019-11-14 2020-02-28 中国电子科技集团公司第五十四研究所 一种毫米波通信系统多波束多用户资源分配方法
CN111431646A (zh) * 2020-03-31 2020-07-17 北京邮电大学 一种毫米波系统中的动态资源分配方法
CN111669834A (zh) * 2020-07-06 2020-09-15 中南大学 毫米波蜂窝网络中基于d2d辅助的资源分配方法
CN112616189A (zh) * 2020-12-10 2021-04-06 北京邮电大学 一种静态和动态相结合的毫米波波束资源分配与优化方法

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
AU2013381337B2 (en) * 2013-04-25 2016-02-25 Intel Corporation Millimeter-wave communication device and method for intelligent control of transmit power and power density
US10798755B2 (en) * 2018-10-18 2020-10-06 Cisco Technology, Inc. Millimeter wave (mmWave) radio resource allocation scheme for vehicle-to-infrastructure (V2I) communications
CN111083668B (zh) * 2020-01-02 2021-06-11 山东大学 一种在车联网中基于联盟博弈算法的d2d资源分配方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110769514A (zh) * 2019-11-08 2020-02-07 山东师范大学 一种异构蜂窝网络d2d通信资源分配方法及系统
CN110856260A (zh) * 2019-11-14 2020-02-28 中国电子科技集团公司第五十四研究所 一种毫米波通信系统多波束多用户资源分配方法
CN111431646A (zh) * 2020-03-31 2020-07-17 北京邮电大学 一种毫米波系统中的动态资源分配方法
CN111669834A (zh) * 2020-07-06 2020-09-15 中南大学 毫米波蜂窝网络中基于d2d辅助的资源分配方法
CN112616189A (zh) * 2020-12-10 2021-04-06 北京邮电大学 一种静态和动态相结合的毫米波波束资源分配与优化方法

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115580879A (zh) * 2022-09-07 2023-01-06 重庆邮电大学 一种基于联邦强化学习的毫米波网络波束管理方法
CN115515100A (zh) * 2022-09-19 2022-12-23 西北工业大学 一种基于ris-uav的无源波束位置优化方法、系统、介质及设备
CN116437451A (zh) * 2023-06-07 2023-07-14 天地信息网络研究院(安徽)有限公司 一种基于时隙序列的定向自组网动态功率分配方法
CN116437451B (zh) * 2023-06-07 2023-08-15 天地信息网络研究院(安徽)有限公司 一种基于时隙序列的定向自组网动态功率分配方法
CN116582842A (zh) * 2023-07-13 2023-08-11 中国科学技术大学 一种irs辅助的车载毫米波广播方法、系统及储存介质
CN116582842B (zh) * 2023-07-13 2023-10-20 中国科学技术大学 一种irs辅助的车载毫米波广播方法、系统及储存介质
CN117255334A (zh) * 2023-11-17 2023-12-19 国网浙江省电力有限公司信息通信分公司 一种面向应急卫星通信的多级协同调度方法及系统
CN117255334B (zh) * 2023-11-17 2024-01-26 国网浙江省电力有限公司信息通信分公司 一种面向应急卫星通信的多级协同调度方法及系统

Also Published As

Publication number Publication date
CN112616189B (zh) 2022-09-09
CN112616189A (zh) 2021-04-06

Similar Documents

Publication Publication Date Title
WO2022121985A1 (zh) 一种静态和动态相结合的毫米波波束资源分配与优化方法
CN109348537B (zh) 多波束自组织网络信道接入控制方法
Duan et al. SDN enabled 5G-VANET: Adaptive vehicle clustering and beamformed transmission for aggregated traffic
Gui et al. Multi-hop relay-aided underlay D2D communications for improving cellular coverage quality
CN109905918B (zh) 一种基于能效的noma蜂窝车联网动态资源调度方法
CN112737837B (zh) 一种高动态网络拓扑下无人机群带宽资源分配方法
Zhong et al. Traffic matching in 5G ultra-dense networks
Biermann et al. How backhaul networks influence the feasibility of coordinated multipoint in cellular networks [accepted from open call]
KR102235763B1 (ko) 멀티 액세스 엣지 컴퓨팅 기반의 이기종 네트워크 시스템
CN110225565B (zh) 一种基于多波束方向性天线的动中组网方法
CN107249202B (zh) 一种分布式无线回程路由算法
CN108093435A (zh) 基于缓存流行内容的蜂窝下行链路网络能效优化系统及方法
Wang et al. QoS-aware cell association in 5G heterogeneous networks with massive MIMO
Ernst et al. The design and evaluation of fair scheduling in wireless mesh networks
Zhou et al. Multi-beam transmission and dual-band cooperation for control/data plane decoupled WLANs
Ma et al. QoS-Oriented joint optimization of resource allocation and concurrent scheduling in 5G millimeter-wave network
Zhou et al. Multi-scale dynamic allocation of licensed and unlicensed spectrum in software-defined HetNets
CN108366426B (zh) 一种超密集环境下基于模拟退火算法的虚拟映射方法
Qi et al. Advanced user association in non-orthogonal multiple access-based fog radio access networks
CN104507153B (zh) 一种ofdma下的小小区功率控制方法
Parichehreh et al. Seamless LTE connectivity in high‐speed trains
Chukhno et al. Unsupervised learning for D2D-assisted multicast scheduling in mmWave networks
Zhou et al. Joint radio resource allocation for decoupled control and data planes in densely deployed coordinated WLANs
Chaudhari et al. Efficient dynamic relay probing and concurrent backhaul link scheduling for mmWave cellular networks
CN110113820B (zh) 多轮排序和请求拆分的太赫兹无线个域网时隙分配方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21902690

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21902690

Country of ref document: EP

Kind code of ref document: A1