A method for dynamic access of unmanned aerial vehicles (uav) in a distributed manner for unknown number of nodes and topology

Abstract

The application discloses a kind of unmanned plane distributed dynamic access methods for unknown node number and topology, comprising: the channel collision rate data in the preset observation window is based on the node to be accessed, dynamically self-adapting adjustment competition bit number configuration parameter k;Based on k-bit binary random sequence, bit-by-bit transceiving and listening distributed elimination competition is executed, and temporary state data is generated;By characteristic node and accessed auxiliary characteristic node according to the competition interval reception state, verification mask is generated, and is executed with the random sequence of temporary node itself according to bit and arbitration, obtains final access right;Successfully accessed node generates lightweight time frame state snapshot by incremental difference comparison and broadcasts, simultaneously based on activity counter monitoring realizes the automatic recovery of suspicious dead time slot.The application realizes collaborative verification and conflict suppression in hidden terminal scene, with the advantages of low conflict, low latency, low overhead, high reliability, especially suitable for high dynamic, large-scale unmanned plane ad hoc network scene.

Description

Technical Field

[0001] This invention relates to the field of UAV self-organizing network and MAC layer protocol design, and in particular to a distributed dynamic access method for UAVs with unknown number of nodes and topology. Background Technology

[0002] Unmanned aerial vehicle (UAV) ad hoc networks are playing an increasingly crucial role in scenarios such as military reconnaissance, disaster emergency rescue, and wide-area environmental monitoring. However, compared with terrestrial mobile ad hoc networks, UAV ad hoc networks face more severe challenges of dual uncertainties: First, the suddenness of missions and the plug-and-play nature of nodes mean that the number of new nodes to be connected is completely unknown before competition occurs; second, the rapid topology changes brought about by three-dimensional high-dynamic flight make it impossible for any node to obtain accurate network topology information before connecting. Under these conditions of "unknown number of new access nodes" and "unknown network topology," how to achieve low-conflict, low-latency, and deterministic distributed access has become the core technical bottleneck restricting the performance of UAV ad hoc networks.

[0003] Existing access mechanisms used in mobile ad hoc networks exhibit significant limitations when migrated to UAV ad hoc networks. Traditional carrier sense multiple access / collision avoidance mechanisms rely on the RTS / CTS handshake mechanism to reserve channels. However, in the high-speed three-dimensional space of UAVs, link lifetime is extremely short. The RTS / CTS handshake may fail before completion due to nodes flying out of communication range, leading to frequent retransmissions and exponentially increasing backoff delays, resulting in a sharp decline in channel utilization.

[0004] For contention-based access, distributed binary contention mechanisms based on time-division multiple access (TDMA) achieve decentralized arbitration by introducing contention-based micro-time slots, avoiding handshake overhead. However, in such mechanisms, the number of contention-based micro-time slots, i.e., the value k, which determines the contention dimension, is usually configured statically. When the number of nodes suddenly accessing the network far exceeds expectations, a fixed k value will result in an excessively small combination space for binary random sequences, leading to a sharp increase in the probability of different nodes generating the same random number. This causes a dramatic increase in the collision rate within the contention interval, resulting in a significant waste of signaling resources. Conversely, when there are few access nodes, an excessively large k value will cause a large number of micro-time slots to be in a state of non-contention and waste, introducing unnecessary signaling overhead. This one-size-fits-all static configuration cannot match the highly dynamic network load of UAVs.

[0005] Furthermore, the hidden terminal problem is a persistent challenge for wireless ad hoc networks, particularly prominent in 3D UAV scenarios. Two nodes that are each other's hidden terminals cannot directly detect each other's signals and may simultaneously win in binary contention, leading to conflicts. Traditional solutions attempt to eliminate hidden terminals by introducing common neighbors, but under highly dynamic flight conditions, the characteristic nodes capable of simultaneously covering multiple pairs of hidden terminals have extremely short lifespans, and their clearing messages may also become invalid due to collisions. This results in incomplete elimination of hidden terminals, producing the so-called false access phenomenon, which in turn causes continuous co-channel interference within data time slots, severely impacting network throughput.

[0006] In summary, existing access mechanisms, under the dual challenges of unknown node numbers and unknown topology, struggle to simultaneously achieve low conflict rates, low signaling overhead, and high hidden terminal elimination rates. There is an urgent need for a distributed access method that can dynamically adapt to network load and ensure access uniqueness through a robust verification mechanism. Summary of the Invention

[0007] The purpose of this invention is to provide a distributed dynamic access method for UAVs with unknown number of nodes and topology, which solves the problems of high collision rate, high signaling overhead and failure of hidden terminal elimination in traditional access mechanisms under the dual uncertainties of unknown number of new access nodes and unknown network topology.

[0008] Technical solution: A distributed dynamic access method for UAVs with unknown number of nodes and topology, comprising the following steps:

[0009] Acquire channel collision event data of the node to be accessed within a preset observation window, and determine the contention bit configuration parameters used to dynamically adjust the contention dimension;

[0010] Based on the competition bit configuration parameters, a distributed bit-by-bit competition elimination process is performed to generate temporary state data indicating that the node has won in this competition;

[0011] The system acquires the verification mask data generated by the feature nodes and auxiliary nodes within the network, and performs collaborative elimination verification processing on the verification mask data and the competing sequence data generated by itself to obtain the final access right determination result.

[0012] When the judgment result is a win, lightweight state snapshot data representing the changes in resource occupancy status within the current time frame is generated and broadcast, and automatic release processing of suspicious time slots is performed based on the activity monitoring of occupied time slots.

[0013] Beneficial effects: Under the dual uncertainty of the total number of new access nodes and the network topology information, the present invention can achieve efficient distributed access with low conflict, low overhead and reliability under the conditions of dynamic adaptive competition dimension, multi-source collaborative hidden terminal verification and lightweight incremental state synchronization. Attached Figure Description

[0014] Figure 1 This is a flowchart of the overall solution of the present invention.

[0015] Figure 2 This is a flowchart of the present invention for determining the configuration parameters of the competition bit depth for dynamically adjusting the competition dimension.

[0016] Figure 3 This is a flowchart of the present invention for generating temporary state data representing the node that won the competition.

[0017] Figure 4 This is a flowchart of the present invention for obtaining verification mask data generated by feature nodes and auxiliary nodes within the network.

[0018] Figure 5 This is a flowchart illustrating how the present invention generates and broadcasts lightweight state snapshot data representing changes in resource occupancy status within the current frame. Detailed Implementation

[0019] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0020] like Figures 1-5 As shown, this embodiment describes in detail the data processing flow of a distributed dynamic access method for UAVs with unknown node numbers and topology, specifically including:

[0021] S1: Obtain channel collision event data of the node to be accessed within the preset observation window, and determine the contention bit configuration parameters used to dynamically adjust the contention dimension.

[0022] In this embodiment, the node to be accessed dynamically calculates the optimal competition dimension parameters adapted to the current network load by continuously observing and statistically analyzing the historical conflict status of the channel, thus providing an adaptive basic configuration for subsequent distributed competition.

[0023] Specifically, the node to be accessed first continuously performs carrier sensing at its physical layer, collecting channel collision event data within a preset observation window. This observation window is defined as a sliding window of length W signaling time slots. Whenever, within a signaling time slot, multiple signals overlap, causing the receiver to be unable to decode correctly, the node records it as a collision event. By accumulating all the signaling time slots where collisions occur within this sliding window, the node obtains a collision count, denoted as N.collision .

[0024] Furthermore, the node collects collision count data N. collision A ratio operation is performed with the length W of the sliding window. The processing converts the original count data into a normalized probability value that reflects the current intensity of channel collisions, i.e., the collision rate data, calculated using the formula P. collision = N collision / W.

[0025] After calculating the collision rate data, the node compares it with preset high collision thresholds T1 and low collision thresholds T2 to generate an adjustment decision. Thresholds T1 and T2 are preset empirically; for example, T1 can be set to 50%, indicating excessively frequent collisions, and T2 can be set to 10%, indicating a relatively idle channel. The comparison logic is as follows: if the collision rate data P... collision If the value is greater than or equal to the high collision threshold T1, it indicates that the current competition is too intense, and the probability of successful access in this round is extremely low because the random number combination space is insufficient. At this point, the node generates its first decision, namely, the decision to increase the number of competing bits, updating the bit configuration parameter k for the next round of competition to min(k+1, K). max ), where K max This is a preset upper limit for the number of bits involved in the contention, designed to prevent the signaling overhead from expanding indefinitely. Conversely, if the collision rate data P... collision The current contention bit depth configuration parameter k is less than or equal to the low collision threshold T2, and is greater than the preset lower limit K. min This indicates that the current network load is light, and an excessively high k value is causing unnecessary waste of signaling resources. At this point, the node generates a second decision, namely, reducing the number of contention bits, updating k to max(k-1, K). min ).

[0026] Further improvements to this scheme aim to prevent network contention logic chaos caused by nodes independently adjusting the k value. The updated contention bit configuration parameter k will not take effect immediately. The contention bit configuration parameter data will be temporarily stored and then encapsulated into a time frame status information data packet for broadcast in subsequent step S4. Neighboring nodes receive this data packet, extract it, and synchronize their local contention bit configuration parameters. This ensures, in a decentralized manner, that all nodes in the network use a unified contention dimension k for access contention in the next time frame period.

[0027] S2: Based on the competition bit configuration parameters, perform distributed bit-by-bit competition elimination processing to generate temporary state data indicating that the node has won in this competition.

[0028] After obtaining the network-wide synchronized competition bit configuration parameter k, a distributed elimination arbitration without centralized comparison is performed by sending and receiving k-bit binary random sequences, ultimately generating temporary status data to identify potential winners.

[0029] In one alternative implementation, the process of this embodiment begins with the generation of the competition sequence. Each node to be connected independently generates a binary random sequence data R of length k using its local pseudo-random number generator. Each bit of sequence R, i.e., b... k (most significant bit), b k-1 ..., b1 (least significant bit) will serve as the node's unique priority credential in this round of contention. Subsequently, a time-series mapping process is performed, retrieving the k-bit sequence data from the most significant bit b... k The least significant bit b1 is mapped sequentially to k consecutive contention-based micro-slots in the time domain, forming a strict time-bit correspondence.

[0030] Next, the core bit-by-bit transmission and reception / listening phase begins. In the j-th micro-timeslot, each node reads the j-th bit value of its sequence data R and performs conditional transmission and reception processing based on the different values ​​of this bit: on the one hand, if the bit value is 1, the node is considered an active contender. Its transmitter will send a preset contention signal in this micro-timeslot to declare its priority. On the other hand, if the bit value is 0, the node enters the receiving / listening state. Its receiver collects the superimposed energy data of all signals on the channel in the current micro-timeslot. Subsequently, the node performs threshold determination processing on the collected channel energy data: if the energy value is higher than a preset physical energy detection threshold, it indicates that another node is transmitting (its current bit in the random sequence is 1), and at this time, the node will generate a collision indication data representing the detected collision signal.

[0031] Based on the above perception, the distributed eviction logic is triggered immediately. A node detects the presence of a higher-priority node only when it is in a receiving state (its current bit is 0) and generates conflict indication data. According to the arbitration principle of binary counting, the node must immediately withdraw from the competition. The node internally generates an eviction command, forcing its transmitter to remain silent and cease participating in the transmission of signals for all subsequent uncontested micro-slots, thus avoiding unnecessary signaling collisions.

[0032] Finally, when a node has traversed all k contention-slot micro-slots and has never generated an elimination instruction due to its own bit being 0 and a conflict being detected, it indicates that its random sequence, in a binary sense, overwhelms all nodes that it directly or indirectly clashed with in each round. At this point, the node is marked as a temporary node, and a corresponding temporary state data is generated as its qualification credential to enter the next processing step.

[0033] S3: Obtain the verification mask data generated by the feature nodes and auxiliary nodes in the network, and perform collaborative elimination verification processing on the verification mask data and the competing sequence data generated by itself to obtain the final access right determination result.

[0034] In a dynamic topology environment, two nodes that are hidden terminals may simultaneously become temporary nodes because they cannot detect each other, leading to a conflict. This embodiment introduces a third-party verification and collaborative arbitration mechanism to perform a secondary screening of temporary nodes, obtaining a decisive final access right determination result.

[0035] First, there is the stage of acquiring the verification mask data. Each node in the network collects and analyzes its signal reception history within the contention interval. If a node (defined as a feature node) can simultaneously receive and decode signals from multiple resident nodes due to its unique physical location, it is qualified to act as an arbitrator. The feature node retrieves its signal reception status data during k contention micro-slots. Based on this data, it generates a k-bit verification mask data M. The generation rule is: traversing from the 1st to the kth micro-slots, if a valid contention signal (regardless of its source) has been received in that micro-slot, the j-th bit of M is set to 1; otherwise, it is set to 0. The verification mask data M characterizes the bit pattern of the local highest priority signal from the perspective of the feature node. At the same time, to enhance the verification coverage and robustness, nodes that have successfully connected to the network, acting as auxiliary feature nodes, also perform the same processing flow and generate auxiliary verification mask data.

[0036] Preferably, before a feature node sends its mask data, a group mapping anti-collision process needs to be performed to avoid new collisions caused by a large number of feature nodes sending HCMs at the same time. The process is as follows: the h micro-slots used for verification elimination are equally divided into multiple transmission groups. Then, the lower n bits of the numerical data in the k-bit random sequence generated by the feature node during the contention phase are extracted. The lower n bits of the numerical data are then converted and mapped to a corresponding group number identifier. Through this mapping, each feature node is uniquely assigned to a specific micro-slot group and strictly sends HCM information containing verification mask data M within the elimination micro-slot corresponding to its group number identifier, realizing non-negotiation, deterministic distributed peak-shifting transmission.

[0037] The core of obtaining the final access right determination lies in the bitwise and arbitration phase. The temporary node remains in a receiving state throughout the entire elimination interval, collecting and demodulating all HCM messages from the feature node and auxiliary feature nodes, extracting n verification mask data from the messages, forming a set {M1, M2, ..., M...}. n}. Then, the crucial numerical logic decision is executed: the temporary node performs a bitwise AND operation on the random sequence data R generated in its own competition with each mask data Mi in the set.

[0038] If for all M i The calculation result is R&M i All values ​​are greater than 0, meaning that in the eyes of each arbitrator, the sequence R of the temporarily reserved nodes overlaps with the highest priority signal they recognize, thus confirming that there are no hidden terminal conflicts. At this point, a result representing the final winning access right determination is generated.

[0039] Conversely, if there exists any verification mask data M x This makes the calculation result R&M x A value of 0 indicates that at least one arbitrator received a hidden terminal signal with a higher priority than the temporarily reserved node, which the temporarily reserved node failed to detect. At this point, an access right determination result representing elimination is generated, and the temporarily reserved node immediately relinquishes its access in this round and returns to its initial waiting state. According to a further improvement of this application, if network sparsity leads to a scarcity of unaccessed feature nodes, already accessed auxiliary feature nodes will retransmit their observed masks as supplementary verification information. The temporarily reserved node must simultaneously satisfy the verification conditions of all feature nodes and auxiliary feature nodes to win, thereby ensuring the coverage of hidden terminal elimination in extreme cases.

[0040] S4: When the judgment result is a win, generate and broadcast lightweight state snapshot data representing the changes in resource occupancy status within the current time frame, and perform automatic release processing of suspicious time slots based on the activity monitoring of occupied time slots.

[0041] Once a node obtains final access rights, it announces its access with minimal signaling overhead and synchronizes the time slot occupancy status across the entire network. At the same time, it reclaims dead time slot resources through an autonomous monitoring mechanism.

[0042] Nodes that successfully connect first generate and broadcast lightweight time frame state information. To reduce signaling overhead, this scheme employs an incremental snapshot mechanism. Nodes retrieve two sets of data from their memory: one is the network-wide time slot state vector data S generated at the end of the previous time frame before the current contention occurred. prev Secondly, after successful access at this node and a change in the time slot occupancy status, the current network-wide time slot status vector data S is updated in real time. curr Both vectors are one-dimensional arrays, with indices representing slot IDs and values ​​indicating whether the slot is idle or occupied. Next, the node pairs S... prev and S currElement-by-element differential comparison processing filters out all slot IDs where state values ​​have changed. These IDs, along with the new state, are extracted to generate a slot set containing only change information, i.e., the incremental state snapshot △S. Finally, this minimal data set △S (reduced from O(N) to O(△N)) is encapsulated into a time frame state information data packet and broadcast in the newly occupied data slots of the node. After receiving this information packet, neighboring nodes only need to partially update a small number of corresponding entries in their state tables to maintain network-wide state synchronization.

[0043] Finally, to ensure the release of dead time slots that are no longer used due to node departure or failure, automatic release processing for suspicious time slots is performed. Each node in the network maintains an independent activity counter locally for each other time slot marked as occupied. In each subsequent time frame period, the node monitors the channel at the physical layer to check whether it has received the time frame status information broadcast or data payload that the time slot owner should periodically send. If detected, its corresponding activity counter is reset. If no activity counter is detected for several consecutive periods, the activity counter value of a time slot will continuously accumulate and eventually reach a preset idle frame number threshold N. idle If a node determines that the owner of a time slot has expired, it will automatically change the local status data of the time slot from occupied to suspected idle and release the time slot resource to the idle resource pool. Afterwards, any node with access needs can re-compete for the time slot, thus achieving robust time slot resource reclamation without centralized control, greatly improving the long-term resource utilization of the network in highly dynamic environments.

[0044] According to another aspect of this application, the competition bit configuration parameters for dynamically adjusting the competition dimension are determined, including:

[0045] Collect and count the total number of signaling time slots that collide within a sliding window of length W, and use this count as the collision count data.

[0046] In this step, the node to be accessed first constructs a sliding observation window for continuously monitoring channel contention status. Specifically, the sliding observation window is defined as a first-in-first-out queue structure with a fixed length of W consecutive signaling slots. When each new signaling slot ends, the node pushes the collision status observation value of that slot into the queue and removes the oldest observation value, thus ensuring that the window always covers the historical data of the most recent W signaling slots.

[0047] For each signaling time slot, a node performs carrier sensing and energy detection at its physical layer to collect signal state data on the channel within that time slot. Within a single signaling time slot, multiple nodes may simultaneously transmit contention signals. When multiple signals overlap in both the time and frequency domains, preventing the receiver from correctly distinguishing any single complete signal through energy detection or preamble matching, the node determines that a signal collision event has occurred within that signaling time slot. Judgment criteria may include: the energy level of the received signal consistently exceeding a preset busy / idle threshold, but correlation detection of the preamble or training sequence consistently fails.

[0048] The node generates a binary collision indicator for the current signaling slot. A collision is marked as 1, and a collision is marked as 0 if the channel is idle or a single signal successfully occupies it. Then, the node performs a cumulative summation operation on the collision indicator indicators for all W signaling slots within the sliding window. This operation iterates through each element in the window queue, counting the number of elements with a value of 1, and finally calculates an integer value, which is the number of collisions within the observation window, denoted as N. collision This intuitively reflects the absolute frequency of access conflicts under the current network density and contention parameter configuration.

[0049] The collision rate of the current window is calculated by performing a ratio operation between the number of collisions and the window length W.

[0050] After obtaining the collision count data N collision Next, the node needs to transform it into a normalized index that is independent of the window length so that it is comparable across different observation scales. To do this, the node performs a ratio calculation.

[0051] Specifically, the node retrieves the preset fixed length value W of the sliding window from its memory, as well as the collision count data N obtained through cumulative summation. collision Then, using both as operands, perform a division operation, with the formula being P. collision = N collision / W.

[0052] This operation converts discrete collision event counts into a probability value that takes continuous values ​​within the interval [0, 1], i.e., collision rate data. For example, if the window length W is set to 100 signaling slots, and 23 collision events are counted in the most recent 100 slots, the calculated collision rate data is 0.23, indicating that 23% of the signaling slots are currently wasted due to collisions. The collision rate data quantifies the degree of insufficiency in the current competition for random number space and the level of waste of channel resources.

[0053] Furthermore, in order to avoid drastic parameter fluctuations due to single statistical fluctuations, in a preferred embodiment, the node can also perform smoothing filtering (such as exponential weighted moving average) on the collision rate data obtained from multiple consecutive calculations to generate a more stable average collision rate data, which can be used as input for subsequent threshold comparison.

[0054] The collision rate data is compared with preset high and low collision thresholds to generate adjustment decisions.

[0055] In this embodiment, the node executes a double-threshold decision logic with hysteresis based on the calculated collision rate data to generate an adjustment decision on the competing bit configuration parameters.

[0056] First, the node reads two preset threshold parameters from its non-volatile memory: a high collision threshold T1 (also known as a low collision threshold T2). These thresholds are set according to engineering experience: T1 is typically set to a high value, such as 0.5 or 0.6, indicating that the collision rate is unacceptably high and the current contention success rate is extremely low; T2 is typically set to a low value, such as 0.08 or 0.1, indicating that the channel is relatively idle and the current contention bit configuration may have redundant overhead. Simultaneously, the node also needs to preset a hard boundary value for the contention bit width k, i.e., an upper limit K. max and lower limit K min This ensures that no matter how extreme the collision rate, the value of k will not exceed the signaling overhead that the system can tolerate.

[0057] Next, the node will calculate the collision rate data P. collision The values ​​of P are compared with those of T1 and T2 in turn: if P collision If T1 is greater than or equal to T1, it indicates that the current competition dimension is too low, and the combination space size of the binary random sequence (2 to the power of k) is insufficient to effectively distinguish a large number of concurrently accessing nodes, leading to frequent collision events where two or more nodes generate the same random number. At this point, the node generates its first adjustment decision, the instruction of which is to increase the value of the competition bit configuration parameter k by 1, and the increased value must not exceed the upper limit K. max That is, the objective k = min(k+1, K) max If P collision Less than or equal to T2, and the current k value is still greater than the preset lower limit K. min This indicates that the current network load is relatively light, and an excessively large k value leads to unnecessary contention for micro-slots, reducing signaling efficiency. At this point, the node generates a second adjustment decision, with the instruction to decrease the value of k by 1, and the decreased value must not be lower than K. min That is, the target k = max(k-1, K) min If P collisionIf the value is between T1 and T2, it indicates that the current competitive bit configuration is basically matched with the network load, the node generation decision is maintained, and the k value remains unchanged.

[0058] Through this bidirectional, bounded adaptive adjustment mechanism, the network can rapidly expand the contention space to suppress conflicts during periods of dense node access and shrink the contention scale to save signaling resources during periods of sparse access.

[0059] According to another aspect of this application, generating temporary state data representing the node's victory in this competition includes:

[0060] A binary random sequence data R of length k with a contention bit configuration parameter is generated using a pseudo-random number generator. Each bit of the R is then mapped sequentially to k consecutive contention micro-slots in the time domain, from the most significant bit to the least significant bit.

[0061] In this embodiment, the node to be accessed first completes the generation and time-series mapping of the competition credential, establishing a time-data correspondence for subsequent distributed arbitration.

[0062] Specifically, the node invokes its internal pseudo-random number generator. This generator can be a hardware circuit based on a linear feedback shift register, or a software module that generates a random bit stream through cryptographic hashing iteration using an initial seed. The generator's initial seed can be derived from the XOR value of a combination of the node's unique hardware identifier (such as a MAC address) and the current frame number, ensuring that the random sequences generated by each node are both unpredictable and statistically independent of each node's sequences.

[0063] The pseudo-random number generator outputs a binary bit stream of length k. The node arranges these k bits sequentially according to their generation order, forming a k-bit binary random sequence data R. Each bit in the sequence has a definite bit weight: the most significant bit (the k-th bit) has the largest binary reduction value, and the least significant bit (the 1st bit) has the smallest reduction value. This sequence data R will serve as the node's unique priority identifier in this round of competition.

[0064] Subsequently, the node performs time-series mapping processing. The contention interval consists of k consecutively arranged contention micro-slots, each with a unique index on the time axis, from the 1st micro-slot to the kth micro-slot. The node maps each bit of the binary random sequence data R, from the most significant bit to the least significant bit, to a micro-slot index from 1 to k. The specific mapping rule is as follows: the most significant bit (the kth bit) of R is mapped to the 1st contention micro-slot of the contention interval, the (k-1)th bit is mapped to the 2nd contention micro-slot, and so on, until the least significant bit (the 1st bit) of R is finally mapped to the kth contention micro-slot of the contention interval.

[0065] Once this mapping is complete, the node establishes an accurate time-bit correspondence table, and subsequent bit-by-bit contention and elimination will be strictly executed according to this table.

[0066] In the j-th micro-time slot, read the j-th bit value of the sequence data R and perform conditional transmit / receive processing: if the bit value is 1, send a contention signal; if it is 0, collect the channel energy data of the current micro-time slot.

[0067] In this embodiment, the node enters the bit-by-bit contention execution phase. For each contentionable micro-slot where the slot index j increments from 1 to k, the node cyclically executes the following conditional send / receive processing.

[0068] First, at the start of each micro-slot, the node reads the bit value corresponding to the micro-slot index j from the sequence data R according to the established timing mapping relationship. This bit value can only be one of two logical states: 0 or 1. Then, based on the read bit value, the node executes two mutually exclusive processing branches:

[0069] If the value of this bit is 1, the node is determined to be an active bidder by the contention rules of the current micro-timeslot. Its transmitter is enabled, and it sends a contention signal of a preset format outward within the micro-timeslot. This contention signal does not need to carry any data information; it can be a preamble sequence with good autocorrelation properties or a simple single-tone pulse. Its only purpose is to inject energy into the channel that can be detected by other nodes, thus announcing its own presence in the contention.

[0070] If the value of this bit is 0, the node is identified as a backoff eavesdropper. Its transmitter remains silent within this micro-time slot, and the receiver is enabled and switched to channel-aware mode. The receiver uses an analog-to-digital converter to collect the superimposed magnetic field strength or power spectral density of all signals on the channel within the current micro-time slot, generating a set of channel energy sampling data. This channel energy data is a scalar or short vector representing the busy level of the channel within this micro-time slot.

[0071] The collected channel energy data is subjected to threshold determination processing. If the energy value is higher than the preset threshold, collision indication data representing the detected collision signal is generated.

[0072] This embodiment is executed only when the node is in a backoff listening state (i.e., the current bit is 0). The node performs further threshold determination processing on the collected channel energy data to identify whether other nodes are sending contention signals in micro-time slots that it is not participating in.

[0073] The node reads a preset energy detection threshold value E from its configuration parameter storage area. thThe threshold value is determined in advance through channel measurement and calibration experiments before system deployment. It needs to be higher than the power level of ambient noise and conventional interference, but it also needs to ensure that it can sensitively detect contention signal pulses from other nodes.

[0074] Then, the node compares the collected channel energy data (usually its average or peak value) with the threshold value E. th A size comparison is performed. This comparison yields two possible results: if the channel energy data is lower than or equal to E... th This indicates that the channel is essentially idle during this micro-slot being monitored by the node, with no other nodes transmitting signals. The node determines that no collision has occurred, generates an indication that the channel is idle, and continues to compete for the next micro-slot. If the channel energy data is higher than E... th This indicates that at least one other node is sending a contention signal within that micro-time slot. Since the node's own current bit is 0, this means there is another node sending a 1 within the same micro-time slot. In the semantics of binary contention rules, this is equivalent to the existence of a higher-priority contender. At this point, the node internally generates contention indication data representing the detected conflict signal.

[0075] When a node's own bit is 0 and it generates conflict indication data, the distributed elimination logic is triggered, an elimination instruction is generated, and the signal transmission of all subsequent micro-slots is stopped.

[0076] In this embodiment, based on the determination result, the node performs a critical elimination process to decide whether it should continue to participate in the competition.

[0077] Specifically, the node continuously maintains a Boolean liveness flag, initially set to true. During the processing cycle of the j-th micro-slot, the node synchronously checks two conditions: first, whether the j-th bit of its own sequence R in the current micro-slot is 0; and second, whether collision indication data has been generated.

[0078] A node discovers a competitor with a 1 in a high-order bit of its own sequence only if both conditions are met. According to binary comparison rules, a 1 is greater than 0, therefore this competitor has a higher priority than the current node. At this point, the node executes distributed elimination logic: it generates an elimination instruction, immediately sets the liveness flag to false, and forcibly disables the transmitter's signal transmission function for all subsequent contention slots within this contention cycle. After receiving the elimination instruction, the node can still continue listening to subsequent slots (for maintenance or observation), but under no circumstances will it transmit contention signals, effectively relinquishing its right to compete in this round.

[0079] Conversely, if a node's current bit is 1 during the entire j-th micro-slot processing cycle, or if it is 0 but no conflict indication data is generated (i.e., energy is below the threshold), the node determines that it is alive at that bit position, does not generate an elimination instruction, and continues to execute the processing cycle of the next micro-slot (j+1).

[0080] After traversing k contention micro-slots, nodes that have not generated any eviction instructions are marked as temporary nodes, and corresponding temporary status data is generated.

[0081] After the processing cycle of the last contention micro-slot (the kth micro-slot) in the contention interval ends, the node performs the final winner determination and status marking process.

[0082] The node checks the status of its internal liveness flag. If, during the traversal of all k contention-slot micro-slots, the node has never generated an elimination command due to meeting the elimination conditions (its own bit is 0 and a collision is detected), meaning the liveness flag remains true, then the following fact holds: in the binary bitwise comparison sequence, the node's random sequence R is numerically at least no less than any of its competitors that directly or indirectly collide with it in the same micro-neighborhood. In other words, the node's random sequence is one of the largest in local contention.

[0083] At this point, the node identifies itself as a temporary node and generates corresponding temporary state data. Inside the node, it switches the protocol state machine from the contention state to the waiting-for-verification state, carrying a complete backup of the original random sequence data R used by the node in the contention.

[0084] It should be noted that due to the broadcast nature of wireless channels and the existence of hidden terminals, multiple temporary nodes located outside each other's communication range may simultaneously gain access. Therefore, the temporary status here is merely a qualification certificate, not a final access permit, and still requires verification through subsequent steps.

[0085] According to another aspect of this application, obtaining verification mask data generated by feature nodes and auxiliary nodes within the network includes:

[0086] The nodes in the network that can simultaneously receive signals from multiple temporary nodes are designated as feature nodes. These feature nodes collect signal reception status data for each micro-time slot within the contention interval.

[0087] In this embodiment, the network performs natural filtering and data collection of feature nodes.

[0088] In an environment with multiple temporary nodes coexisting, these nodes act as hidden terminals for each other, unable to directly detect each other's presence through carrier sensing. However, due to the geometric characteristics of the network topology, there must exist some third-party nodes physically located between or near these hidden terminal pairs. These nodes are within the one-hop communication range of multiple temporary nodes and can simultaneously receive signals from them. Nodes with such special topological conditions are defined as characteristic nodes.

[0089] Throughout the entire contention period, all nodes in the network, regardless of whether they are actively participating in the contention, continuously maintain a signal reception state at the physical layer. For each potential feature node, within each contention micro-slot, they acquire signal reception state data through analog-to-digital conversion and correlation detection. This signal reception state data is not specific bit content, but a binary indicator recording whether a valid contention signal was detected within the micro-slot. Valid signal detection is achieved through an energy threshold combined with preamble correlation peak detection; if the correlation peak exceeds a set threshold, a signal is determined to exist within the micro-slot.

[0090] These signal reception status data are recorded in a buffer of length k in micro-time slots, forming a complete reception status sequence that corresponds one-to-one with the timing of the competing micro-time slots.

[0091] Based on the signal reception status data, generate k-bit verification mask data M. The generation rule is: if a valid signal is received in the j-th micro-time slot, the j-th bit of the verification mask data M is set to 1; otherwise, it is set to 0.

[0092] In this embodiment, the feature node performs format conversion and encoding processing on the collected signal reception status data to generate structured verification mask data M.

[0093] Specifically, the feature node creates a blank mask data register of length k bits, initially with all bits set to 0. Then, it iterates through the k contention micro-slots of the entire contention interval, incrementing the step variable j from 1 to k. In each iteration, the feature node reads the signal reception record corresponding to the j-th micro-slot from its signal reception status data buffer. According to a preset encoding generation rule, the feature node judges the record value and performs a bit-setting operation: if the record value is valid (i.e., a signal is detected), the feature node determines that within that micro-slot, there is at least one higher-priority node within its signal coverage area (this node's random bit in that micro-slot is 1). At this time, the feature node assigns the j-th bit of the mask data M to 1. If the record value is idle (i.e., no signal is detected), the feature node determines that within that micro-slot, the corresponding bits of all potential competitors within its signal coverage area are 0. At this time, the j-th bit of the mask data remains 0.

[0094] After the traversal is complete, the k-bit string in the register constitutes the verification mask data M. Mask M represents the logical OR of the random sequences of all competing temporary nodes in the local electromagnetic environment, as seen from the third-party perspective of the feature node. In other words, all the bits that are 1 in mask M represent the bit pattern of the random sequence possessed by the virtual competitor with the strongest signal, from the perspective of the feature node.

[0095] The auxiliary feature nodes that have successfully connected to the network also collect and generate auxiliary verification mask data.

[0096] This embodiment performs parallel data acquisition and mask generation processing for auxiliary feature nodes to compensate for the insufficient verification coverage that may result from relying solely on unconnected feature nodes.

[0097] During network operation, nodes that have successfully acquired time slot resources and are in normal transmission and reception mode have the same hardware capabilities as nodes that have not connected, and also possess the ability to listen to the channel. These nodes are defined as auxiliary feature nodes. When a new signaling time slot arrives, the auxiliary feature nodes will reuse their receivers and maintain a listening state throughout the contention interval, collecting signal reception status data micro-slot by micro-slot.

[0098] Then, based on the collected reception status data, the auxiliary feature node generates auxiliary verification mask data belonging to that auxiliary feature node. The core difference between the auxiliary feature node and the ordinary feature node lies in the timing and channel of their message transmission: the ordinary feature node transmits the HCM in the subsequent dedicated elimination micro-slot, while the auxiliary feature node's supplementary verification mask data is encapsulated in the accompanying control signaling in its allocated data slot and transmitted in a piggyback manner.

[0099] By introducing supplementary verification with auxiliary feature nodes, this solution effectively addresses the pain point in highly dynamic sparse networking scenarios where the number of unconnected nodes that can be used as ordinary feature nodes is scarce due to geographical location, making it impossible to fully cover all hidden terminal pairs, thus enhancing the robustness of the verification mechanism.

[0100] According to another aspect of this application, a cooperative elimination verification process is performed on the verification mask data and the competing sequence data generated by itself, including:

[0101] The temporary node collects all verification information from the feature node and auxiliary feature node, and extracts a set of n verification mask data {M1, M2, ..., M}. n}

[0102] In this embodiment, the temporary node performs the aggregation and data extraction processing of multi-source verification information.

[0103] After the contention period ends, the temporary node does not immediately declare access. Instead, it tunes its receiver to the immediately following hidden terminal elimination period and maintains a receiving state throughout the entire period. This elimination period contains h micro-time slots used to carry HCM messages from various feature nodes. At the same time, the temporary node also listens for accompanying control signaling from auxiliary feature nodes in the data time slots of the subsequently accessed nodes.

[0104] For each HCM message received within the elimination interval, the temporary node performs demodulation and decoding processing, extracting the k-bit verification mask data M generated by that feature node from the message's payload field. i It is then stored in a temporary mask set buffer.

[0105] For auxiliary verification information detected in the occupied data time slot, the temporary node performs the same extraction operation to obtain auxiliary verification mask data and appends it to the same mask set buffer.

[0106] At the end of the entire elimination interval and the specified auxiliary information collection window, the temporary node checks its buffer. Assuming it has successfully received and parsed n valid masked messages from all listening sources, the mask set buffer stores n k-bit verification mask data. The temporary node organizes this data into a set data structure, denoted as {M1, M2, ..., M...}. n}

[0107] Each extracted mask data Mi is processed by performing a bitwise AND operation with the random sequence data R generated during the competition.

[0108] In this embodiment, the temporary node takes its own competing credentials and external verification information as input to perform core, decisive numerical logic operations.

[0109] First, the temporary node retrieves two key data objects from memory: one is the k-bit binary random sequence data R of this node; the other is the set {M1, M2, ..., Mn} consisting of n verification mask data. n}

[0110] Then, the temporary node initiates a loop iteration process, with the loop index i traversing from 1 to n. In each iteration, the node combines the random sequence data R with the current mask data M. i The two bit operands, of equal length, are fed into the arithmetic logic unit (ALU) for a bitwise AND operation. The rules for the bitwise AND operation are as follows: for each bit from the 1st bit to the kth bit, if the corresponding bit of R is 1 and the corresponding bit of Mi is also 1, then the corresponding bit of the result is 1; otherwise, the corresponding bit of the result is 0. After the operation, a k-bit result value of the same length as the corresponding bit is generated.

[0111] The node performs further zero-value checks on the result value: it checks if at least one bit in the k bits of the result value is 1. If any bit is 1, it means R&M i The arithmetic value of R is greater than 0, indicating that R and M are related. i All bits are 1, meaning the sequence R of the temporary nodes and the mask M of the feature nodes are both 1. i There is an intersection. If all bits are 0, the result of the operation is equal to 0, indicating that R and M have an intersection. i There is no overlap at any bit position, meaning that the signal represented by the sequence R of the temporary node has no corresponding counterpart in the strongest virtual signal observed at the feature node.

[0112] If for all M i All satisfy the calculation result R&M i If the value is greater than 0, then a result representing the final winning access right will be generated.

[0113] In this embodiment, the temporary node performs a summarizing logical conjunction judgment on the generated bitwise AND operation results to generate the final winning conclusion.

[0114] After the loop iteration ends, the temporary node maintains a decision flag array, which records the decision flags for each M. i Does it meet R&M requirements? i The condition is > 0. This embodiment performs a global conjunction process: checking whether all elements in the flag array are true.

[0115] If for each mask M in the set i The bitwise AND result of the result with R is greater than 0, indicating that an important global fact has been unanimously confirmed by a third party: from the perspective of all arbitrable feature nodes and auxiliary feature nodes, the random sequence R generated by the temporary node coincides with the highest priority signal pattern detected by each of them within their respective local ranges at least one bit; this excludes the possibility that there is another node with a higher priority whose signal was captured by the feature node (reflected in M), but the temporary node failed to perceive it (its R is 0 at this bit, resulting in no intersection with M). In other words, the conjunction condition is true, that is, it is determined that there is no hidden terminal with a higher priority than the temporary node.

[0116] At this point, this embodiment generates a result representing the final winning access right determination, which is a decisive internal signal that transitions the protocol state machine of the driving node from the waiting verification state to the successful access state.

[0117] If there exists any M x This makes the calculation result R&M x If == 0, then a result representing the elimination of access rights will be generated, and the user will withdraw from this competition.

[0118] In this embodiment, the temporary node performs the detection and processing of elimination conditions.

[0119] During the traversal of the flag array, once any mask M is found... x If, (regardless of whether it originates from a feature node or an auxiliary feature node), the bitwise AND operation between the feature node and its own random sequence R equals 0, then the global conjunction condition is immediately violated. At this point, the decision logic inside the node immediately captures this event.

[0120] The result of the calculation is R&M x == 0 indicates that the transmit mask M x A third-party node, within one or more micro-time slots of its monitored contention interval, detected a valid contention signal (therefore M). x The corresponding bit of the temporary node is 1), but the random sequence R of the temporary node is exactly 0 in all these bits. This proves that there is at least one other node in the network that sent a 1 signal in those micro-slots, while the temporary node itself did not send a signal because its corresponding bit is 0. Since the binary race rule determines that the sender of 1 has higher priority, this means that in the characteristic node M... x Within the coverage area, there exists a hidden competitor with a higher priority than the current temporary node, and this competitor has not been detected by the temporary node's listening mechanism.

[0121] Therefore, an event with a result of 0 is interpreted as the hidden terminal being exposed. At this point, this embodiment generates an access right determination result representing elimination. Upon receiving this elimination result, the temporarily reserved node immediately performs the exit process: relinquishing all qualifications obtained in this signaling time slot, clearing its temporary state data and random sequence R, reverting the protocol state machine to the initial idle waiting state, and beginning preparations for the next possible round of competition.

[0122] According to another aspect of this application, lightweight state snapshot data characterizing changes in resource occupancy status within the current frame is generated and broadcast, including:

[0123] The node that successfully connects obtains the network-wide time slot state vector data S from the previous frame. prev And the network-wide time slot state vector data S of the current frame curr .

[0124] In this embodiment, the node that successfully passes verification and obtains final access rights begins to execute its first resource management task after joining the network. The first step of this task is to obtain two key basic data sets used to generate state synchronization information.

[0125] Each node maintains a network-wide timeslot status database in its local memory. This database is organized in tabular form, with the table indexed by the timeslot identifier (Slot). ID The table values ​​represent the current occupancy status of the time slot (e.g., enumerated values ​​such as "idle", "occupied", "conflicting", etc.). This database is a one-dimensional vector structure whose length is equal to the total number of data time slots contained in a time frame.

[0126] Specifically, the node first retrieves the copy of data that it persistently stored by performing a snapshot save operation on its database at the end of the previous time frame period. This data records which node address each time slot was assigned to, or marked as idle, under the global observation of the previous frame. This data is called the network-wide time slot state vector data S of the previous frame. prev Then, the node obtains the real-time database status at the current moment. After successfully connecting and possibly receiving the latest FI broadcast from other nodes, its local database has been updated accordingly. The node serializes each entry in the current database to generate the current network-wide timeslot state vector data S. curr .

[0127] S prev and S curr These two vector data have exactly the same dimensions and index order, and their differences precisely reflect which time slot resources in the network have undergone state changes in the most recent signaling time slot period.

[0128] For S prev and S curr Perform differential comparison processing to extract the time slot set data where the state has changed, i.e., the incremental state snapshot △S.

[0129] In this embodiment, the node is represented by S. prev and S curr As input, an efficient incremental difference extraction algorithm is executed to minimize the amount of state synchronization data that needs to be broadcast externally.

[0130] The node starts a loop, with the loop index incrementing from 0 to the total length of the vector minus one. In each iteration, the node reads S... prev and S curr The state values ​​of the two vectors are at the same index position (i.e., slot ID). Then, an element-wise comparison operation is performed on these two state values.

[0131] If the comparison result shows S prev [Slot ID ] and S curr [Slot IDIf the state values ​​are the same, the state of that time slot has not changed, the node ignores it, does not perform any additional processing, and continues to the next loop.

[0132] If the comparison results show that the two state values ​​are different, it indicates that the time slot has undergone a state transition during this contention cycle, such as changing from "idle" to "occupied" or from "occupied" to "idle". At this time, the node will assign the time slot identifier "Slot". ID Together with its presence in S curr The new state value, State, is used as a tuple data item (Slot). ID , State), add to an initially empty result set.

[0133] After traversing all time slots, the result set contains information on all and only those time slots where the state changed. This set of data is the incremental state snapshot, denoted as ΔS. In large-scale networks, the time slots where the state changes in each cycle usually account for only a small portion of the total number of time slots. Therefore, the amount of data in ΔS is much smaller than the complete network state vector (reduced from O(N) to O(ΔN)), effectively compressing signaling overhead.

[0134] The incremental state snapshot △S is encapsulated into a time frame state information data packet and broadcast.

[0135] In this embodiment, the node encapsulates the generated differential dataset using a protocol and sends it to all its one-hop neighbors via a wireless channel.

[0136] First, the node invokes its protocol stack's packet assembly function to create a new time frame state information data packet. The packet's header contains necessary fields such as a synchronization preamble, source node address, and packet type identifier. In the payload field, the node serializes the incremental state snapshot △S set data according to a preset TLV (Type-Length-Value) or similar structured format. The encoded content clearly identifies the ID of each changing time slot and its latest state.

[0137] Then, based on its newly acquired data time slot allocation table, the node delivers the time frame status information data packet to the physical layer for modulation, up-conversion, and power amplification at the boundary of the specific time slot it is authorized to send, and finally broadcasts it outward through the antenna.

[0138] All neighboring nodes within one hop of this node, either in a receiving state or periodically listening, will capture this data packet. The neighboring nodes demodulate and decode the packet to extract the ΔS data, and based on this, update only a small number of corresponding entries in their local state database. This incremental synchronization mechanism avoids the enormous overhead of periodically flooding the entire state table, improving control signaling efficiency in high-density networks. Simultaneously, if a new contention-driven bit configuration parameter k is generated, this parameter data will also be encapsulated in this data packet, achieving seamless parameter synchronization.

[0139] According to another aspect of this application, the automatic release process for suspicious time slots includes:

[0140] Maintain an activity counter for each occupied time slot recorded locally.

[0141] This embodiment describes the infrastructure construction of a distributed time-slot resource activity monitoring mechanism.

[0142] In the local memory of each network node, in addition to maintaining a time slot occupancy status table, a dedicated activity counter is associated with and maintained for each time slot marked as "occupied" in its entries. This counter is an integer variable, initially set to 0, representing the number of consecutive time frame cycles that have elapsed since the last successful reception of valid information from the time slot owner. This counter is maintained independently for each "occupied" time slot, forming an activity count table, which is stored alongside the status table.

[0143] Within each time frame period, monitor whether the time frame status information broadcast or payload data corresponding to the owner of the time slot is received.

[0144] In this embodiment, the node periodically performs monitoring and judgment on the activity of the "occupied" time slots.

[0145] In each new time frame period, after completing the transmission and reception processing of all time slots within the current frame, the node begins to execute a liveness check loop. This loop iterates through all entries in its status table marked as "occupied".

[0146] For each "occupied" time slot and its associated time slot owner address (recorded when the time slot is initially allocated), the node checks its reception records within the current time frame period. Specifically, the node queries whether it has successfully received and decoded valid data from the time slot owner address within the predetermined data time slot window, including but not limited to regular data payloads, or the time frame status information (FI) broadcast packet sent by that owner. Due to its periodic transmission, the FI broadcast packet is the most direct heartbeat signal for determining whether a node is still "alive."

[0147] If any valid signal is received from the owner during this period, the owner of that time slot is considered to still be active.

[0148] If the activity counter in a time slot reaches the preset idle frame threshold N idle If no valid data is detected by then, the status data of the time slot will be changed from occupied to suspicious idle.

[0149] In this embodiment, the node updates the counter based on the monitoring results and makes a preliminary judgment on the state change.

[0150] For each check item in the loop, the processing logic is as follows: If valid data from the time slot owner is successfully received in this frame, the node determines that the time slot is active. At this time, the node performs a counter reset operation, clearing the activity counter associated with the time slot to zero and restoring it to its initial value of 0 to reflect its latest activity status. If no valid data from the time slot owner is received in this frame, the node determines that the time slot may be inactive. At this time, the node performs a counter increment operation, increasing the value of the activity counter associated with the time slot by 1.

[0151] After the accumulation operation is executed, the node immediately updates the counter value and compares it with the preset idle frame threshold N. idle Perform comparison processing. N idle It is a globally predefined constant whose value determines the system's tolerance for a node being silent or leaving; for example, it can be selected as 3, 5, or 10 time frame periods.

[0152] If the updated counter value still does not reach N idle This indicates that the silence time is short, possibly just a brief channel obstruction or sudden interference, and the node does not take further action. However, if the updated counter value is greater than or equal to N... idle This means that the occupant of this time slot has been N consecutive times. idle If a node remains completely silent for an entire period, it can determine that the occupant has likely left the network, experienced a malfunction, or is in a state of prolonged disconnection. In this case, the node proactively performs a state modification operation, permanently changing the status data of that time slot in its local state table from "occupied" to "suspiciously idle".

[0153] Time slot resources marked as suspiciously idle will be automatically released to the idle resource pool for other nodes to compete for again.

[0154] In this embodiment, once a node completes the final action of resource reclamation, the time slots that were determined to be dead will be returned to available resources.

[0155] Once the status data of a time slot is updated to "suspiciously idle," the resource management label for that time slot undergoes a fundamental change. The time slot ID is removed from the node's internal "list of occupied resources" and re-added to a global "idle resource pool." This idle resource pool is the only legitimate input source for a node's competing module when selecting a target time slot.

[0156] Subsequently, when the node needs to send new data or reserve resources for other services, it can compete for the reclaimed time slot as a legitimate target. Simultaneously, in the next time frame period, the time slot's state change from "occupied" to "idle" will be captured by the node's differential comparison processing and encapsulated in the next transmitted incremental state snapshot △S for broadcast. In this way, the information about the released time slot will quickly spread throughout the network. Any neighboring node that hears this FI broadcast will also update its local state table accordingly, marking the time slot as idle. This fully distributed mechanism achieves consistent and timely recycling and reuse of dead resources across the entire network, improving the spectrum utilization efficiency of long-endurance missions.

[0157] This invention solves the access challenge of UAV ad hoc networks under the dual uncertainties of unknown node number and unknown topology through a three-in-one distributed access mechanism of "dynamic competition, collaborative verification, and lightweight synchronization". Its advantages are as follows: First, the conflict-aware k-value adaptive adjustment algorithm enables the competition dimension to match the dynamic changes in network load in real time. It automatically expands the random number space to suppress collisions during dense, bursty node access and automatically shrinks the competition scale to save signaling overhead when nodes are sparse, achieving an autonomous balance between competition efficiency and resource consumption. Second, the hidden terminal elimination mechanism based on feature node grouping mapping and auxiliary node collaboration effectively solves the problem of feature node scarcity and HCM collisions in highly dynamic scenarios by using low-n-bit random number staggered HCM transmission and retransmission verification of already accessed nodes, achieving collaborative verification and conflict suppression in hidden terminal scenarios. Third, the lightweight state snapshot synchronization based on incremental differential comparison and the automatic release mechanism of suspicious time slots based on activity monitoring reduce the network-wide state synchronization overhead from O(N) to O(ΔN), and improves spectrum utilization and network robustness under long-endurance missions through the autonomous reclamation of dead resources. In summary, this invention achieves one-stop distributed access with low conflict, low latency, low overhead, and high reliability without requiring any centralized control or prior knowledge of the network topology. It is especially suitable for highly dynamic, large-scale, and rapidly changing UAV self-organizing network application scenarios.

[0158] The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited to the specific details in the above embodiments. Within the scope of the technical concept of the present invention, various equivalent transformations can be made to the technical solutions of the present invention, and these equivalent transformations all fall within the protection scope of the present invention.

Claims

1. A distributed dynamic access method for unmanned aerial vehicles (UAVs) with unknown number of nodes and topology, characterized in that, include: Acquire channel collision event data of the node to be accessed within a preset observation window, and determine the contention bit configuration parameters used to dynamically adjust the contention dimension; Based on the competition bit configuration parameters, a distributed bit-by-bit competition elimination process is performed to generate temporary state data indicating that the node has won in this competition; The system acquires the verification mask data generated by the feature nodes and auxiliary nodes within the network, and performs collaborative elimination verification processing on the verification mask data and the competing sequence data generated by itself to obtain the final access right determination result. When the judgment result is a win, lightweight state snapshot data representing the changes in resource occupancy status within the current time frame is generated and broadcast, and automatic release processing of suspicious time slots is performed based on the activity monitoring of occupied time slots.

2. The method according to claim 1, characterized in that, Determine the contention bit configuration parameters used for dynamically adjusting the contention dimension, including: Collect and count the total number of signaling slots that collide within a sliding window of length W signaling slots, and use this count as the collision count data. Perform a ratio calculation between the number of collisions and the window length W to obtain the collision rate data for the current window; The collision rate data is compared with preset high and low collision thresholds to generate adjustment decisions.

3. The method according to claim 1, characterized in that, Generate temporary state data representing the node's victory in this competition, including: A binary random sequence data R of length k with a contention bit configuration parameter is generated using a pseudo-random number generator. Each bit of the R is then mapped sequentially to k consecutive contention micro-slots in the time domain, from the most significant bit to the least significant bit. In the j-th micro-time slot, read the j-th bit value of the sequence data R and perform conditional transmit / receive processing: if the bit value is 1, send a contention signal; if it is 0, collect the channel energy data of the current micro-time slot. The collected channel energy data is subjected to threshold determination processing. If the energy value is higher than the preset threshold, collision indication data representing the detected collision signal is generated. When a node's own bit is 0 and it generates conflict indication data, the distributed elimination logic is triggered, an elimination instruction is generated, and the signal transmission of all subsequent micro-slots is stopped. After traversing k contention micro-slots, nodes that have not generated any eviction instructions are marked as temporary nodes, and corresponding temporary status data is generated.

4. The method according to claim 1, characterized in that, Obtain the verification mask data generated by feature nodes and auxiliary nodes within the network, including: The nodes in the network that can simultaneously receive signals from multiple temporary nodes are designated as feature nodes. The feature nodes collect signal reception status data for each micro-time slot within the contention interval. Based on the signal reception status data, generate k-bit verification mask data M. The generation rule is: if a valid signal is received in the j-th micro-time slot, set the j-th bit of the verification mask data M to 1; otherwise, set it to 0. The auxiliary feature nodes that have successfully connected to the network also collect and generate auxiliary verification mask data.

5. The method according to claim 4, characterized in that, Perform collaborative elimination verification processing on the verification mask data and its own generated competing sequence data, including: The temporary node collects all verification information from the feature node and auxiliary feature node, and extracts a set of n verification mask data {M1, M2, ..., M}. n }; Each extracted mask data M i Perform a bitwise AND logical operation on the random sequence data R generated in competition with itself; If for all M i All satisfy the calculation result R&M i If the value is greater than 0, then a result representing the final winning access right will be generated; If there exists any M x This makes the calculation result R&M x If == 0, then a result representing the elimination of access rights will be generated, and the user will withdraw from this competition.

6. The method according to claim 5, characterized in that, Before extracting the verification mask data, the process also includes grouping and mapping feature nodes to prevent collisions. The h elimination micro-slots used for hidden terminal elimination are divided into multiple transmission groups, and the lower n bits of the numerical data in the k-bit random number generated by the feature node during the competition phase are extracted. The low n-bit numerical data is mapped to the corresponding group number identifier, so that the feature node sends information containing the verification mask data M strictly within the elimination micro-slot corresponding to its group number identifier, thereby realizing distributed peak-shifting transmission.

7. The method according to claim 1, characterized in that, Generate and broadcast lightweight state snapshot data representing changes in resource occupancy status within the current frame, including: The node that successfully connects obtains the network-wide time slot state vector data S from the previous frame. prev And the network-wide time slot state vector data S of the current frame curr ; For S prev and S curr Perform differential comparison processing to extract the time slot set data where the state has changed, i.e., the incremental state snapshot △S; The incremental state snapshot △S is encapsulated into a time frame state information data packet and broadcast.

8. The method according to claim 1, characterized in that, Perform automatic release processing for suspicious time slots, including: Maintain an activity counter for each occupied time slot recorded locally; Within each time frame period, monitor whether the time frame status information broadcast or payload data corresponding to the owner of the time slot is received; If the activity counter in a time slot reaches the preset idle frame threshold N idle If no valid data is detected by then, the status data of the time slot will be changed from occupied to suspiciously idle. Time slot resources marked as suspiciously idle will be automatically released to the idle resource pool for other nodes to compete for again.

9. The method according to claim 2, characterized in that, After generating the adjustment decision, the process also includes: encapsulating the updated contention bit configuration parameter k in lightweight state snapshot data and broadcasting it together, so that neighboring nodes can receive and synchronize their local contention bit configuration parameters to ensure the uniformity of contention dimensions across the entire network.

10. The method according to claim 5, characterized in that, When performing collaborative elimination verification, a temporary node must simultaneously satisfy all verification mask conditions from both the feature node and the auxiliary feature node in order to obtain final access rights. If there are few unaccessible feature nodes, the auxiliary feature node will resend the temporary node mask it observes within the data range as supplementary verification information to enhance the coverage of hidden terminal elimination.

Images