A console voice interaction method based on artificial intelligence

Abstract

The application discloses a console voice interaction method based on artificial intelligence and relates to the technical field of computers.The application unifies heterogeneous API protocols into a computable topological structure, combines the dynamic recombination capability of a neural protocol compiler, eliminates protocol faults between the tree-shaped verification of a public security video platform and the linear execution of a fire-fighting system, fills parameters in a depth-first manner and uses a queue type delivery mechanism to ensure the parameter integrity of complex instructions and avoid instruction analysis failure caused by interface differences, automatically injects default values into missing parameters and records abnormal logs to further provide fault tolerance protection, adapt to the actual combat environment with frequent interface upgrades, crack instruction blocking caused by resource competition, dynamically bind network bandwidth, GPU computing power, authorized tokens and other key competition dimensions in a resource coupling relationship graph, accurately identify the concurrent conflict between video stream transmission and voice broadcast, and use a conflict probability prediction model based on a resource competition function to quantitatively evaluate the risk level and provide a decision basis for a dynamic programming algorithm.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and in particular to a console voice interaction method based on artificial intelligence. Background Technology

[0002] During emergency response, the command center needs to simultaneously call upon various heterogeneous systems such as public security monitoring, fire communication, and medical dispatch. Commanders use voice commands to achieve rapid cross-platform responses, such as "retrieving monitoring data for a certain road section and notifying nearby fire trucks to provide support." Such scenarios require commands to be precisely broken down and delivered to different vendors' systems in parallel. However, protocol differences and resource competition between multiple systems lead to inherent conflicts in the command execution chain.

[0003] Current mainstream solutions employ dynamic routing protocol conversion and instruction priority queue mechanisms; they map voice instructions to standardized middleware instructions through a pre-built rule base before distributing them to the target platform; some solutions introduce asynchronous call models, allowing non-dependent instructions to execute in parallel; however, in actual cross-system collaboration, they still face core bottlenecks: First, tree-structured APIs, such as those for public security video platforms, require multi-layer parameter validation, while linear execution APIs, such as those for fire protection systems, suffer from protocol breaks in a single request, both of which can easily lead to parameter loss during transmission; second, there is a lack of global resource awareness when multiple instructions are concurrent, with operations such as video stream retrieval and voice broadcasting competing for bandwidth resources, resulting in delays or drops of critical instructions.

[0004] Some solutions abstract interface differences through a protocol adaptation layer and configure dedicated threads for high-priority commands; for example, when a "notify fire department" command is recognized, non-emergency video stream requests are forcibly suspended. However, this method relies on manually preset rules and cannot dynamically identify implicit resource coupling between commands. For example, video analysis requires GPU computing power, which indirectly blocks speech synthesis. At the same time, due to frequent upgrades of vendor interfaces, the protocol conversion process requires continuous maintenance of the mapping rule base, and there is still a risk of response gaps in the handling of complex events. Summary of the Invention

[0005] In view of the aforementioned existing problems, the present invention is proposed.

[0006] This invention provides an AI-based console voice interaction method to address three major bottlenecks in cross-platform command execution for emergency command: protocol fragmentation, resource contention, and difficulty in interface adaptation. Existing solutions rely on manual rules and cannot dynamically detect hidden conflicts.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0008] This invention provides an artificial intelligence-based console voice interaction method, which includes:

[0009] Step S1: Receive the voice command stream and extract the operation object, action type and associated parameters in the command through a cross-modal semantic parsing model;

[0010] Step S2: Construct a resource coupling relationship graph and access the device status data, network bandwidth, and computing resource utilization of each controlled platform in real time;

[0011] Step S3: Predict resource contention nodes during instruction execution based on the resource coupling relationship graph;

[0012] Step S4: When resource contention is detected, an instruction scheduling sequence is generated based on a dynamic programming algorithm;

[0013] Step S5: The scheduled instructions are converted into an executable instruction set for the target platform using a neural protocol compiler;

[0014] Step S6: Distribute the instruction set to the target platform according to the instruction scheduling sequence.

[0015] As a preferred embodiment of the artificial intelligence-based console voice interaction method described in this invention, step S2 includes:

[0016] The tree-like API protocol of the public security video platform is abstracted into a multi-layered directed graph with verification.

[0017] The linear API protocol of the fire communication system is abstracted into a single-chain execution path;

[0018] Dynamically bind network bandwidth fluctuation data with GPU computing unit load metrics.

[0019] As a preferred embodiment of the artificial intelligence-based console voice interaction method described in this invention, step S3 includes:

[0020] Identify the types of shared resources between concurrent instruction pairs;

[0021] When the shared resource is a network transmission channel, predict the peak conflict between video streaming and voice broadcasting;

[0022] Calculate the conflict probability based on the resource contention function.

[0023] As a preferred embodiment of the AI-based console voice interaction method described in this invention, step S3, the calculation of the resource contention probability, includes:

[0024] Establish a demand matrix and distribute the concurrent instruction flow (i,j) across various shared resources. The instantaneous demand on d is denoted as d. ij,r And take the current available upper limit B. r With real-time occupancy rate u r(t) serves as the basis for subsequent calculations;

[0025] Construct a competition intensity function, when the remaining resource (B) r -u r When (t) decreases, the competitive intensity increases rapidly, expressed as:

[0026]

[0027] Where, θ ij,r (t) represents the instantaneous contention intensity of command pair (i,j) on resource r at time t, d ij,r Let k be the instantaneous demand of the instruction pair (i,j) for resource r. r B is the buffer coefficient, dimensionless. r u is the total available bandwidth of resource r. r (t) represents the real-time occupancy rate of resource r at time t;

[0028] Correcting k using recent jitter variance r :

[0029] k r =1+γ·Var(u r (t-τ:t)),

[0030] Where γ is the empirical weight, Var(u r (t-τ:t)) represents the variance of occupancy rate within the time window [t-τ,t], where τ is the length of the statistical window in seconds.

[0031] Logarithmic odds are used to compress extreme values:

[0032]

[0033] Where σ(x) is the normalized value after mapping, and x is the input variable, which is then taken as θ. ij,r (t), where α is the slope adjustment factor and β is the translation factor;

[0034] The comprehensive competition index is obtained by weighted summation of the competition intensity for multiple resources:

[0035]

[0036] Among them, S ij (t) represents the overall competitiveness index of the instruction pair. For a collection of shared resources, w r The weight of resource r, σ(θ) ij,r (t) represents the normalized competition intensity, and (θ) represents the competition intensity after normalization. ij,r (t) represents the contention intensity of instruction pair (i,j) on resource r, and is dimensionless;

[0037] Assume that the race event follows a Poisson failure rate, then:

[0038] C ij (t)=1-exp(-S ij (t)),

[0039] Among them, C ij S(t) represents the resource conflict probability of concurrent instruction pairs at time t. ij (t) represents the comprehensive competitiveness index calculated in the previous step.

[0040] As a preferred embodiment of the AI-based console voice interaction method described in this invention, the shared resource types include:

[0041] Network transmission channel;

[0042] GPU computing unit;

[0043] Platform authorization token.

[0044] As a preferred embodiment of the AI-based console voice interaction method described in this invention, step S4 includes:

[0045] Recursively break down the parameter verification path of the tree-structured API command until the leaf node;

[0046] When resource contention is detected between linear API instructions and tree-structured API instructions, a dynamically adjusted delay window is automatically inserted.

[0047] The length of the delay window is adjusted according to the real-time network jitter rate.

[0048] As a preferred embodiment of the AI-based console voice interaction method described in this invention, the length of the delay time window in step S4 is adjusted according to the real-time network jitter rate, specifically including:

[0049] The round-trip delay of the most recent N data packets is l. n (t) Calculate the average normalized jump:

[0050]

[0051] Where ρ(t) represents the jitter rate, which is dimensionless, N is the number of statistical packets, and l n (t) represents the round-trip delay of the nth packet, and ε represents the minimum value for preventing zero.

[0052] The time window is scaled using an S-shaped saturation function based on the jitter rate.

[0053]

[0054] Among them, T w(t) represents the length of the current delay window to be inserted, T min T is the smallest insertable time window. max κ is the maximum time window, ρ0 is the steepness coefficient, and ρ0 is the inflection point jitter rate.

[0055] Introducing smoothing filtering:

[0056]

[0057] in, For the actual smoothing time window, λ is the smoothed time window value obtained from the previous scheduling cycle, λ is the historical weight, and Δt is the scheduling cycle interval in seconds.

[0058] As a preferred embodiment of the AI-based console voice interaction method described in this invention, the recursive parameter verification path includes:

[0059] When a leaf node fails to be validated, backtrack upwards to the nearest optional parameter branch;

[0060] Cut off redundant parameters whose depth exceeds the protocol-defined layer number and generate alarm logs.

[0061] As a preferred embodiment of the artificial intelligence-based console voice interaction method described in this invention, step S5 includes:

[0062] Map the semantically parsed action units to intermediate instruction templates;

[0063] For tree-structured API protocols, a depth-first traversal is used to fill in the parameters.

[0064] Queued parameter passing is used for linear API protocols;

[0065] Default values ​​are automatically injected when a required parameter is detected to be missing.

[0066] As a preferred embodiment of the AI-based console voice interaction method described in this invention, the depth-first traversal filling parameters include:

[0067] Fill in the required parameters of the lower-level components according to the protocol hierarchy;

[0068] Establish a skip flag for optional parameter branches;

[0069] Generate a parameter integrity verification summary and append it to the end of the instruction set.

[0070] The beneficial effects of this invention are as follows: This invention significantly improves the robustness of voice interaction in emergency command scenarios through a multi-level intelligent decision-making mechanism.

[0071] This invention achieves lossless collaboration of cross-platform commands by unifying heterogeneous API protocols into a computable topology and combining it with the dynamic reorganization capabilities of a neural protocol compiler to eliminate protocol gaps between the tree-like verification of public security video platforms and the linear execution of fire protection systems. Depth-first parameter filling and queue-based transmission mechanisms ensure the parameter integrity of complex commands and avoid command parsing failures caused by interface differences. Automatic injection of default values ​​for missing parameters and exception logging further provide fault tolerance and adapt to real-world environments with frequent interface upgrades.

[0072] This invention addresses instruction blocking caused by resource contention. It dynamically binds key contention dimensions such as network bandwidth, GPU computing power, and authorization tokens to a resource coupling graph, accurately identifying concurrent conflicts between video streaming and voice broadcasting. A conflict probability prediction model based on a resource contention function quantitatively assesses risk levels, providing a decision-making basis for dynamic programming algorithms. By recursively decomposing tree-like API instruction paths and inserting adaptive delay windows, it intelligently schedules instruction sequences when resources are overloaded, fundamentally avoiding instruction loss caused by bandwidth contention or computing power depletion.

[0073] This invention constructs an adaptive intelligent scheduling hub, which, unlike traditional static rule bases, achieves continuous optimization through a real-time feedback mechanism: network jitter rate drives the dynamic expansion and contraction of the delay window, historical occupancy variance corrects for resource contention intensity, and parameter integrity verification provides feedback on protocol conversion quality. This forms a closed-loop decision flow of monitoring-prediction-execution-verification, ensuring the system maintains high responsiveness and reliability even in complex and ever-changing emergency scenarios. Attached Figure Description

[0074] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0075] Figure 1 This is a flowchart illustrating the AI-based console voice interaction method in Example 1. Detailed Implementation

[0076] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0077] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0078] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0079] Example 1, referring to Figure 1 This embodiment provides an artificial intelligence-based console voice interaction method, including:

[0080] Step S1: Receive the voice command stream and extract the operation object, action type and associated parameters in the command through a cross-modal semantic parsing model;

[0081] Step S2: Construct a resource coupling relationship graph and access the device status data, network bandwidth, and computing resource utilization of each controlled platform in real time;

[0082] Step S2 includes:

[0083] The tree-like API protocol of the public security video platform is abstracted into a multi-layered directed graph with verification.

[0084] The linear API protocol of the fire communication system is abstracted into a single-chain execution path;

[0085] Dynamically link network bandwidth fluctuation data with GPU computing unit load metrics;

[0086] Step S3: Predict resource contention nodes during instruction execution based on the resource coupling relationship graph;

[0087] Step S3 includes:

[0088] Identify the types of shared resources between concurrent instruction pairs;

[0089] When the shared resource is a network transmission channel, predict the peak conflict between video streaming and voice broadcasting;

[0090] Calculate the conflict probability based on the resource contention function;

[0091] The calculation steps for the resource contention probability in step S3 include:

[0092] Establish a demand matrix, and denote the instantaneous demand of concurrent instruction streams (i,j) on each shared resource r∈R as d. ij,r And take the current available upper limit B. r With real-time occupancy rate u r (t) serves as the basis for subsequent calculations;

[0093] Construct a competition intensity function, when the remaining resource (B)r -u r When (t) decreases, the competitive intensity increases rapidly, expressed as:

[0094]

[0095] Where, θ ij,r (t) represents the instantaneous contention intensity of command pair (i,j) on resource r at time t, d ij,r Let k be the instantaneous demand (Mb, GFLOPS, etc.) of resource r for instruction pair (i,j). r B is the buffer coefficient, dimensionless. r u is the total available bandwidth of resource r. r (t) represents the real-time occupancy rate of resource r at time t;

[0096] Correcting k using recent jitter variance r :

[0097] k r =1+γ·Var(u r (t-τ:t)),

[0098] Where γ is the empirical weight, ranging from 0.5 to 1.2, Var(u r (t-τ:t)) represents the variance of occupancy rate within the time window [t-τ,t], where τ is the statistical window length in seconds. Var(·) calculates the variance of occupancy rate with a sampling period of 50ms and a sliding window length of τ.

[0099] Logarithmic odds are used to compress extreme values:

[0100]

[0101] Where σ(x) is the normalized value after mapping, and x is the input variable, which is then taken as θ. ij,r (t), where α is the slope adjustment factor, 1-3, and β is the translation factor, -1-1;

[0102] The comprehensive competition index is obtained by weighted summation of the competition intensity for multiple resources:

[0103]

[0104] Among them, S ij (t) represents the comprehensive competitiveness index of the instruction pair, R represents the set of shared resources, and w r The weight of resource r is 0-1, with values ​​closer to 1 indicating greater criticality, σ(θ) ij,r (t) represents the normalized competition intensity, and (θ) represents the competition intensity after normalization. ij,r (t) represents the contention intensity of instruction pair (i,j) on resource r, and is dimensionless;

[0105] Assume that the race event follows a Poisson failure rate, then:

[0106] C ij (t)=1-exp(-S ij (t)),

[0107] Among them, C ij S(t) represents the resource conflict probability of concurrent instruction pairs at time t. ij (t) represents the comprehensive competitiveness index calculated in the previous step;

[0108] Specifically, this step forms a closed loop from the demand level to the probability output: collecting the instantaneous demand of concurrent instructions across multiple resource dimensions, and then constructing the competition intensity based on the remaining resources, so that the more scarce the resources and the greater the demand, the higher the θ value. ij,r (t) approaches infinity; adaptive buffer coefficient k r As real-time jitter expands, the logarithmic odds function σ compresses the intensity to [0,1] and maintains monotonicity to suppress artificially high competition values ​​caused by resource peaks, ensuring that extreme inputs do not overwhelm the contributions of other resources; multi-resource weights w r The comprehensive index S is obtained by combining business priority and scarcity through normalization and superposition. ij (t) reflects both single resource bottlenecks and global congestion; Poisson probability mapping directly transforms the exponent into 0-1 probabilities, making it convenient to set thresholds at the scheduling layer to trigger reordering or delay, thereby achieving coordinated avoidance of network bandwidth and GPU load.

[0109] Shared resource types include:

[0110] Network transmission channel;

[0111] GPU computing unit;

[0112] Platform authorization token;

[0113] Step S4: When resource contention is detected, an instruction scheduling sequence is generated based on a dynamic programming algorithm;

[0114] Step S4 includes:

[0115] Recursively break down the parameter verification path of the tree-structured API command until the leaf node;

[0116] When resource contention is detected between linear API instructions and tree-structured API instructions, a dynamically adjusted delay window is automatically inserted.

[0117] The length of the delay window is adjusted based on the real-time network jitter rate;

[0118] The length of the delay window in step S4 is adjusted according to the real-time network jitter rate, specifically including:

[0119] The round-trip delay of the most recent N data packets is l. n (t) Calculate the average normalized jump:

[0120]

[0121] Where ρ(t) represents the jitter rate, which is dimensionless, N is the number of statistical packets, and l n (t) represents the round-trip time delay of the nth packet, and ε is the minimum value for preventing zero, ε≈10. -3 ms;

[0122] The time window is scaled using an S-shaped saturation function based on the jitter rate.

[0123]

[0124] Among them, T w (t) represents the length of the current delay window to be inserted, T min T is the smallest insertable time window. max The maximum time window is κ, the steepness coefficient is 2-6, and ρ0 is the inflection point jitter rate.

[0125] Introducing smoothing filtering:

[0126]

[0127] in, For the actual smoothing time window, λ is the smoothed time window value obtained from the previous scheduling cycle, λ is the historical weight, which takes the value of 0.6-0.9, and Δt is the scheduling cycle interval, in seconds.

[0128] Specifically, we first use normalized jumps to measure instantaneous jitter, excluding the proportion of fixed delays, to obtain ρ(t) which is directly related to network fluctuations; then, we map the jitter rate to [T] using a saturation-limited sigmoid function. min ,T max When ρ(t) is below the inflection point ρ0, the time window remains close to T. min To ensure low latency, after the jitter increases and crosses the inflection point, the exponential term decreases rapidly, and the window length approaches T. max Sufficient buffering is provided for subsequent instruction release; finally, exponential smoothing is used to suppress high-frequency fluctuations, and the parameter λ controls the depth of historical memory, so that the window length changes gradually, avoiding additional queuing delays caused by frequent plugging and unplugging, and improving the throughput and success rate of multi-instruction scheduling in unstable network scenarios.

[0129] The recursive decomposition parameter verification path includes:

[0130] When a leaf node fails to be validated, backtrack upwards to the nearest optional parameter branch;

[0131] Cut off redundant parameters whose depth exceeds the protocol-defined layer number and generate alarm logs;

[0132] Step S5: The scheduled instructions are converted into an executable instruction set for the target platform using a neural protocol compiler;

[0133] Step S5 includes:

[0134] Map the semantically parsed action units to intermediate instruction templates;

[0135] For tree-structured API protocols, a depth-first traversal is used to fill in the parameters.

[0136] Queued parameter passing is used for linear API protocols;

[0137] Automatically inject default values ​​when a required parameter is detected to be missing;

[0138] The depth-first traversal filling parameters include:

[0139] Fill in the required parameters of the lower-level components according to the protocol hierarchy;

[0140] Establish a skip flag for optional parameter branches;

[0141] Generate a parameter integrity verification digest and append it to the end of the instruction set;

[0142] Step S6: Distribute the instruction set to the target platform according to the instruction scheduling sequence;

[0143] Unless otherwise specified, γ, α, β, w are used in this article r The parameters such as λ and their value ranges (e.g., 0.5 to 1.2, 1 to 3, etc.) can be determined by those skilled in the art based on historical system data, simulation results, or online adaptive optimization. The listed values ​​are only preferred examples and do not constitute a limitation of the present invention.

[0144] It should be noted that the above 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 preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A voice interaction method for a console based on artificial intelligence, characterized in that, Includes the following steps: Step S1: Receive the voice command stream and extract the operation object, action type and associated parameters in the command through a cross-modal semantic parsing model; Step S2: Construct a resource coupling relationship graph and access the device status data, network bandwidth, and computing resource utilization of each controlled platform in real time. The resource coupling relationship graph is constructed by abstracting the tree-like API protocol of the public security video platform into a multi-layer verification directed graph, abstracting the linear API protocol of the fire communication system into a single-chain execution path, and dynamically binding network bandwidth fluctuation data with GPU computing unit load indicators. Step S3: Predict resource contention nodes during instruction execution based on the resource coupling relationship graph; Step S4: When resource contention is detected, an instruction scheduling sequence is generated based on a dynamic programming algorithm; Step S5: The scheduled instructions are converted into an executable instruction set for the target platform using a neural protocol compiler; Step S6: Distribute the instruction set to the target platform according to the instruction scheduling sequence; Step S3 includes: Identify the types of shared resources between concurrent instruction pairs; When the shared resource is a network transmission channel, predict the peak conflict between video streaming and voice broadcasting; Calculate the conflict probability based on the resource contention function; The calculation steps for the resource contention probability in step S3 include: Establish a demand matrix and combine concurrent instruction streams In various shared resources The instantaneous demand on is denoted as and take the current available upper limit. With real-time occupancy As a benchmark for subsequent calculations; Construct a competition intensity function when the remaining resources are... When the price decreases, the intensity of competition increases rapidly, as shown below: ， in, Indicates time t instruction pair In resources Instantaneous competition intensity For instruction pairs Resources Instantaneous demand The buffer coefficient is dimensionless. For resources Total available bandwidth, For a moment resource Real-time occupancy rate; Correction using recent jitter variance : ， in, For experience weight, For time windows Internal occupancy variance The statistical window length is expressed in seconds (s). Logarithmic odds are used to compress extreme values: ， in, The normalized value after mapping. For input variables, we will take them here later. , As the slope adjustment factor, The translation factor; The comprehensive competition index is obtained by weighted summation of the competition intensity for multiple resources: ， in, The instruction is for the overall competitiveness index. For a collection of shared resources, For resources The weight, The normalized competitive intensity, Representing resources The above command pair The intensity of competition is dimensionless; Assume that the race event follows a Poisson failure rate, then: ， in, For concurrent instructions at time The probability of resource conflicts, This is the comprehensive competitiveness index calculated in the previous step.

2. The AI-based console voice interaction method as described in claim 1, characterized in that, The types of shared resources include: Network transmission channel; GPU computing unit; Platform authorization token.

3. The AI-based console voice interaction method as described in claim 1, characterized in that, Step S4 includes: Recursively break down the parameter verification path of the tree-structured API command until the leaf node; When resource contention is detected between linear API instructions and tree-structured API instructions, a dynamically adjusted delay window is automatically inserted. The length of the delay window is adjusted according to the real-time network jitter rate.

4. The AI-based console voice interaction method as described in claim 3, characterized in that, The length of the delay window in step S4 is adjusted according to the real-time network jitter rate, specifically including: Recently Round-trip latency of one data packet Calculate the average normalized jump: ， in, Represents jitter rate, dimensionless. To count the number of packages, For the first Round-trip delay of each package, To prevent extremely small quantities; The time window is scaled using an S-shaped saturation function based on the jitter rate. ， in, The current delay window length to be inserted. The smallest insertable time window. For the maximum time window, This is the steepness coefficient. The inflection point jitter rate; Introducing smoothing filtering: ， in, For the actual smoothing time window, The smoothed time window value obtained from the previous scheduling cycle. For historical weighting, The scheduling cycle interval is measured in seconds (s).

5. The AI-based console voice interaction method as described in claim 3, characterized in that, The recursive decomposition parameter verification path includes: When a leaf node fails to be validated, backtrack upwards to the nearest optional parameter branch; Cut off redundant parameters whose depth exceeds the protocol-defined layer number and generate alarm logs.

6. The AI-based console voice interaction method as described in claim 1, characterized in that, Step S5 includes: Map the semantically parsed action units to intermediate instruction templates; For tree-structured API protocols, a depth-first traversal is used to fill in the parameters. Queued parameter passing is used for linear API protocols; Default values ​​are automatically injected when a required parameter is detected to be missing.

7. The AI-based console voice interaction method as described in claim 6, characterized in that, The depth-first traversal filling parameters include: Fill in the required parameters of the lower-level components according to the protocol hierarchy; Establish a skip flag for optional parameter branches; Generate a parameter integrity verification summary and append it to the end of the instruction set.

Images