A multi-scenario adaptive intercom communication method and system based on a radio mesh

Abstract

The application relates to the technical field of intercom communication, and discloses a multi-scene adaptive intercom communication method and system based on a radio Mesh, wherein the method comprises the following steps: identifying the scene type code of a current intercom node, and broadcasting a node discovery frame to surrounding intercom nodes; calculating the comprehensive score of a neighbor link quality according to the node discovery frame, and determining a link mode flag word together with the reference signal receiving power detected by a cellular network module; carrying out noise reduction on a first voice signal collected by a microphone according to the scene type code, obtaining a second voice signal, carrying out encryption, generating an encrypted audio frame, and transmitting the encrypted audio frame through a radio Mesh frequency hopping spread spectrum channel or a cellular network channel pointed by the link mode flag word; and the application solves the problem that the existing Bluetooth chain topology single node disconnection leads to full link fracture, and realizes adaptive call quality guarantee in multiple types of differentiated noise scenes such as motorcycle riding, extremely cold skiing, surfing and referee law enforcement.

Description

Technical Field

[0001] This invention relates to the field of intercom communication technology, and in particular to a multi-scenario adaptive intercom communication method and system based on radio mesh. Background Technology

[0002] Wireless intercom communication technology has a wide range of security collaboration needs in various scenarios such as outdoor sports, industrial operations, and sporting events. Traditional Bluetooth intercom technology uses a point-to-point chain-like serial topology. When multiple nodes are talking, if any intermediate node falls out of coverage due to an obstacle, the entire link will break. Moreover, after the link is broken, the user must manually re-pair, which seriously affects the continuity and security of communication in high-dynamic scenarios such as motorcycle riding, skiing, and refereeing.

[0003] The communication range of existing Bluetooth intercom technology is limited by the characteristics of radio wave propagation. In open environments, the maximum communication range is less than 800 meters, and it further decreases to less than 200 meters in complex environments with obstructions such as tunnels, curves, and industrial steel structures. This cannot meet the ultra-long-range communication needs of large-scale group cycling or remote mountainous areas and extremely cold snow-covered mountains where there is no cellular network coverage. Moreover, existing local short-range intercom and wide-area cellular intercom are independent of each other and lack the ability to seamlessly and automatically switch between the two standards. In addition, existing technology lacks the ability to adaptively process audio for different application scenarios. In special noise environments such as the strong wind noise and engine harmonic noise generated by high-speed motorcycle travel, the broadband environmental noise in extreme cold skiing scenarios, and the whistle interference in refereeing scenarios, the intelligibility of voice calls is severely degraded. Furthermore, in scenarios where hands cannot operate the device, such as skiing or surfing, existing technology cannot automatically complete the keyless voice activation transmission. Summary of the Invention

[0004] This invention provides a multi-scenario adaptive intercom communication method and system based on radio mesh. This invention solves the problem of the entire link breaking due to single node disconnection in the existing Bluetooth chain topology, and realizes adaptive call quality assurance in various differentiated noise scenarios such as motorcycle riding, extreme cold skiing, surfing, and referee enforcement.

[0005] In a first aspect, the present invention provides a multi-scenario adaptive intercom communication method based on radio mesh, the multi-scenario adaptive intercom communication method based on radio mesh comprising:

[0006] Identify the scene type code of the current intercom node and broadcast a node discovery frame to surrounding intercom nodes;

[0007] The link mode flag of the current intercom node is determined by calculating the comprehensive score of the neighbor link quality based on the node discovery frame and by comparing it with the reference signal received power detected by the cellular network module.

[0008] The first voice signal collected by the microphone is denoised according to the scene type code to obtain a second voice signal. The second voice signal is then encrypted to generate an encrypted audio frame, which is then transmitted according to the radio mesh frequency hopping spread spectrum channel or cellular network channel indicated by the link mode flag.

[0009] In conjunction with the first aspect, in a first implementation of the first aspect of the present invention, the step of identifying the scene type code of the current intercom node and broadcasting a node discovery frame to surrounding intercom nodes includes:

[0010] Collect background noise signals from the microphone during the silent zone, and calculate the low-frequency energy integral value, mid-frequency energy integral value, and whistle characteristic frequency band energy integral value of the background noise signals;

[0011] When the low-frequency energy integral value exceeds the preset energy detection threshold and is greater than the mid-frequency energy integral value, the output scene type code is "motorcycle riding scene"; when the mid-frequency energy integral value exceeds the preset energy detection threshold and is greater than both the low-frequency energy integral value and the whistle characteristic frequency band energy integral value, the output scene type code is "skiing or surfing scene"; when the whistle characteristic frequency band energy integral value exceeds the preset energy detection threshold and is greater than the low-frequency energy integral value, the output scene type code is "referee enforcement scene".

[0012] Configure the frequency hopping spread spectrum radio frequency module corresponding to the current intercom node based on the scenario type code, and periodically broadcast node discovery frames to surrounding intercom nodes.

[0013] In conjunction with the first aspect, in a second implementation of the first aspect of the present invention, the step of configuring the frequency hopping spread spectrum radio frequency module corresponding to the current intercom node based on the scene type code and periodically broadcasting node discovery frames to surrounding intercom nodes includes:

[0014] Based on the device identifier of the current intercom node, the network seed shared by each node in the intercom group is used as the initial seed of the shift register to complete the configuration of the frequency hopping spread spectrum radio frequency module.

[0015] After completing the configuration of the frequency hopping spread spectrum radio frequency module, a node discovery frame is periodically broadcast to surrounding intercom nodes. The node discovery frame carries the device identifier, network identifier, frequency hopping phase offset, remaining battery power, scene type code, and transmission power level.

[0016] In conjunction with the first aspect, in a third implementation of the first aspect of the present invention, the step of configuring the frequency hopping spread spectrum radio frequency module by using the network seed shared by all nodes in the intercom group as the initial seed of the shift register based on the device identifier of the current intercom node includes:

[0017] Based on the device identifier of the current intercom node, a frequency hopping sequence is generated through a shift register;

[0018] The network seed of the intercom group initiating node broadcast is replaced with the initial seed of the shift register, so that the phase of the frequency hopping sequence of all intercom nodes in the intercom group is aligned, and the aligned frequency hopping sequence is written into the frequency hopping spread spectrum radio frequency module to complete the frequency hopping sequence configuration.

[0019] In conjunction with the first aspect, in a fourth implementation of the first aspect of the present invention, the step of calculating the comprehensive score of neighbor link quality based on the node discovery frame and determining the link mode flag of the current intercom node with the reference signal received power detected by the cellular network module includes:

[0020] The received signal strength, signal-to-noise ratio, and packet error rate carried in the node discovery frame are weighted and summed to obtain a comprehensive score of the neighbor link quality.

[0021] Based on the comprehensive quality score of the neighbor links, a primary path and a backup path are selected among all nodes in the network, and a multi-hop routing table is constructed based on the primary path and the backup path.

[0022] The link mode flag of the current intercom node is determined based on the comprehensive quality score of the neighbor links of the main path in the multi-hop routing table and the reference signal received power detected by the cellular network module.

[0023] In conjunction with the first aspect, the fifth implementation of the first aspect of the present invention further includes:

[0024] Send link keep-alive probe frames to each hop node of the main path;

[0025] The link keep-alive response frames returned by each hop node are continuously counted. When the number of consecutive failures to receive a link keep-alive response frame from the same hop node reaches a preset number, the path is switched from the primary path to the corresponding backup path in the multi-hop routing table.

[0026] In conjunction with the first aspect, in a sixth implementation of the first aspect of the present invention, the step of determining the link mode flag of the current intercom node based on the comprehensive quality score of neighbor links of the primary path in the multi-hop routing table and the reference signal received power detected by the cellular network module includes:

[0027] The average comprehensive score is calculated based on the comprehensive quality score of the neighbor links of the main path in the multi-hop routing table.

[0028] The average comprehensive score is compared with the Mesh link switching threshold, and the reference signal received power periodically detected by the cellular network module is compared with the available power threshold of the cellular network.

[0029] When the average comprehensive score is not lower than the Mesh link switching threshold and the reference signal received power is not lower than the cellular network available power threshold, the link mode flag is set to hybrid mode; when the average comprehensive score is not lower than the Mesh link switching threshold and the reference signal received power is lower than the cellular network available power threshold, the link mode flag is set to radio Mesh mode; when the average comprehensive score is lower than the Mesh link switching threshold and the reference signal received power is not lower than the cellular network available power threshold, the link mode flag is set to cellular network mode; when the average comprehensive score is lower than the Mesh link switching threshold and the reference signal received power is lower than the cellular network available power threshold, the link mode flag remains unchanged from the previous cycle and enters a waiting-for-re-evaluation state.

[0030] In conjunction with the first aspect, in the seventh implementation of the first aspect of the present invention, the step of denoising the first voice signal collected by the microphone according to the scene type code to obtain a second voice signal, encrypting the second voice signal to generate an encrypted audio frame, and transmitting the encrypted audio frame according to the radio mesh frequency hopping spread spectrum channel or cellular network channel pointed to by the link mode identifier includes:

[0031] Based on the scene type code, the first voice signal collected by the microphone is filtered and noise-reduced to obtain the second voice signal;

[0032] The second voice signal is subjected to voice activity detection according to the voice activity detection threshold corresponding to the scene type code, and the voice frames that pass the voice activity detection are extracted.

[0033] The voice frames detected by voice activity are encoded and AES encrypted and packaged to obtain encrypted audio frames;

[0034] The encrypted audio frame is transmitted through the radio mesh frequency hopping spread spectrum channel or cellular network channel indicated by the link mode identifier.

[0035] In conjunction with the first aspect, in an eighth implementation of the first aspect of the present invention, transmitting the encrypted audio frame according to the radio mesh frequency hopping spread spectrum channel or cellular network channel pointed to by the link mode identifier includes:

[0036] When the link mode flag is in radio mesh mode, the encrypted audio frame is written into the transmit buffer of the frequency hopping spread spectrum radio frequency module for directional unicast forwarding until the encrypted audio frame reaches all target nodes in the intercom group.

[0037] When the link mode identifier is cellular network mode, the encrypted audio frame is uploaded to the cloud relay server, and the cloud relay server broadcasts the encrypted audio frame to all registered nodes in the same intercom group room.

[0038] When the link mode flag is in mixed mode, the gateway node, which simultaneously holds a valid radio mesh connection and a valid cellular network connection, performs deduplication and backhaul suppression on the audio stream received from the frequency hopping spread spectrum channel and the audio stream sent from the cloud relay server, and then performs superposition to obtain a mixed audio stream. The mixed audio stream is then re-uploaded to the cloud relay server for distribution to the cellular network side nodes.

[0039] Secondly, the present invention provides a multi-scenario adaptive intercom communication system based on radio mesh, the multi-scenario adaptive intercom communication system based on radio mesh comprising:

[0040] The scene recognition module is used to identify the scene type code of the current intercom node and broadcast the node discovery frame to surrounding intercom nodes;

[0041] The mode detection module is used to calculate the comprehensive score of neighbor link quality based on the node discovery frame and determine the link mode flag of the current intercom node by comparing it with the reference signal received power detected by the cellular network module.

[0042] An audio transmission module is used to reduce noise in the first voice signal collected by the microphone according to the scene type code to obtain a second voice signal, encrypt the second voice signal to generate an encrypted audio frame, and transmit the encrypted audio frame according to the radio mesh frequency hopping spread spectrum channel or cellular network channel pointed to by the link mode flag.

[0043] At the routing topology level, this invention constructs a distributed multi-hop routing table containing primary and backup paths, and introduces a maximum bottleneck path selection mechanism based on the maximum comprehensive quality score of the weakest link neighbor of each candidate path. Combined with a continuously running link keep-alive detection mechanism, this allows the intercom node to quickly switch the audio data frame forwarding path to the backup path when any intermediate node falls out of coverage due to obstruction. This solves the problem of the entire link breaking due to a single node disconnection in the existing Bluetooth chain topology, and supports large-scale nodes in the intercom group to be online simultaneously. At the heterogeneous network integration level, by comparing the average comprehensive score of radio mesh link quality with the mesh link switching threshold, and by comparing the received power of the cellular network reference signal with the available power threshold of the cellular network, the link mode flag is automatically determined and transparent switching between radio mesh mode, cellular network mode, and hybrid mode is driven. This enables two-way voice communication between members of the radio mesh local area network and remote members of the cellular network within the intercom group after mixing and bridging through the gateway node. This extends the limited communication distance of a single radio mesh standard to long-distance communication within the coverage area of ​​the cellular network, and the entire switching process is completely transparent to the upper-layer intercom application. At the audio processing level, by performing high-pass filtering on the first voice signal according to the scene type code, and completing buttonless voice activation transmission according to the voice activity detection threshold corresponding to the scene type code, adaptive call quality assurance is achieved under various differentiated noise scenarios such as motorcycle riding, extreme skiing, surfing, and referee enforcement.

[0044] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained in accordance with the structures particularly pointed out in the description, claims and drawings.

[0045] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0046] Figure 1 This is a schematic diagram of an embodiment of the multi-scenario adaptive intercom communication method based on radio mesh in this invention.

[0047] Figure 2 This is a schematic diagram of the scene type recognition process in an embodiment of the present invention;

[0048] Figure 3 This is a schematic diagram of the link mode detection process in an embodiment of the present invention;

[0049] Figure 4 This is a schematic diagram illustrating the audio encryption and transmission process in an embodiment of the present invention;

[0050] Figure 5 This is a schematic diagram of an embodiment of a multi-scenario adaptive intercom communication system based on radio mesh according to the present invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] The terms "comprising" and "having," and any variations thereof, used in the embodiments of this invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the steps or units listed, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0053] To facilitate understanding of this embodiment, a multi-scenario adaptive intercom communication method based on radio mesh, disclosed in this embodiment of the invention, will first be described in detail. For example... Figure 1 As shown, this method includes the following steps:

[0054] 101. Identify the scene type code of the current intercom node and broadcast the node discovery frame to surrounding intercom nodes;

[0055] 102. Calculate the comprehensive score of neighbor link quality based on the node discovery frame and determine the link mode flag of the current intercom node by comparing it with the reference signal received power detected by the cellular network module;

[0056] 103. Based on the scene type code, the first voice signal collected by the microphone is denoised to obtain the second voice signal. The second voice signal is then encrypted to generate an encrypted audio frame. The encrypted audio frame is then transmitted according to the radio mesh frequency hopping spread spectrum channel or cellular network channel pointed to by the link mode flag.

[0057] In one specific embodiment, such as Figure 2 The process of executing step 101 may specifically include the following steps:

[0058] 201. Collect the background noise signal of the microphone in the silent section, and calculate the low-frequency energy integral value, mid-frequency energy integral value and whistle characteristic frequency band energy integral value of the background noise signal;

[0059] 202. When the low-frequency energy integral value exceeds the preset energy detection threshold and is greater than the mid-frequency energy integral value, the output scene type code is "motorcycle riding scene"; when the mid-frequency energy integral value exceeds the preset energy detection threshold and is greater than both the low-frequency energy integral value and the whistle characteristic frequency band energy integral value, the output scene type code is "skiing or surfing scene"; when the whistle characteristic frequency band energy integral value exceeds the preset energy detection threshold and is greater than the low-frequency energy integral value, the output scene type code is "referee enforcement scene".

[0060] 203. Configure the frequency hopping spread spectrum radio frequency module corresponding to the current intercom node based on the scenario type code, and periodically broadcast node discovery frames to surrounding intercom nodes.

[0061] Specifically, after the current intercom node starts, it first checks whether a valid scene type code is already stored in the non-volatile memory. If a valid preset scene type code is stored in the non-volatile memory, it directly reads the preset scene type code as the current scene type code and calls the corresponding scene parameter set accordingly. If no valid preset scene type code is stored in the non-volatile memory, the automatic recognition mechanism is activated, and the microphone is controlled to collect background noise signals during the silent period. The automatic recognition mechanism is used to output the motorcycle riding scene type code, skiing or surfing scene type code, and referee enforcement scene type code. The coach training scene type code, industrial obstacle scene type code, and further subdivisions between the skiing scene type code and the surfing scene type code are obtained after being pre-configured and written to the non-volatile memory. When no valid scene type code is stored in the non-volatile memory, the automatic recognition mechanism is activated, and the microphone is controlled to collect background noise signals during the silent period. If no voice activation event is triggered within a continuous 500ms time window, and the short-term energy of the microphone input signal... Below the fixed silent threshold ,in, During dBFS, the current time window is determined to be a silent period. After the silent period is confirmed, the microphone acquires background noise samples at a sampling rate of 16kHz and a quantization precision of 16bit. Since 500ms corresponds to 8000 sampling points, the acquired background noise samples are sent to the spectrum analysis module, and a 512-point FFT operation is performed to obtain the background noise power spectral density. The energy integral score in the low-frequency band is expressed as: The energy integral value in the mid-frequency band is expressed as: The energy integral score of the whistle characteristic frequency band is expressed as: .in, Used to characterize the energy level of concentrated engine noise and low-frequency mechanical noise. Used to characterize the energy levels of broadband components of wind noise and wave noise. This is used to characterize the energy level of high-frequency whistle components in the refereeing environment. The energy integral values ​​of each characteristic frequency band, along with a preset energy detection threshold and the relative magnitudes of these integral values, are used as the judgment criteria to improve the stability and resistance to false positives in scene recognition. When the low-frequency energy integral value... Greater than the energy integral value in the mid-frequency band And the energy integral value in the low-frequency band Exceed When the output scene type code is "motorcycle riding scene", that is... This judgment logic corresponds to the background noise characteristics during motorcycle riding, where low-frequency harmonics from the engine and mechanical vibration energy dominate. The mid-frequency energy integral value... Greater than the energy integral value in the low-frequency band And simultaneously greater than the energy integral value of the whistle characteristic frequency band. And the mid-frequency energy integral value Exceed When the output scene type code is skiing or surfing scene, that is... or Because skiing and surfing scenarios have strong similarities in their background noise spectrum, the default setting for automatic identification output is... When further differentiation into skiing or surfing scenarios is needed, the mobile device will pre-configure the subdivision. (The last sentence appears to be incomplete and unrelated to the preceding text. It mentions "whistle characteristic frequency band energy integral value," but without further context, it's impossible to translate accurately.) Greater than the energy integral value in the low-frequency band And the energy integral value of the whistle characteristic frequency band Exceed At that time, the output scenario type code is the referee enforcement scenario, that is... This judgment logic corresponds to the noise distribution characteristics where high-frequency whistle components are relatively prominent in the judicial enforcement environment. When none of the above conditions are met, to prevent the current intercom node from remaining in an undefined scene state for an extended period, the scene type code will revert to the default value. After the scene type code is determined, it is synchronously written to non-volatile memory and used as a unified index key for calling the voice processing parameter set. Based on the scene type code, the frequency hopping spread spectrum RF module corresponding to the current intercom node is configured accordingly, and after configuration, a node discovery frame is periodically broadcast to surrounding intercom nodes, so that surrounding nodes can promptly know the existence status, scene attributes, and RF operating status of the current intercom node. Based on the device identifier of the current intercom node, the network seed shared by all nodes in the intercom group is used as the initial seed of the shift register to complete the configuration of the frequency hopping spread spectrum RF module, so that the current intercom node is consistent with other nodes in the same group on the frequency hopping reference. The current intercom node enters the node discovery broadcast stage and sends node discovery frames to surrounding intercom nodes at a fixed period. The broadcast period is 50ms. The node discovery frame carries fields such as device identifier, network identifier, frequency hopping phase offset, remaining battery power, scene type code, and transmit power level. The device identifier is 48 bits long, the network identifier is 16 bits long, the frequency hopping phase offset is 8 bits long, the remaining battery power is 8 bits long, the scene type code is 4 bits long, the transmit power level is 4 bits long, and the CRC16 is 16 bits long, for a total frame length of 18 bytes. With this configuration, surrounding intercom nodes can discover neighboring nodes upon receiving the node discovery frame, and can also detect them based on the scene type code, transmit power level, and frequency hopping phase offset carried in the node discovery frame.

[0062] In one specific embodiment, the process of performing step 203 may specifically include the following steps:

[0063] (1) Based on the device identifier of the current intercom node, the network seed shared by each node in the intercom group is used as the initial seed of the shift register to complete the configuration of the frequency hopping spread spectrum radio frequency module;

[0064] (2) After completing the configuration of the frequency hopping spread spectrum radio frequency module, the node discovery frame is periodically broadcast to the surrounding intercom nodes. The node discovery frame carries the device identifier, network identifier, frequency hopping phase offset, remaining power, scene type code and transmission power level.

[0065] Specifically, the current intercom node reads its own device identifier and uses it as the basic input for generating the local frequency hopping sequence, mapping the device identifier to the participation parameters of the pseudo-random sequence generator. Simultaneously, the intercom group has already formed a shared network seed during the network handshake phase. This network seed is randomly generated by the initiating node during the initial network setup, with a value length of 32 bits, and is distributed to joining nodes via the node discovery frame. Therefore, upon receiving the network seed, the current intercom node no longer simply uses the local default initial value, but instead replaces the initial seed in the shift register with the network seed. This ensures that the frequency hopping sequence generated by the current intercom node is phase-aligned with the frequency hopping sequences of other intercom nodes within the same intercom group. This allows all intercom nodes to enter the same channel switching sequence using the same frequency hopping rhythm in the 2.4GHz band, achieving frame-level aligned reception and preventing unstable transmission and reception due to inconsistent frequency hopping phases even though nodes are operating within the same intercom group. The generation of the frequency hopping sequence is accomplished by a 15th-order linear feedback shift register, with the corresponding generator polynomial being x. 15 +x 14 +1. After the shift register completes the initial seed loading, it continuously outputs a pseudo-random frequency hopping sequence and writes the phase-aligned frequency hopping sequence into the frequency hopping spread spectrum RF module, thus completing the configuration of the frequency hopping spread spectrum RF module for the current intercom node. After the frequency hopping spread spectrum RF module configuration is complete, the current intercom node enters the node discovery broadcast state and continuously sends node discovery frames to surrounding intercom nodes at a fixed period, enabling surrounding nodes to promptly obtain the identity information, network affiliation, and current working status of the current intercom node. The node discovery frame adopts a periodic broadcast method with a broadcast period set to 50ms. The node discovery frame carries the device identifier, network identifier, frequency hopping phase offset, remaining battery power, scene type code, and transmit power level. The device identifier is 48 bits long, the network identifier is 16 bits long, the frequency hopping phase offset is 8 bits long, the remaining battery power is 8 bits long, the scene type code is 4 bits long, the transmit power level is 4 bits long, and a 16-bit cyclic redundancy check field is added. The total length of the entire frame is 18 bytes. The remaining battery level field indicates the current battery status of the intercom node, expressed as 1%. The transmit power level field indicates the current radio frequency transmit power level of the intercom node, with 16 levels in total, a maximum transmit power of 20dBm, and a minimum transmit power of 4dBm. The frequency hopping phase offset field reflects the phase position of the current intercom node relative to the group's common frequency hopping reference, enabling the receiving node to more accurately perform timing alignment and link quality measurement when a node discovery frame is received. Since the node discovery frame is continuously broadcast, surrounding intercom nodes can establish neighbor node perception results based on the received node discovery frame and use the node discovery frame as the raw input for constructing the neighbor link quality scoring table.

[0066] In one specific embodiment, the process of configuring the frequency hopping spread spectrum radio frequency module based on the device identifier of the current intercom node and using the network seed shared by all nodes in the intercom group as the initial seed of the shift register can specifically include the following steps:

[0067] (1) Based on the device identifier of the current intercom node, a frequency hopping sequence is generated through a shift register;

[0068] (2) Receive the node discovery frame broadcast by the intercom group on the fixed initial channel, and extract the network seed and frequency hopping phase offset;

[0069] (3) Replace the initial seed of the shift register with the networking seed of the intercom group initiating node broadcast, so that the phase of the frequency hopping sequence of all intercom nodes in the intercom group is aligned, and write the aligned frequency hopping sequence into the frequency hopping spread spectrum radio frequency module to complete the frequency hopping sequence configuration.

[0070] Specifically, the current intercom node reads its own device identifier and uses it as input for the pseudo-random sequence generation process. The device identifier is a 48-bit address derived value programmed at the factory, and the frequency hopping sequence is generated by a pseudo-random sequence generator. The pseudo-random sequence generator uses a 15th-order linear feedback shift register, with a corresponding generator polynomial of x. 15 +x 14+1. The device identifier is involved in the feedback link and state loading process of mapping to the shift register, so that the shift register continuously outputs a pseudo-random frequency hopping symbol sequence under clock drive. Then, according to the preset channel mapping relationship, the continuously output symbol sequence is transformed into the frequency hopping sequence of the current intercom node. After setting, the current intercom node can obtain the basic frequency hopping result associated with its own identity, thereby avoiding different nodes from using the same local frequency hopping starting point before unified networking, and ensuring that the frequency hopping control process has node identity distinguishing characteristics. After the initial frequency hopping sequence based on the device identifier is generated, the network seed replacement process is executed. All intercom nodes in the same intercom group share the same network seed during the network handshake phase. The network seed is randomly generated by the intercom group initiating node during the first networking, with a length of 32 bits, and is distributed in plaintext through the node discovery frame. After receiving the network seed, the joining node replaces the initial seed of its local pseudo-random sequence generator with the network seed. Therefore, upon receiving the network seed broadcast by the initiating node of the intercom group, the current intercom node no longer continues to use the initial register state formed solely by the local device identifier. Instead, it writes the network seed into the initial state register of the shift register and uses this network seed as the new unified starting state to re-drive the 15th-order linear feedback shift register to output the frequency hopping symbol sequence. Since all intercom nodes in the intercom group use the same network seed to replace the initial seed of their respective shift registers, the frequency hopping sequences generated by all intercom nodes can maintain consistency in the initial phase, thus achieving frequency hopping sequence phase alignment. The aligned frequency hopping sequence is written into the frequency hopping spread spectrum RF module to complete the frequency hopping sequence configuration. The frequency hopping sequence output in real time by the shift register is loaded into the frequency hopping control register, channel index register, and phase offset register, enabling the frequency hopping spread spectrum RF module to perform dynamic channel switching in the 2.4GHz band according to the aligned frequency hopping rhythm. After the writing process is completed, the frequency hopping spread spectrum RF module of the current intercom node switches from the static standby state to the initialized frequency hopping working state.

[0071] In one specific embodiment, such as Figure 3 The process of executing step 102 can specifically include the following steps:

[0072] 301. The received signal strength, signal-to-noise ratio and packet error rate carried in the node discovery frame are weighted and summed to obtain the comprehensive score of the neighbor link quality;

[0073] 302. Based on the comprehensive quality score of the neighbor links, select the primary path and the backup path among all nodes in the network, and construct a multi-hop routing table based on the primary path and the backup path;

[0074] 303. Determine the link mode flag of the current intercom node based on the comprehensive quality score of the neighbor links of the main path in the multi-hop routing table and the reference signal received power detected by the cellular network module.

[0075] Specifically, after receiving a node discovery frame from a neighboring node, the current intercom node records the neighboring node's presence status and simultaneously measures and records three link quality indicators during the same frequency hopping reception event. The received signal strength indicator (RSS) is measured independently on each frequency hopping channel by the RF front-end automatic gain control circuit, and the arithmetic mean of the RSS measurements from the most recent 20 consecutive frequency hopping events is used as the current RSS indicator. The signal-to-noise ratio (SNR) is calculated by the receiver's digital signal processing module as the ratio of the current demodulated signal power to the background noise floor, and the moving average of the most recent 20 measurements is used. The packet error rate (BER) is calculated as the proportion of frames with failed cyclic redundancy check (CRC) checks out of the total number of frames in the most recent 100 consecutive data frames. Based on these three indicators, the comprehensive neighbor link quality scoring formula is as follows: ,in, This represents the overall quality score of the neighbor link from the current intercom node A to its neighbor node B. This represents the received signal strength weighting coefficient, with a value of 0.4. This represents the signal-to-noise ratio weighting coefficient, with a value of 0.35. This represents the error packet rate weighting coefficient, with a value of 0.25. This represents the average received signal strength indicator (SSM) of the current intercom node A when it receives a discovery frame from neighbor node B, in dBm. This represents the lower limit reference value for received signal strength, with a value of -90dBm. This indicates the upper limit reference value for received signal strength, with a value of -30dBm. This represents the average signal-to-noise ratio (SNR) between the current intercom node A and its neighbor node B, expressed in dB. This indicates the upper limit reference value for the signal-to-noise ratio, with a value of 30dB. This represents the packet error rate from the current intercom node A to its neighbor node B, ranging from 0 to 1. The current intercom node writes the comprehensive neighbor link quality score of all valid neighbors into its local neighbor link quality score table in real time and refreshes the table every 100ms. For neighbor records that have not received a node discovery frame for more than 300ms, the corresponding record is marked as expired and removed from the set of valid neighbors. Based on the comprehensive neighbor link quality score, each node uses a distributed maximum bottleneck path algorithm to calculate the optimal multi-hop path to other nodes in the entire network. The path quality evaluation criterion is... ,in, Representing a path Path quality value, Representing a path Any hop link, This represents the overall quality score of the neighboring links of the corresponding link. This criterion indicates that the quality of a path is not determined by the simple summation or multiplication of the scores of each hop, but rather by the overall score of the weakest hop in the entire path. Therefore, during route discovery, nodes broadcast route discovery frames to the target node. These frames carry the source node's device identifier, the target node's device identifier, the current path's minimum overall score, a list of passed node identifiers, and a hop count. Each intermediate node, when forwarding, updates the current path's minimum overall score to the smaller of the "recorded minimum value" and the "current hop link overall score," and appends its device identifier to the list of passed node identifiers. It then continues broadcasting and forwarding as long as the hop count is less than 8. After collecting all arriving candidate paths, the target node selects the path with the largest current path minimum overall score as the primary path and the path with the second largest current path minimum overall score that does not share a common intermediate node with the primary path as the backup path. The primary path, backup path, and their corresponding path quality values ​​are then written into the multi-hop routing table. The current intercom node calculates the average comprehensive score based on the path quality value corresponding to the main path in the multi-hop routing table. The calculation method for the average Mesh coverage quality score of the entire group is as follows: ,in, This represents the average overall score of the current intercom group. This indicates the total number of target nodes within the current intercom group. This indicates that the current intercom node is connected to the target node. The main path quality value is read every 100ms, and the average comprehensive score is calculated. Simultaneously, the cellular network module queries the reference signal received power of the currently camped cell every 500ms. When the average comprehensive score is lower than the Mesh link handover threshold and this state persists for more than 200ms, the current radio mesh coverage quality is determined to be below the minimum intercom requirement; the Mesh link handover threshold is 0.25. When the reference signal received power detected by the cellular network module is not lower than the cellular network available power threshold, the cellular network is determined to be available; the cellular network available power threshold is -110dBm. Based on the above dual-condition comparison relationship, when the average comprehensive score is not lower than 0.25, the link mode flag is set to radio mesh mode; when the average comprehensive score is lower than 0.25 and the reference signal received power is not lower than -110dBm, the link mode flag is set to cellular network mode; when the average comprehensive score is not lower than 0.25 and the reference signal received power is also not lower than -110dBm, the link mode flag is set to hybrid mode. The link mode flag is further represented as Mesh_Mode=0x01, 5G_Mode=0x02, and Hybrid_Mode=0x03.

[0076] In one specific embodiment, it further includes:

[0077] (1) Send link keep-alive probe frames to each hop node of the main path;

[0078] (2) Continuously count the link keep-alive response frames returned by each hop node. When the number of consecutive failures to receive a link keep-alive response frame from the same hop node reaches a preset number, switch from the main path to the corresponding backup path in the multi-hop routing table.

[0079] Specifically, the current intercom node reads the main path from the multi-hop routing table and identifies the relationships between adjacent nodes on the main path hop by hop. Then, it sends link keep-alive probe frames to each hop node on the main path at a fixed period. That is, each node sends a link keep-alive probe frame to each hop link on the current main path at a period of 10ms. The link keep-alive probe frame is 6 bytes long, consisting of a 2-byte header, a 2-byte target device identifier, and a 2-byte sequence number, and is transmitted via a frequency-hopping spread spectrum channel. Upon receiving the link keep-alive probe frame, the peer node immediately returns a link keep-alive response frame. The link keep-alive response frame is 4 bytes long, consisting of a 2-byte header and a 2-byte echo of the corresponding sequence number. The sending node can then determine whether the specified link keep-alive probe frame has received a corresponding link keep-alive response frame based on the consistency between the sent sequence number and the echoed sequence number. This transforms each hop link on the main path from being considered usable only based on historical routing results to a dynamic state of being continuously verified as usable during actual forwarding. During continuous counting, an independent unanswered counter is established for each hop node. After each link keep-alive probe frame is sent, the current intercom node must initiate a corresponding waiting window for the target hop node. When a link keep-alive response frame with the same sequence number is received within the current period, the unanswered counter for that target hop node is cleared, and the main path remains unchanged. If no link keep-alive response frame with the corresponding sequence number is received within the current period, the unanswered counter for that target hop node is incremented. If a node does not receive a link keep-alive response frame from the same neighbor node within three consecutive link keep-alive probe periods, the link from the current node to that neighbor node is considered interrupted. Since the link keep-alive probe period is 10ms, the link failure detection window for three consecutive unanswered attempts is 30ms. When the number of consecutive failures to receive a link keep-alive response frame from the same hop node reaches a preset number (3 times), the current intercom node immediately triggers a route switching process. It reads the backup path corresponding to the current primary path from the multi-hop routing table, replaces the next-hop address of subsequent audio frames from the faulty node of the original primary path to the first-hop node of the backup path, and simultaneously updates the data link layer forwarding table entries. This ensures that subsequent voice data is no longer forwarded along the original primary path but is instead transmitted via the backup path. The route switching action itself is a local memory-level operation, mainly involving reading the backup path from the multi-hop routing table, completing the next-hop address replacement, and refreshing the forwarding table entries. Its time consumption is no more than one frequency hopping cycle (approximately 0.625ms). Adding the processor instruction execution time for multi-hop routing table reading and address replacement (no more than 0.1ms), the total switching time is less than 5ms. This 5ms corresponds to the "switching execution time after fault determination," and is not equivalent to the fault detection window time; the fault detection window is still 30ms, consisting of three link keep-alive detection cycles.

[0080] In one specific embodiment, the process of determining the link mode identifier of the current intercom node based on the comprehensive quality score of the neighbor links of the main path in the multi-hop routing table and the reference signal received power detected by the cellular network module can specifically include the following steps:

[0081] (1) Calculate the average comprehensive score based on the comprehensive quality score of the neighbor links of the main path in the multi-hop routing table;

[0082] (2) Compare the average comprehensive score with the Mesh link switching threshold, and simultaneously compare the reference signal received power periodically detected by the cellular network module with the cellular network available power threshold;

[0083] (3) When the average comprehensive score is not lower than the Mesh link switching threshold and the reference signal received power is not lower than the cellular network available power threshold, the link mode flag is set to hybrid mode; when the average comprehensive score is not lower than the Mesh link switching threshold and the reference signal received power is lower than the cellular network available power threshold, the link mode flag is set to radio Mesh mode; when the average comprehensive score is lower than the Mesh link switching threshold and the reference signal received power is not lower than the cellular network available power threshold, the link mode flag is set to cellular network mode; when the average comprehensive score is lower than the Mesh link switching threshold and the reference signal received power is lower than the cellular network available power threshold, the link mode flag of the previous cycle remains unchanged and enters the waiting for re-evaluation state.

[0084] Specifically, the current intercom node reads the main path records for each target node from its locally maintained multi-hop routing table, extracts the comprehensive quality score of neighbor links or path quality value corresponding to each main path, and then calculates the average of the corresponding comprehensive scores. The average mesh coverage quality score for the entire group is the average of the main path quality values ​​from the current intercom node to each target node, which is taken as the average of the comprehensive scores of the current intercom node. Its expression is ,in, This represents the average comprehensive score calculated by the current intercom node based on the main path of the multi-hop routing table. This indicates the total number of target nodes within the current intercom group. This indicates that the current intercom node is connected to the target node. The main path quality value. The average comprehensive score reflects the mesh coverage quality level of the current intercom node within the entire intercom group. The average comprehensive score is compared with the mesh link handover threshold, and simultaneously the reference signal received power periodically detected by the cellular network module is compared with the available cellular network power threshold. The mesh quality monitoring process is executed continuously at a period of 100ms, while the cellular network signal monitoring process queries the reference signal received power of the currently camped cell at a period of 500ms. The reference signal received power uses... The value is expressed in dBm; where the Mesh link handover threshold is... Take 0.25, the reference signal received power can be taken as a threshold. .when Furthermore, if this condition persists for more than 200ms, it is determined that the current Mesh coverage quality is below the minimum requirements for intercom calls; when At that time, the cellular network is determined to be in an available state. When the average comprehensive score is... Not lower than the Mesh link switching threshold When the current intercom node establishes a radio mesh main path through a multi-hop routing table, it indicates that the local intercom quality requirements can still be met. Therefore, the link mode flag is set to radio mesh mode. When the average comprehensive score is... Below the Mesh link switching threshold And the reference signal received power detected by the cellular network module Not lower than the available power threshold of the cellular network When the current radio mesh coverage quality is insufficient, but the cellular network has stable carrying capacity, the link mode flag is set to cellular network mode; when the average comprehensive score is... Not lower than the Mesh link switching threshold And the reference signal received power When the power level is not lower than the available cellular network power threshold, it indicates that the current intercom node maintains both a valid radio mesh connection and a valid cellular network connection. In this case, the link mode flag is set to hybrid mode, and the gateway node undertakes the audio bridging task between the local mesh subnet and the remote cellular subnet. Based on the defined values ​​of the link mode flag (Mesh_Mode=0x01, 5G_Mode=0x02, Hybrid_Mode=0x03), the above determination results are directly written to the local link state register.

[0085] In one specific embodiment, such as Figure 4 The process of executing step 103 can specifically include the following steps:

[0086] 401. Based on the scene type code, the first voice signal collected by the microphone is filtered and noise-reduced to obtain the second voice signal;

[0087] 402. Perform voice activity detection on the second voice signal according to the voice activity detection threshold corresponding to the scene type code and extract the voice frames that pass the voice activity detection;

[0088] 403. Encode and AES-encrypt the voice frames detected by voice activity to obtain encrypted audio frames;

[0089] 404. Transmit the encrypted audio frame according to the radio mesh frequency hopping spread spectrum channel or cellular network channel indicated by the link mode identifier.

[0090] Specifically, the microphone acquires the first voice signal, which enters the general audio pre-processing chain. The microphone analog signal first undergoes 26dB gain adjustment by an operational amplifier, then is input to an analog-to-digital converter (ADC). The ADC has a sampling rate of 16kHz, a quantization accuracy of 16bit, and a signal-to-noise ratio (SNR) of no less than 90dB. The pulse-code modulation (PCM) data stream output by the ADC enters the digital signal processing chain and undergoes high-pass filtering by a hardware digital filtering circuit. The high-pass filter has a cutoff frequency of 150Hz, uses a second-order Butterworth structure, and has an attenuation rate of 12dB / octave. This is used to filter out low-frequency harmonics from the engine, low-frequency wind noise components, and low-frequency noise introduced by microphone installation vibration. The current intercom node performs scene-specific filtering and noise reduction processing on the first voice signal according to the scene type code to obtain the second voice signal. The second stage of the digital signal processing chain is a scene-specific deep neural network noise suppression module. This module takes a 257-dimensional short-time Fourier transform amplitude spectrum as input, passes it through three fully connected hidden layers, and outputs a 257-dimensional ideal ratio masking value. The ideal ratio masking value is then multiplied element-wise with the noisy amplitude spectrum, and combined with the original phase spectrum, an inverse short-time Fourier transform is performed to reconstruct the net speech time-domain waveform, thus obtaining the second speech signal. Regarding the corresponding parameters, the frame length is 20ms, the fast Fourier transform number is 256 points, and the inference time per frame is approximately 6.5ms and no more than 7ms, thus meeting the requirements for real-time intercom processing. For referee enforcement scenarios, a dedicated processing procedure is further defined. Before obtaining the second voice signal, the first voice signal undergoes high-pass filtering and notch filtering of the whistle characteristic frequency band sequentially. The notch filtering of the whistle characteristic frequency band is performed by multiple cascaded second-order infinite impulse response notch filters. The center frequency of each notch filter corresponds to multiple peak frequency points within the fundamental frequency range of the referee's whistle. Three cascaded second-order notch filters are used, with center frequencies set to 2500Hz, 3000Hz, and 3500Hz respectively, a quality factor of 15 for each, and a notch depth of no less than 40dB. This first suppresses the whistle energy, and then the remaining background noise is removed by a deep neural network noise suppression module. Selective extraction is performed according to the voice activity detection threshold corresponding to the scene type code. The second voice signal is input into the voice activity detector to obtain the voice activity probability value of the current audio frame. The audio frame length corresponding to the voice activity detector is 20ms, outputting the net voice probability, with an inference time of less than 1ms. The current intercom node reads the corresponding threshold from the pre-stored scene voice activity detection threshold table according to the scene type code. The threshold values ​​for motorcycle riding scenarios and referee enforcement scenarios are higher than those for skiing or surfing scenarios. Specifically, the threshold value for motorcycle riding scenarios is 0.85, for skiing scenarios it is 0.80, for surfing scenarios it is 0.88, for referee enforcement scenarios it is 0.85, and for coach training scenarios and industrial obstacle scenarios it is also 0.85.When the probability value of voice activity in the current audio frame is not lower than the corresponding threshold, the frame is extracted as a voice frame; when the probability value of voice activity is continuously lower than the corresponding threshold and the duration exceeds 500ms, it is determined to enter the voice silence state, and the voice frame extraction is stopped, while the transmission state of the frequency hopping spread spectrum RF module is turned off. The net voice pulse code modulation frame after scene-based audio processing is input to the encoder according to the frame structure of 16kHz, 16bit, 20ms. OPUS encoding is performed first, with encoding parameters of 16kbps, 20ms frame, and mono; then AES-128-CTR symmetric encryption is performed. The encryption key is derived from the network seed and device identifier through the HMAC-SHA256 algorithm. The derivation relationship is key = HMAC-SHA256(network seed || device identifier)[0:16], thus obtaining a 128-bit encryption key. The network seed is renegotiation each time the intercom group re-networks to ensure key isolation between different intercom groups. After encryption, a frame header is added, including a 16-bit network identifier, a 16-bit sequence number, a 4-bit priority flag, and a 4-bit reserved bit. The total length of the frame header is 5 bytes, thus forming an encrypted audio frame. For referee enforcement scenarios, a broadcast priority flag is added to the encrypted audio frame generated for this scenario, ensuring that all intercom nodes in the intercom group unconditionally decrypt and play the encrypted audio frame carrying the broadcast priority flag upon receipt, enabling unlimited broadcasting of referee instructions. The encrypted audio frame is transmitted according to the radio mesh frequency hopping spread spectrum channel or cellular network channel indicated by the link mode flag. When the link mode flag is radio mesh mode, the current intercom node writes the encrypted audio frame into the frequency hopping spread spectrum radio frequency transmit buffer and transmits it on the corresponding channel according to the current frequency hopping phase. The transmission power is configured according to preset scenario parameters, especially in industrial obstacle scenarios and skiing scenarios where the battery cell temperature is below a certain level. At 5℃, the transmit power is set to the maximum level of 20dBm. When the link mode flag is set to cellular network mode, encrypted audio frames are no longer transmitted through the frequency-hopping spread spectrum RF module. Instead, they are encapsulated by DTLS / AES-256-GCM and transmitted to the cloud relay via the UDP / 5G channel. The frequency-hopping spread spectrum RF module is also turned off to save power. When the link mode flag is set to mixed mode, the gateway node simultaneously performs two parallel transmission processes: frequency-hopping spread spectrum RF transmission and 5G data upload. The digital signal processor writes to both transmit buffers in parallel through direct memory access, enabling bidirectional audio bridging between the radio mesh local area network and remote members of the cellular network.

[0091] The step of detecting voice activity in the second voice signal according to the voice activity detection threshold corresponding to the scene type code and extracting the voice frames that pass the voice activity detection includes: inputting the second voice signal into a voice activity detector to obtain the voice activity probability value of the current audio frame; reading the corresponding voice activity detection threshold from a pre-stored scene voice activity detection threshold table according to the scene type code, wherein the voice activity detection threshold corresponding to each scene in the scene voice activity detection threshold table is determined based on the optimal operating point of the ROC curve formed by the false activation rate and the missed activation rate of the voice activity detector on the scene noise test set, wherein the voice activity detection threshold corresponding to the motorcycle riding scene and the referee enforcement scene is higher than the voice activity detection threshold corresponding to the skiing or surfing scene; comparing the voice activity probability value with the voice activity detection threshold, extracting the frame as a voice frame when the voice activity probability value is not lower than the voice activity detection threshold, and stopping the voice frame extraction and turning off the transmission state of the frequency hopping spread spectrum radio frequency module when the voice activity probability value is lower than the voice activity detection threshold and the duration exceeds the preset silence judgment duration.

[0092] In one specific embodiment, the process of transmitting the encrypted audio frame according to the radio mesh frequency hopping spread spectrum channel or cellular network channel pointed to by the link mode flag may specifically include the following steps:

[0093] (1) When the link mode flag is radio mesh mode, the encrypted audio frame is written into the transmit buffer of the frequency hopping spread spectrum radio frequency module for directional unicast forwarding until the encrypted audio frame reaches all target nodes in the intercom group;

[0094] (2) When the link mode identifier is cellular network mode, the encrypted audio frame is uploaded to the cloud relay server, and the cloud relay server broadcasts the encrypted audio frame to all registered nodes in the same intercom group room;

[0095] (3) When the link mode flag is a mixed mode, the gateway node that simultaneously holds a valid radio mesh connection and a valid cellular network connection performs deduplication and backhaul suppression on the audio stream received from the frequency hopping spread spectrum channel and the audio stream sent from the cloud relay server, and then performs superposition to obtain a mixed audio stream. The mixed audio stream is then re-uploaded to the cloud relay server for distribution to the cellular network side node.

[0096] Specifically, the encrypted audio frame transmission module is designed as a branch execution structure directly driven by the link mode flag. After voice frame encoding, AES encryption, and frame header encapsulation, the current intercom node first reads the link mode flag and then selects either a radio mesh directional unicast forwarding path, a cellular network cloud relay broadcast path, or a local mesh and cellular network parallel bridging path. When the link mode flag is in radio mesh mode, the current intercom node reads the main path entry corresponding to the target node set from the multi-hop routing table and writes the encrypted audio frame into the transmit buffer of the frequency-hopping spread spectrum RF module according to the next-hop address in the main path. In this mode, the transmission relies on the already converged main path to perform hop-by-hop directional unicast forwarding. Each hop node, after receiving the encrypted audio frame, forwards it only to the next-hop node in the current path according to its local multi-hop routing table, until the encrypted audio frame reaches the corresponding target node along the main path. When multiple target nodes exist within the same intercom group, the current intercom node can query the main path according to the target node dimension and perform targeted unicast to the forwarding branches corresponding to different target nodes one by one. This restricts the encrypted audio frame to propagate within the effective main path range, avoiding indiscriminate diffusion that occupies frequency-hopping spread spectrum channel resources and maintaining channel utilization efficiency under large-scale networking conditions. For audio frames with additional broadcast priority tags, the routing layer increases the priority forwarding order, allowing nodes within the group to complete playback first when the reception conditions are met. When the link mode flag is cellular network mode, the current intercom node no longer continues to send the encrypted audio frame along the local multi-hop path through the frequency-hopping spread spectrum radio frequency module, but switches to the cellular network upload link. When radio mesh coverage is insufficient but cellular network availability is available, a node first initiates a group registration request to the cloud relay server. The cloud relay server assigns a unique intercom room identifier to the current network identifier and distributes session keys to registered nodes within the group, thereby establishing an encrypted real-time audio transmission channel based on UDP / DTLS. After the channel is established, the current intercom node uploads the encoded audio data to the cloud relay server after DTLS encryption and encapsulation. The cloud relay server then broadcasts this data to all online and registered nodes within the same intercom room. With this setup, all members on the cellular network side do not need to be aware of the details of the local radio mesh link; they only need to maintain registration and session continuity with the cloud relay server to receive the encrypted audio frames uploaded by the current intercom node, thus achieving intra-group voice distribution within the long-distance cellular network coverage area. When the link mode flag is in mixed mode, a subset of nodes with effective local radio mesh coverage and a subset of remote cellular network nodes already exist in the network. Therefore, the gateway node with dual-sided connectivity assumes the bridging responsibility.The gateway node receives audio streams uploaded from the radio mesh subnet on the frequency-hopping spread spectrum channel side and audio streams sent from the cloud relay server on the cellular network side. It performs consistency comparison on the source identifiers, sequence numbers, or timestamps of the two audio streams, performs deduplication processing on audio frames with the same source identifier or duplicate timestamps, and simultaneously performs backfeed suppression on the audio transmitted back to its local end. Then, in the local audio processing module, it performs superposition processing on the two retained audio streams to obtain a mixed audio stream. After the mixed audio stream is generated, the gateway node re-uploads it to the cloud relay server, which then distributes it to the cellular network side nodes. Simultaneously, the gateway node can also write the mixed audio stream into the transmit buffer of the frequency-hopping spread spectrum radio frequency module and forward it along the main path of the radio mesh to the target nodes within the local radio mesh subnet, thereby achieving bidirectional voice communication between the radio mesh side nodes and the cellular network side nodes. To reduce processing latency during the parallel forwarding process of dual links, the gateway node uses the direct memory access mechanism of the digital signal processor to write the audio data to be transmitted to the radio mesh transmit buffer and the cellular network transmit buffer in parallel, so that the mixed audio stream enters the forwarding process synchronously on the local radio mesh link and the cellular network link.

[0097] When the scene type code is a referee enforcement scene, the step of filtering and noise reduction processing on the first voice signal collected by the microphone according to the scene type code to obtain the second voice signal includes: performing high-pass filtering and whistle feature frequency band notch filtering processing on the first voice signal in sequence. The whistle feature frequency band notch filtering processing is performed by multiple cascaded second-order infinite impulse response notch filters. The center frequency of each notch filter corresponds to multiple peak frequency points in the fundamental frequency range of the motion referee's whistle, to obtain a whistle-suppressed voice signal; inputting the whistle-suppressed voice signal into a deep neural network noise suppression module, performing ideal ratio masking prediction on the short-time Fourier transform amplitude spectrum of each frame of voice signal, multiplying the ideal ratio masking element-wise with the noisy amplitude spectrum, and reconstructing it through inverse short-time Fourier transform to obtain the second voice signal; adding a broadcast priority mark to the encrypted audio frame generated when the scene type code is a referee enforcement scene. All intercom nodes in the intercom group unconditionally decrypt and play the encrypted audio frame carrying the broadcast priority mark when they receive it, without being limited by the upper limit of the number of people talking at the same time.

[0098] When there are walkie-talkies of different brands connected to the current walkie-talkie node via Bluetooth hands-free protocol within the walkie-talkie group, the process further includes: performing sampling rate conversion on the audio stream output by the walkie-talkie of different brands via the Bluetooth hands-free protocol channel, so that the converted audio sampling rate is consistent with the sampling rate of the radio mesh network audio processing flow, to obtain a resampled audio stream; processing the resampled audio stream according to the steps of filtering and noise reduction, encoding and AES encryption packaging to obtain a cross-brand encrypted audio frame; writing the cross-brand encrypted audio frame into the transmit buffer of the frequency hopping spread spectrum radio frequency module, and performing directional unicast forwarding according to the hop-by-hop node device identifier recorded in the main path of the multi-hop routing table, until the cross-brand encrypted audio frame reaches all target nodes within the walkie-talkie group; performing AES decryption and decoding on the encrypted audio frames received by the current walkie-talkie node from other walkie-talkie nodes within the walkie-talkie group via the frequency hopping spread spectrum channel to obtain a decoded audio stream, and transmitting the decoded audio stream back to the walkie-talkie of different brands via the Bluetooth hands-free protocol channel to complete the two-way audio bridging.

[0099] The multi-scenario adaptive intercom communication method based on radio mesh in the embodiments of the present invention has been described above. The multi-scenario adaptive intercom communication system based on radio mesh in the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 5 One embodiment of the multi-scenario adaptive intercom communication system based on radio mesh in this invention includes:

[0100] The scene recognition module 501 is used to identify the scene type code of the current intercom node and broadcast the node discovery frame to the surrounding intercom nodes.

[0101] The mode detection module 502 is used to calculate the comprehensive score of neighbor link quality based on the node discovery frame and determine the link mode flag of the current intercom node with the reference signal received power detected by the cellular network module.

[0102] The audio transmission module 503 is used to reduce the noise of the first voice signal collected by the microphone according to the scene type code to obtain a second voice signal, encrypt the second voice signal to generate an encrypted audio frame, and transmit the encrypted audio frame according to the radio mesh frequency hopping spread spectrum channel or cellular network channel pointed to by the link mode flag.

[0103] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0104] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0105] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multi-scenario adaptive intercom communication method based on radio mesh, characterized in that, include: Identify the scene type code of the current intercom node and broadcast a node discovery frame to surrounding intercom nodes; The link mode flag of the current intercom node is determined by calculating the comprehensive score of the neighbor link quality based on the node discovery frame and by comparing it with the reference signal received power detected by the cellular network module. The first voice signal collected by the microphone is denoised according to the scene type code to obtain the second voice signal, and the second voice signal is encrypted to generate an encrypted audio frame. When the link mode identifier is in radio mesh mode, the encrypted audio frame is written into the transmit buffer of the frequency hopping spread spectrum radio module for directional unicast forwarding until the encrypted audio frame reaches all target nodes in the intercom group; when the link mode identifier is in cellular network mode, the encrypted audio frame is uploaded to the cloud relay server, which then broadcasts the encrypted audio frame to all registered nodes in the same intercom group room; when the link mode identifier is in mixed mode, the gateway node, which simultaneously holds a valid radio mesh connection and a valid cellular network connection, performs deduplication and backhaul suppression on the audio stream received from the frequency hopping spread spectrum channel and the audio stream sent from the cloud relay server, and then performs superposition to obtain a mixed audio stream, which is then re-uploaded to the cloud relay server for distribution to the cellular network side nodes.

2. The multi-scenario adaptive intercom communication method based on radio mesh according to claim 1, characterized in that, The step of identifying the scene type code of the current intercom node and broadcasting the node discovery frame to surrounding intercom nodes includes: The background noise signal of the microphone in the silent zone is collected, and the low-frequency energy integral value, mid-frequency energy integral value and whistle characteristic frequency band energy integral value of the background noise signal are calculated. When the low-frequency energy integral value exceeds the preset energy detection threshold and is greater than the mid-frequency energy integral value, the output scene type code is "motorcycle riding scene"; when the mid-frequency energy integral value exceeds the preset energy detection threshold and is greater than both the low-frequency energy integral value and the whistle characteristic frequency band energy integral value, the output scene type code is "skiing or surfing scene"; when the whistle characteristic frequency band energy integral value exceeds the preset energy detection threshold and is greater than the low-frequency energy integral value, the output scene type code is "referee enforcement scene". Configure the frequency hopping spread spectrum radio frequency module corresponding to the current intercom node based on the scenario type code, and periodically broadcast node discovery frames to surrounding intercom nodes.

3. The multi-scenario adaptive intercom communication method based on radio mesh according to claim 2, characterized in that, The step of configuring the frequency hopping spread spectrum radio frequency module corresponding to the current intercom node based on the scenario type code, and periodically broadcasting node discovery frames to surrounding intercom nodes, includes: Based on the device identifier of the current intercom node, the network seed shared by each node in the intercom group is used as the initial seed of the shift register to complete the configuration of the frequency hopping spread spectrum radio frequency module. After completing the configuration of the frequency hopping spread spectrum radio frequency module, a node discovery frame is periodically broadcast to surrounding intercom nodes. The node discovery frame carries the device identifier, network identifier, frequency hopping phase offset, remaining battery power, scene type code, and transmission power level.

4. The multi-scenario adaptive intercom communication method based on radio mesh according to claim 3, characterized in that, The step of configuring the frequency hopping spread spectrum radio frequency module based on the device identifier of the current intercom node, using the network seed shared by all nodes in the intercom group as the initial seed of the shift register, includes: Based on the device identifier of the current intercom node, a frequency hopping sequence is generated through a shift register; The network seed of the intercom group initiating node broadcast is replaced with the initial seed of the shift register, so that the phase of the frequency hopping sequence of all intercom nodes in the intercom group is aligned, and the aligned frequency hopping sequence is written into the frequency hopping spread spectrum radio frequency module to complete the frequency hopping sequence configuration.

5. The multi-scenario adaptive intercom communication method based on radio mesh according to claim 4, characterized in that, The step of calculating the comprehensive neighbor link quality score based on the node discovery frame and determining the link mode flag of the current intercom node based on the reference signal received power detected by the cellular network module includes: The received signal strength, signal-to-noise ratio, and packet error rate carried in the node discovery frame are weighted and summed to obtain a comprehensive score of the neighbor link quality. Based on the comprehensive quality score of the neighbor links, a primary path and a backup path are selected among all nodes in the network, and a multi-hop routing table is constructed based on the primary path and the backup path. The link mode flag of the current intercom node is determined based on the comprehensive quality score of the neighbor links of the main path in the multi-hop routing table and the reference signal received power detected by the cellular network module.

6. The multi-scenario adaptive intercom communication method based on radio mesh according to claim 5, characterized in that, Also includes: Send link keep-alive probe frames to each hop node of the main path; The link keep-alive response frames returned by each hop node are continuously counted. When the number of consecutive failures to receive a link keep-alive response frame from the same hop node reaches a preset number, the path is switched from the primary path to the corresponding backup path in the multi-hop routing table.

7. The multi-scenario adaptive intercom communication method based on radio mesh according to claim 5, characterized in that, The step of determining the link mode flag of the current intercom node based on the comprehensive quality score of neighboring links of the primary path in the multi-hop routing table and the reference signal received power detected by the cellular network module includes: The average comprehensive score is calculated based on the comprehensive quality score of the neighbor links of the main path in the multi-hop routing table. The average comprehensive score is compared with the Mesh link switching threshold, and the reference signal received power periodically detected by the cellular network module is compared with the available power threshold of the cellular network. When the average comprehensive score is not lower than the Mesh link switching threshold and the reference signal received power is not lower than the cellular network available power threshold, the link mode flag is set to hybrid mode; when the average comprehensive score is not lower than the Mesh link switching threshold and the reference signal received power is lower than the cellular network available power threshold, the link mode flag is set to radio Mesh mode; when the average comprehensive score is lower than the Mesh link switching threshold and the reference signal received power is not lower than the cellular network available power threshold, the link mode flag is set to cellular network mode; when the average comprehensive score is lower than the Mesh link switching threshold and the reference signal received power is lower than the cellular network available power threshold, the link mode flag remains unchanged from the previous cycle and enters a waiting-for-re-evaluation state.

8. The multi-scenario adaptive intercom communication method based on radio mesh according to claim 7, characterized in that, The step of denoising the first speech signal collected by the microphone according to the scene type code to obtain a second speech signal, encrypting the second speech signal, and generating an encrypted audio frame includes: Based on the scene type code, the first voice signal collected by the microphone is filtered and noise-reduced to obtain the second voice signal; The second voice signal is subjected to voice activity detection according to the voice activity detection threshold corresponding to the scene type code, and the voice frames that pass the voice activity detection are extracted. The voice frames detected by voice activity are encoded and AES encrypted and packaged to obtain encrypted audio frames.

9. A multi-scenario adaptive intercom communication system based on radio mesh, characterized in that, The method for performing the multi-scenario adaptive intercom communication method based on radio mesh as described in any one of claims 1-8 includes: The scene recognition module is used to identify the scene type code of the current intercom node and broadcast the node discovery frame to surrounding intercom nodes; The mode detection module is used to calculate the comprehensive score of neighbor link quality based on the node discovery frame and determine the link mode flag of the current intercom node by comparing it with the reference signal received power detected by the cellular network module. An audio transmission module is used to reduce noise in the first voice signal collected by the microphone according to the scene type code to obtain a second voice signal, encrypt the second voice signal, and generate an encrypted audio frame. When the link mode identifier is radio mesh mode, the encrypted audio frame is written into the transmit buffer of the frequency hopping spread spectrum radio frequency module for directional unicast forwarding until the encrypted audio frame reaches all target nodes in the intercom group. When the link mode identifier is cellular network mode, the encrypted audio frame is uploaded to the cloud relay server, and the cloud relay server broadcasts the encrypted audio frame to all registered nodes in the same intercom group room. When the link mode identifier is mixed mode, the gateway node, which simultaneously holds a valid radio mesh connection and a valid cellular network connection, performs deduplication and backhaul suppression on the audio stream received from the frequency hopping spread spectrum channel and the audio stream sent from the cloud relay server, and then performs superposition to obtain a mixed audio stream, and re-uploads the mixed audio stream to the cloud relay server for distribution to the cellular network side nodes.

Images