Method and system for recovering conflicting data packets under background noise in a multi-gateway LoRa network

Abstract

This invention discloses a method and system for recovering collision packets below the noise floor in a multi-gateway LoRa network. To address the technical problem of eliminating interference and recovering original packets in a LoRa network, the invention first utilizes the incoherence of signals and noise across multiple gateways to detect and extract the frequency characteristics of collision packets hidden within the noise floor. Then, a novel gateway selection strategy is employed to select an appropriate set of gateways for packet recovery based on the packet power domain characteristics extracted from collision detection. This eliminates interference and recovers the original packets, improving packet recovery performance. This invention is suitable for the Internet of Things and wireless communications.

Description

Technical Field

[0001] The present invention relates to the field of wireless communications, and in particular to the field of recovering conflicting data packets below background noise in a multi-gateway LoRa network. Background Art

[0002] Low-power wide-area networks (LPWANs) can provide long-distance, low-power, and low-cost communication services for many large-scale IoT systems. As a representative long-distance communication technology, the LoRa network has been widely studied in academia and industry due to its unique advantages in high openness and flexibility.

[0003] A LoRa packet consists of four parts: a preamble, a synchronization word, a start frame delimiter (SFD), and the payload. The preamble is a series of rising basic symbols (a signal with increasing frequency over time and an initial frequency of 0) that indicates the beginning of the packet. The number of rising symbols is determined by the user. The synchronization word consists of two rising symbols and is used to isolate devices belonging to different LoRa networks. The SFD is 2.25 falling basic symbols (a signal with decreasing frequency over time and an initial frequency at the upper limit of the bandwidth) that separates the payload from the preamble and synchronization word. The payload carries data bytes using rising symbols with a specified initial frequency.

[0004] In a LoRa network, node devices use the ALOHA protocol, which directly transmits data packets without monitoring the channel before sending. In practical applications, nodes are often densely deployed, and the communication range of the nodes is wide. Packet conflicts caused by their concurrent transmission have become one of the main challenges of the LoRa network.

[0005] Existing packet recovery research usually assumes that the colliding packet signal is above the noise floor. However, considering the large-scale and low-power characteristics of LoRa communication, many colliding packets are below the noise floor, so existing packet recovery schemes perform poorly in actual LoRa networks.

[0006] Based on this, the present invention proposes a collaborative data packet recovery mechanism, which aims to recover data packet collisions below the noise floor. Summary of the Invention

[0007] In order to solve or partially solve some or all of the above technical problems, the present invention is achieved through the following technical solutions:

[0008] A method for recovering collision packets below the noise floor in a multi-gateway LoRa network, the method comprising the following steps: a node sends a data packet to a gateway; the gateway performs collision detection on the data packet and, upon detecting a data packet collision, attempts to recover the collision packet locally. If local recovery is not possible, the gateway aligns the FFT window and uploads the FFT result and metadata to a server; upon receiving the FFT result uploaded by the gateway, the server searches the record to check whether the collision packet has been received or recovered; if the collision packet has not been received or recovered, the server groups the FFT results from multiple gateways according to the arrival time and collision time of the collision packet, selects a gateway combination that can be used to recover the data packet based on the FFT results, and accumulates the FFT results to recover the collision packet.

[0009] In a preferred embodiment, if the server receives a normal data packet, it records the data packet content and arrival time;

[0010] The metadata includes: data packet arrival time, collision window position C p , and power domain features extracted from the collision packets.

[0011] In a preferred embodiment, the gateway receives the data packet, calculates the FFT result of each symbol of the preamble using a sliding window, and then superimposes consecutive FFT amplitude results to obtain the FFT amplitude sum;

[0012] Using window-by-window FFT results, different FFT modes are used to identify packet collisions.

[0013] A sliding window of 2.25 basic descending symbols is used to search for the window where the cross-correlation value with the data packet reaches a peak value as the starting frame delimiter window, and the collision position C is determined. p ;

[0014] The gateway at least uploads the collision alignment FFT result, collision position C p and the preamble window p, as well as the linear regression results to the server;

[0015] Based on the linear regression results, the index for gateway selection is determined; the server checks the collision position C of the uploaded FFT results p and preamble window p, and select a gateway from the gateway set <C p ,p>The maximum cardinality in the tuple is used as a reference value for collision, and other gateway sets update the indicator for gateway selection according to this reference value;

[0016] The server selects a gateway combination that can be used to restore the data packet, and the indicator used for gateway selection of the FFT result of the target gateway does not exceed a given threshold;

[0017] The server accumulates the magnitude of the selected FFT results window by window to recover the collided packets.

[0018] In a preferred embodiment, the linear regression result includes: a residual indicating noise power, and a slope of the linear regression.

[0019] In a preferred embodiment, the first arriving data packet received in the gateway is P A , the second arriving data packet is denoted as P B ;

[0020] In the process of identifying data packet collisions, the sliding window of the collision signal includes three symbols: P A The leading code basic rising symbol, P A The payload rise sign, and P B The leading code is basically a rising symbol;

[0021] When there are peaks that appear continuously with the moving sliding window, the energy of the peak is recorded at each step;

[0022] When the maximum value of the peak energy is detected, linear regression is performed on the left and right sides of the sliding window position corresponding to the maximum value.

[0023] In a preferred embodiment, in performing the linear regression, the slopes K1, K2, and the residuals σ1, σ2 of two linear regressions are obtained;

[0024] The indicators used for gateway selection are:

[0025] In a preferred embodiment, when the server selects a gateway combination that can be used to recover a data packet, the server first checks the collision position C of the uploaded FFT result. p and the preamble window p; if all FFT results are different from each other, the recovery will be terminated and retransmitted directly; and the server will <C p ,p>The similarity of tuples divides gateways into different sets.

[0026] In a preferred embodiment, for each gateway, the FFT window is enlarged to a value of (f s / B)*f s *T s Determine the fixed window size, where f s represents the sampling rate, B represents the symbol bandwidth, T s Indicates the symbol duration.

[0027] In a preferred embodiment, the FFT window is connected to the P AAfter aligning the signal boundaries, the amplitude of the FFT result is normalized to the range of 0 to 1 before uploading it to the server.

[0028] A system for recovering conflicting data packets below the noise floor in a multi-gateway LoRa network, the system comprising nodes, gateways, and servers, wherein the method for recovering conflicting data packets below the noise floor in a multi-gateway LoRa network as described in any one of the preceding items is applied in the nodes, gateways, and servers.

[0029] The various embodiments disclosed in the present invention each have one or more of the following beneficial technical effects. More beneficial effects will be described in the specific embodiments:

[0030] The multi-gateway collaborative LoRa data packet collision recovery method proposed in the present invention can achieve collision data packet recovery under background noise through (1) a collision data packet detection method based on the incoherence of multi-gateway signals and noise, (2) a new gateway selection strategy, and (3) a novel data alignment and data packet collaborative recovery method. Compared with the existing technology, the data packet recovery effect is better in practical applications.

[0031] The above technical solutions / features are intended to summarize the technical solutions and features described in the detailed description, and therefore the scope of the description may not be exactly the same. However, the new technical solutions and features disclosed in this section, in combination with the technical features disclosed in the subsequent detailed description, disclose more technical solutions.

[0032] The technical solution formed by combining all the technical features disclosed anywhere in the present invention is used for summarizing the technical solution, modifying the patent document, and disclosing the technical solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a schematic diagram of the structure of the conflict data packet recovery system of the present invention;

[0034] Figure 2 It is the normal demodulation process and the demodulation process of collision data packets;

[0035] Figure 3 is the linear regression result of the window-by-window FFT mode;

[0036] Figure 4 It is a schematic diagram of the sliding window forward step process. DETAILED DESCRIPTION

[0037] In conjunction with the various method steps or modules described in the embodiments disclosed herein, they can be implemented in hardware, software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the steps and components of each embodiment have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application or design constraints of the technical solution. A person of ordinary skill in the art may use different methods to implement the described functions for each specific application, but such implementation should not be considered outside the scope of protection claimed by the present invention.

[0038] The serial numbers such as "first" and "second" in any position of the present invention are only used as distinguishing marks for description, and do not imply an absolute order in time or space, nor do they imply that the terms prefixed with such serial numbers and the same terms prefixed with other attributives necessarily have different references.

[0039] Because it is impossible to exhaustively describe various alternative solutions, the following will clearly and completely describe the key points of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Other technical solutions and details not disclosed in detail below generally fall within the technical objectives or technical features that can be achieved through conventional means in the art. Due to space limitations, the present invention will not describe them in detail.

[0040] This disclosure describes various key points that can be combined into various specific embodiments, which will be incorporated into various methods and products. In this disclosure, even if a key point is described only when introducing a method / product solution, it means that the corresponding product / method solution also explicitly includes the technical features.

[0041] When a step, module, or feature is described as existing or included at any position in the present invention, it does not imply that such existence is exclusive and unique. Those skilled in the art can obtain other embodiments based on the technical solution disclosed in the present invention and supplemented by other technical means. Based on the key points described in the specific embodiments of the present invention, those skilled in the art can replace, delete, add, combine, reorder, and other means of certain technical features to obtain a technical solution that still follows the concept of the present invention. These solutions that do not deviate from the technical concept of the present invention are also within the scope of protection of the present invention.

[0042] The present invention focuses on the case of two data packets colliding, and can use the solution of two data packets colliding to sequentially resolve the case of three or more data packets colliding in a recursive manner. Figure 1 Specifically, the technical solution proposed in the present invention first introduces the three main roles of the LoRa network: node (also known as client), gateway and server.

[0043] (1) Node: The present invention does not make any modifications to the LoRa node, so it can be easily deployed in the actual LoRa network.

[0044] (2) Gateway: The gateway needs to perform packet collision detection. Each gateway is able to connect to a large number of clients, collect data packets and forward them to the server. In the present invention, the gateway not only needs to perform LoRa packet decoding, but also needs to detect potential packet collisions. Once a packet collision is detected, the gateway first attempts to recover the colliding data packets locally based on the existing single-gateway packet recovery mechanism. For collisions that cannot be recovered locally, the gateway will call the collaborative recovery function, align the fast Fourier transform (FFT) window and upload the FFT results and metadata (such as packet arrival time, packet collision coordinates, and power domain features extracted from the collision packet) to the server.

[0045] (3) Server: The server needs to perform gateway selection and packet collision recovery. All LoRa gateways are connected and synchronized to the server. When a normal data packet is received, the server records the packet content and arrival time. After receiving the FFT results uploaded by the gateway that detected the packet collision, the server will first search the records to check whether the collision packet has been received or recovered. If the collision packet has not been received or recovered, the server will group the FFT results from multiple gateways according to the arrival time and collision time of the collision packet, select the appropriate gateway combination based on the FFT results, and accumulate their FFT results to recover the collision packet.

[0046] The method for recovering conflicting data packets of the present invention comprises the following stages:

[0047] 1) Collision detection stage

[0048] Denote the first arriving packet as P A , denote the second arriving packet as P B Collision detection has two goals: to identify P B The existence of P B Noise levels higher than the signal level make achieving both of these goals more difficult.

[0049] The present invention uses a sliding window to calculate the FFT result of each symbol of the preamble, and then superimposes the consecutive FFT amplitude results to obtain the sum of the FFT amplitudes. The length of the sliding window is determined by the preamble length of the LoRa data packet. By analyzing the FFT patterns of all possible collision patterns, different conditions are set to distinguish different FFT patterns for collision detection. Specifically, the detection includes the following three processes:

[0050] 1. Accumulate FFT results: The gateway uses an additional FFT window, slides along the signal direction and accumulates the FFT amplitude results of the sliding window. The size of the sliding window is determined by the leading length n of the data packet. pre When sliding into the SFD portion, the window does not skip the last 0.25 basic descending symbols to prevent the PB preamble from ignoring the overlap with these 0.25 basic descending symbols.

[0051] 2. Detect packet collision: Utilize the window-by-window FFT results and use different FFT modes to identify collisions. There are three main types of symbols in the sliding window of the collision signal, namely P A The leading code basic rising symbol, P A The payload rise symbol, P B The leading code is basically a rising symbol. A The peak value of the FFT result of the basic rising symbol of the leading code is always at 0; A The FFT stack peak energy of the payload rising symbol will suddenly increase and maintain until the sliding window moves away; B The peak energy of the FFT stack of the preamble's basic rising symbols first increases and then decreases at the same rate. The present invention uses linear regression to extract these unique energy variation patterns. Once peaks appear continuously as the sliding window moves, the energy of these peaks is recorded at each step. After finding the maximum peak energy, linear regression is performed on the left and right sides of the value.

[0052] Figure 3 The linear regression results of the window-by-window FFT mode are shown. (a) and (b) show the P A The payload and P B The linear regression results of the preamble of . The marked points show the energy changes as the sliding window moves. As shown in (c), P B The side lobes of the preamble can also produce the same pattern, but only the clear symmetric pattern in (b) can be used to detect two collisions. To solve this problem, the present invention identifies the main lobe by comparing their maximum energy with the window (bin). Since the energy of the side lobes is always lower than the main lobe and the window is close to the main lobe, they are regarded as a collision and the window with higher energy is selected as P. B The preamble window position.

[0053] The specific operation of packet collision detection is as follows: First, each gateway records the peak amplitude obtained and performs linear regression on it. Then, the slopes K1, K2 and residuals σ1, σ2 of the two linear regressions are used to distinguish different patterns, and the P is detected by preamble detection. B existence.

[0054] The peak value will be considered as P if and only if |K1+K2|<∈*max(K1,|K2|),|K| / σ>τ B The preamble of , where ∈ describes the symmetric pattern of the two linear regression results, and τ is the threshold for distinguishing the main lobe from the side lobe.

[0055] 3. Locate the collision position: Assume P A and P B The symbol shift between them is α, and the detected P B The preamble window is p. When the sliding window moves to position a, the height of p reaches the maximum value in the cumulative FFT result. At this time, the collision position C p It is given by:

[0056]

[0057] The window position of the preamble symbol can be given by:

[0058] p=round((1-α)*2 SF )

[0059] When p≥2 f -1, α≤0.5, and vice versa. When α≈0.5, (n pre -α)*E and [(n pre –1)+α]*E will be very small. When the signal-to-noise ratio is low, stronger noise may flip the height relationship between the two window positions, resulting in the inference of the wrong collision position. To solve this problem, the present invention uses 2.25 basic descending symbols and P B The correlation between the SFD parts is used to correct C p . Use a sliding window of 2.25 basic descending symbols to p +n pre +2 positions to C p +n pre +4 position search for the window where the correlation value reaches the peak, and find the SFD window C sfd After that, the conflict window is marked as C sfd -n pre -2.

[0060] 2) Gateway selection stage

[0061] In the present invention, the key idea of the gateway selection step for collaborative packet recovery is to select a gateway that contains a signal that is stronger than the noise, thereby avoiding introducing more noise than signal into the cumulative FFT result at the server side. A In the calculation of the signal-to-noise ratio, P BThe signal is treated as noise, so it is not advisable to use the signal-to-noise ratio to select the gateway. Further comparison of P B Peak and noise peaks.

[0062] The present invention utilizes fine-grained collision information to select gateways because this information implicitly conveys the P B This requires the gateway to upload not only the aligned FFT results of the collision, but also the fine-grained information in the collision detection, including the collision location C p And the preamble window p, linear regression results K1, K2, σ1, σ2.

[0063] like Figure 4 As shown, when the sliding window steps forward, the entire symbol of the interfering packet moves into or out of the window at the collision position, which in turn increases or decreases the amplitude of the collision peak. Therefore, the gradients K1 and K2 can represent P B The energy of the preamble code. The larger K1 and K2 are, the higher the P B The higher the signal power, the greater the residual error. Similarly, the residuals σ1 and σ2 can indicate the noise power. Noise will cause the collision peak amplitude to fluctuate. The stronger the noise, the more severe the amplitude fluctuation and the larger the residual error.

[0064] Based on the above indicators, the present invention proposes an integrated indicator M for gateway selection, which implicitly reflects P B The noise-to-signal ratio is:

[0065]

[0066] Through M, we can obtain the interference P B An approximate estimate of the noise-to-signal ratio between the signal and the noise.

[0067] Gateway selection works as follows: First, the server checks the collision position C of these uploaded FFT results. p and preamble window p. If all FFT results are different from each other, the recovery will be terminated and retransmission will be performed. The server will <C p ,p>tuple similarity divides the gateways into different sets. Then, the server selects <C p ,p> is used as the reference value for collision, and other gateway sets are required to update the index M according to this reference value. These gateways need to find the actual <C p ,p>The amplitude change of the peak. After performing linear regression again, the gateway updates its index M. The server then selects a suitable gateway whose index M of the FFT result does not exceed the given threshold τ gsThe server then accumulates the magnitude of the selected FFT results window by window. By applying the existing local decoding scheme to the sum of the selected FFT results, it is able to recover the collided packets that cannot be recovered at any single gateway.

[0068] 3) Data alignment and data packet collaborative recovery phase

[0069] Due to carrier frequency offset (CFO) and sampling time offset (STO), the FFT window cannot be accurately aligned with the signal boundary, resulting in frequency offset. Existing research uses the basic rising symbols of the preamble and the basic falling symbols of the SFD to determine CFO and STO. Both CFO and STO will cause window offsets in the FFT results. However, CFO will cause the same window shift for the basic rising and falling symbols, while STO will cause the opposite window shift. The estimates of the window shift by CFO and STO are given by the following equations, respectively:

[0070] ΔBin cfo =(Bin up +Bin down ) / 2

[0071] ΔBin sto =(Bin down -Bin up ) / 2

[0072] Where ΔBin cfo Bin value offset caused by CFO, ΔBin sto The bin value offset caused by STO, the estimation of CFO and STO is determined by the following formula:

[0073] CFO=ΔBin cfo *f r

[0074] STO=ΔBin sto *f r *T s / B

[0075] where f r represents the frequency resolution of FFT, B represents the bandwidth, and T s represents the symbol duration. According to the above equation, the alignment accuracy is highly dependent on the frequency resolution f r After compensating for CFO and STO, the alignment offset is still less than f r *Ts / B. Specifically, when each gateway combines the FFT window with P A The start time t sa When aligned, there is less than T s / 2 SF The local alignment offset ofSF , where T s =2 SF / B is the symbol duration. Although the offset has limited impact on single-gateway LoRa demodulation, it can affect multi-gateway packet recovery. To address this issue, the present invention utilizes FFT zero-padding to improve the frequency resolution of the FFT result and eliminate the ambiguous local time offset. Therefore, P B The accumulated offset can be reduced later. Specifically, each gateway enlarges the FFT window to the value given by (f s / B)*f s *T s A fixed window size is determined to maximize frequency resolution. s represents the sampling rate, and B represents the symbol bandwidth. After that, the bins in the FFT result will be larger than f s / B is twice as large and the frequency resolution is increased. Therefore, the gateway can align the signal boundaries more accurately. A After aligning the signal boundaries, the amplitude of the FFT result is normalized to the range of 0 to 1 before uploading it to the server. Through normalization, the differences between signal amplitudes are eliminated while maintaining the signal-to-noise ratio of different gateways.

[0076] In addition, the present invention also discloses a system for recovering conflicting data packets below the noise floor in a multi-gateway LoRa network. The system includes a node, a gateway, and a server. The method for recovering conflicting data packets below the noise floor in a multi-gateway LoRa network as described in any one of the above items is applied in the node, gateway, and server.

[0077] Although the present invention has been described with reference to specific features and embodiments thereof, various modifications, combinations, and substitutions may be made thereto without departing from the present invention. The scope of protection of the present invention is not intended to be limited to the specific embodiments of the processes, machines, manufactures, compositions of matter, devices, methods, and steps described in the specification, and such methods and modules may also be implemented in one or more related, interdependent, cooperative, or preceding or subsequent products and methods.

[0078] Some of the technical features mentioned in the attached claims may have alternative technical features, or the order of certain technical processes or the order of material organization may be reorganized. After becoming aware of the present invention, a person of ordinary skill in the art will easily conceive of such alternative means, or change the order of the technical processes or the order of material organization, and then adopt substantially the same means to solve substantially the same technical problems and achieve substantially the same technical effects. Therefore, even if the claims explicitly define the aforementioned means and / or order, such modifications, changes, and substitutions should fall within the scope of protection of the claims in accordance with the doctrine of equivalents.

Claims

1. A method for recovering conflicting data packets under background noise in a multi-gateway LoRa network, characterized in that: The method comprises the following steps: The node sends a data packet to the gateway; The gateway receives the data packet, calculates the FFT result of each symbol of the preamble using a sliding window, and then superimposes consecutive FFT amplitude results to obtain the FFT amplitude sum; Using window-by-window FFT results, different FFT modes are used to identify packet collisions. A sliding window of 2.25 basic descending symbols is used to search for the window where the cross-correlation value with the data packet reaches a peak value as the starting frame delimiter window, and the collision position C is determined. p ; The gateway performs collision detection on the data packets and attempts to recover the colliding data packets locally after detecting a data packet collision. If local recovery is not possible, the gateway aligns the FFT windows and uploads the FFT results and collision detection metadata to the server. The gateway at least uploads the collision alignment FFT result, collision position C p and the preamble window p, as well as the linear regression results to the server; If the server receives the FFT result uploaded by the gateway, searching the record to check whether the collision data packet has been received or recovered; If the collision packet is not received or recovered, the server determines the index for gateway selection based on the linear regression result; the server checks the collision position C of the uploaded FFT result. p and preamble window p, and select a gateway from the gateway set <C p The maximum cardinality in the tuple is used as a reference value for collision, and other gateway sets update the indicator for gateway selection based on this reference value; the server also groups the FFT results from multiple gateways according to the arrival time and collision time of the collision data packet, and selects a gateway combination that can be used to recover the data packet based on the FFT results; The server selects a gateway that finds a data packet conflict, and the index used for gateway selection of the FFT result of the target gateway does not exceed a given threshold and is not greater than zero; The server normalizes the amplitude of the accumulated selected FFT results window by window to recover the collided data packets.

2. The method for recovering conflicting data packets under background noise in a multi-gateway LoRa network according to claim 1, wherein: If the server receives a normal data packet, it records the data packet content and arrival time; The metadata includes: data packet arrival time, collision window position C p , and power domain features extracted from the collision packets.

3. The method for recovering conflicting data packets under background noise in a multi-gateway LoRa network according to claim 1, wherein: The linear regression result includes: a residual indicating noise power, and a slope of the linear regression.

4. The method for recovering conflicting data packets under background noise in a multi-gateway LoRa network according to claim 3, wherein: The first arriving data packet received in the gateway is , the second arriving packet is represented as ; In the process of identifying packet collisions, the sliding window of the collision signal includes three types of symbols: The leading code basic rising symbol, The payload rise symbol, and The leading code is basically a rising symbol; When there are peaks that appear continuously with the moving sliding window, the energy of the peak is recorded at each step; When the maximum value of the peak energy is detected, linear regression is performed on the left and right sides of the sliding window position corresponding to the maximum value.

5. The method for recovering conflicting data packets under background noise in a multi-gateway LoRa network according to claim 4, characterized in that: In performing the linear regression, the slopes K1, K2, and residuals of the two linear regressions are obtained. ; The indicators used for gateway selection are: .

6. The method for recovering conflicting data packets under background noise in a multi-gateway LoRa network according to claim 1 or 5, characterized in that: When the server selects a gateway combination that can be used to recover a data packet, the server first checks the collision position C of the uploaded FFT result. p and the preamble window p; if all FFT results are different from each other, the recovery will be terminated and retransmission will be performed; And the server is based on the gateway <C p , p>The similarity of tuples divides gateways into different sets.

7. The method for recovering conflicting data packets under background noise in a multi-gateway LoRa network according to claim 6, characterized in that: For each gateway, the FFT window is enlarged to the size of Determine the fixed window size, where f s represents the sampling rate, B represents the symbol bandwidth, T s Indicates the symbol duration.

8. The method for recovering conflicting data packets under background noise in a multi-gateway LoRa network according to claim 7, characterized in that: Combine the FFT window with P A After aligning the signal boundaries, the amplitude of the FFT result is normalized to the range of 0 to 1 before uploading it to the server.

9. A system for recovering conflicting data packets under background noise in a multi-gateway LoRa network, the system comprising nodes, gateways, and servers, characterized in that: The method for recovering conflicting data packets under background noise in a multi-gateway LoRa network as described in any one of claims 1 to 8 is applied in the node, gateway, and server.

Images