Space intelligent environment perception and decision method and system based on consistency check

Abstract

The application discloses a space intelligent environment perception and decision method and system based on consistency check, relates to the field of space intelligence, solves the problem of insufficient precision of existing multi-sensor fusion perception, and comprises the following steps: S1, collecting data of a space environment; S2, acquiring local perception data, classifying the local perception data, obtaining a time series data list, and judging data accuracy of different sensors; S3, integrating collected data of each sensor to obtain optimized space data; S4, performing space modeling according to the optimized space data to obtain a navigation world model, constructing a space semantic graph based on the navigation world model; S5, acquiring task information, extracting a route from the space semantic graph to obtain a task route, and simulating the route through the navigation world model; and S6, feeding back the navigation world model according to a simulation result, so that the accuracy of space perception can be effectively improved.

Description

Technical Field

[0001] This invention belongs to the field of spatial intelligence, and more specifically relates to multi-source perception fusion decision-making technology, specifically a spatial intelligent environment perception and decision-making method and system based on consistency verification. Background Technology

[0002] The current methods for environmental perception and decision-making have the following specific shortcomings:

[0003] 1. Existing multi-sensor fusion methods mostly rely on simple weighted averages or probabilistic statistical models for fusion. These methods lack spatial evidence consistency verification mechanisms, which can lead to judgment biases or even fusion errors in complex or dynamic scenarios. At the same time, the weights of each sensor are mostly fixed and cannot be dynamically adjusted with changes in perception quality or environment. This results in the fusion results being severely affected by noise and makes it difficult to achieve stable and robust perception output.

[0004] 2. Existing spatial intelligence systems lack constraints on physical logic and spatial causal relationships; this can easily lead to the system outputting contradictory judgments simultaneously, such as "clear ahead" and "obstacle detected", or generating geometrically infeasible trajectories in path planning.

[0005] 3. Most perception and decision-making systems are based on a unidirectional reasoning architecture, which cannot achieve self-checking and self-correction. At the same time, perception and navigation planning are independent of each other. This can easily lead to situations where the perception module identifies new obstacles, but the navigation model fails to update the map in time, and the system still follows the old path, resulting in navigation deviation or task failure. The lack of a perception-navigation collaborative optimization mechanism makes it impossible for the system to achieve globally consistent spatial decision-making.

[0006] To this end, we propose a spatial intelligent environment perception and decision-making method and system based on consistency verification. Summary of the Invention

[0007] In view of the shortcomings of existing technologies, the purpose of this invention is to provide a spatial intelligent environment perception and decision-making method and system based on consistency verification, and to improve the accuracy of spatial perception.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: a spatial intelligent environment perception and decision-making method and system based on consistency verification, the specific working process of each step of which is as follows:

[0009] Step S1: Collect data on the spatial environment to obtain spatial unit and spatial sensing data;

[0010] Step S2: Extract the spatial sensing data corresponding to the spatial unit to obtain local sensing data, classify the local sensing data to obtain a time series data list; use the time series data list to verify the data of the spatial unit and determine the data accuracy of different sensors;

[0011] Step S3: Analyze the data accuracy of the sensors to obtain a weight list of the sensors; integrate the collected data of each sensor based on the weight list to obtain optimized spatial data;

[0012] Step S4: Perform spatial modeling based on optimized spatial data to obtain a navigation world model; extract spatial nodes from the navigation world model, analyze the relationships between different spatial nodes, statistically analyze the spatial nodes and their relationships, and construct a spatial semantic graph;

[0013] Step S5: Collect optimized spatial data in real time and transmit it to the navigation world model for real-time updates; obtain task information, extract routes from the spatial semantic map to obtain the task route, and simulate the task route through the navigation world model to obtain simulation results;

[0014] Step S6: Feedback is provided to the navigation world model based on the simulation results, and the mission route is updated and simulated in conjunction with the spatial semantic map.

[0015] Furthermore, the specific steps of step S2 are as follows:

[0016] Step S21: Extract the spatial sensing data corresponding to each spatial unit to obtain local sensing data, collect the data source of the local sensing data, denote the type of data source as as, classify and statistically analyze the local sensing data jbs(a) according to the data source, and construct a local sensing data list; perform time-series processing on the local sensing data list to obtain a time-series data list.

[0017] Step S22: Analyze the local sensing data based on the time-series data list, construct a spatial evidence consistency matrix, verify the consistency of the local sensing data using the spatial evidence consistency matrix, and analyze the accuracy of the sensor data based on the verification results.

[0018] Furthermore, the specific steps of step S22 are as follows:

[0019] Step S221: Extract local sensing data from different data sources under the same time series based on the time series data list to obtain the heterogeneous synchronous data list ytb(t); traverse the heterogeneous synchronous data list and collect the spatial geometric data kjh(a,t) from each local sensing data.

[0020] The spatial geometric data is traversed to extract a data point as the standard data bzs; the standard data is then used to calculate the benchmark judgment value jzp(a) with other spatial geometric data.

[0021] Extract the standard data corresponding to the smallest benchmark judgment value jzp(a) as the spatial geometric standard kbz; calculate the spatial position deviation using the spatial geometric standard and spatial geometric data to obtain kjp(a,t).

[0022] Step S222: Traverse the heterogeneous synchronization data list and collect the spatial semantic data kyy(a,t) from each local sensing data.

[0023] The semantic data categories are statistically analyzed to obtain the semantic quantity ys; the proportion of each semantic data category to all semantic data is extracted to obtain the semantic confidence; the semantic confidence under different data sources is statistically analyzed based on the spatial semantic data kyy(a,t) to obtain the semantic confidence yyz(a,t).

[0024] Furthermore, the subsequent steps of step S222 are as follows:

[0025] Step S223: Traverse the time series of the heterogeneous synchronization data list to obtain the spatial location deviation and semantic confidence under different time series. Judge the time series stability of the data source by the spatial location deviation and semantic confidence under different time series to obtain the time series stability value sxw(a).

[0026] Step S224: Construct a spatial evidence consistency matrix from spatial semantic data kyy(a,t), semantic confidence yyz(a,t), and time-series stable value sxw(a); Based on the spatial evidence consistency matrix, combine the spatial semantic data kyy(a,t), semantic confidence yyz(a,t), and time-series stable value sxw(a) to obtain the data accuracy zqd(a).

[0027] Furthermore, the specific steps of step S3 are as follows:

[0028] Step S31: Real-time acquisition of data accuracy zqd(a) from different sensors; Calculation of weights for different sensors based on their data accuracy zqd(a) to obtain weight values ​​qzz(a);

[0029] The weight values ​​are statistically analyzed to obtain the sensor weight list qzl = [qzz(1) to qzz(as)];

[0030] Step S32: Obtain the collected data cjs(a) of each sensor, and perform weighted processing on the collected data by combining the sensor weight list qzl = [qzz(1) to qzz(as)] to obtain the weighted data jqs(a);

[0031] The weighted data is fused, the range of fused data is recorded, and the data is traversed based on the range of fused data. The fused data is denoted as rhs. The data collected by each sensor, cjs(a), is combined with the sensor weights to calculate the fused data judgment value rpd.

[0032] The fused data corresponding to the smallest fused data judgment value is statistically analyzed to obtain optimized spatial data.

[0033] Furthermore, the specific steps of step S4 are as follows:

[0034] Step S41: Construct a 3D spatial model. Import the optimized spatial data into the 3D spatial model to construct a navigation world model. Traverse the navigation world model and extract spatial nodes. Count the spatial nodes to obtain the number of spatial nodes.

[0035] Step S42: Extract a spatial node as the initial node, connect the initial node with other spatial nodes, obtain the local perception data of the connected spatial nodes, extract the spatial geometric data and spatial semantic data of the above spatial nodes, and judge the connection relationship of the spatial nodes based on the spatial geometric data and spatial semantic data. If the spatial geometric data of the above spatial nodes is continuous, it indicates that the above spatial nodes have topological edges. If the spatial semantic data of the above spatial nodes have logical connection relationships, it indicates that the above spatial nodes have semantic edges. Connect the spatial nodes with topological edges and semantic edges and record them.

[0036] Step S43: Based on the number of spatial nodes, traverse the initial node and the connected nodes of the initial node to obtain the connection relationship between all nodes; count the connection relationship between spatial nodes to construct a spatial semantic graph.

[0037] Furthermore, the specific steps of step S5 are as follows:

[0038] Step S51: Real-time acquisition of monitoring data from different sensors, real-time judgment and verification of optimized spatial data based on real-time monitoring data from different sensors, obtaining real-time optimized spatial data, transmitting real-time optimized spatial data to the navigation world model, and updating the navigation world model in real time;

[0039] Step S52: Based on the task information, traverse the navigation world model, extract the starting and ending positions of the task information on the navigation world model, extract the corresponding spatial nodes from the starting and ending positions, and obtain the initial node and the termination node.

[0040] Furthermore, the subsequent steps of step S52 are as follows:

[0041] Step S53: Traverse the initial node and the termination node on the spatial semantic graph. Based on the connection relationship between spatial nodes on the spatial semantic graph, count the connected spatial nodes, perform route planning, extract the spatial node routes that are consistent with the spatial geometric relationship and logical relationship, and obtain the task route.

[0042] Step S54: Simulate the navigation world model according to the mission route, and record the spatial passability and logical passability of the spatial nodes passed through during the route simulation; statistically analyze the spatial passability and logical passability of all spatial nodes on the mission route to obtain the simulation results.

[0043] Furthermore, the specific steps of step S6 are as follows:

[0044] Step S61: Obtain the spatial passability and logical passability of all spatial nodes based on the simulation nodes. If the spatial passability and logical passability of a spatial node are abnormal, record the spatial node as an abnormal node; record the cause of the abnormality of the abnormal node and upload it to the staff.

[0045] Step S62: Extract abnormal nodes, treat them as unconnectable nodes, re-traverse the spatial state graph, update the task route, and repeatedly simulate and judge the task route. When the spatial passability and logical passability of all spatial nodes in the simulation results are normal, the simulation process is completed, and the task route is output.

[0046] A spatial intelligent environment perception and decision-making system based on consistency verification, the decision-making system including:

[0047] Data acquisition module: Collects data from the space environment to obtain spatial unit and spatial sensing data;

[0048] Sensor judgment module: Extracts spatial sensing data corresponding to spatial units to obtain local sensing data, classifies local sensing data to obtain a time-series data list; uses the time-series data list to verify the data of spatial units and judge the data accuracy of different sensors;

[0049] Data integration module: Analyzes the accuracy of sensor data to obtain a weight list of sensors; integrates the data collected by each sensor based on the weight list to obtain optimized spatial data;

[0050] Semantic graph construction module: Based on optimized spatial data, spatial modeling is performed to obtain a navigation world model; spatial nodes are extracted from the navigation world model, the relationships between different spatial nodes are analyzed, and the spatial nodes and their relationships are statistically analyzed to construct a spatial semantic graph;

[0051] Route simulation module: Real-time acquisition of optimized spatial data, transmission to the navigation world model for real-time updates; acquisition of task information, extraction of routes from the spatial semantic map to obtain the task route, simulation of the task route through the navigation world model to obtain simulation results;

[0052] Feedback Analysis Module: Provides feedback to the navigation world model based on simulation results, updates the mission route and conducts simulations in conjunction with the spatial semantic map.

[0053] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0054] 1. This invention improves the data accuracy of sensors by multi-sensor fusion, and analyzes the spatial geometric data and spatial semantic data of each sensor to determine the data accuracy of each sensor. Based on the data accuracy, dynamic weights are set for the sensors, and the data is calculated by the weights to ensure the stability and availability of the data.

[0055] 2. This invention analyzes the spatial geometric data and spatial semantic data of the environment to determine the geometric and logical relationships between different spatial nodes. It performs a dual analysis of the connection relationships between spatial nodes based on the geometric and logical relationships, ensuring the consistency of physical logic and spatial causality between spatial nodes.

[0056] 3. This invention updates the model by collecting data in real time to ensure the model's accuracy, simulates and plans routes using the model, and analyzes the feasibility and stability of routes in real time; it also provides feedback to the model based on the analysis results to achieve adaptive optimization of routes. Attached Figure Description

[0057] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0058] Figure 1 This is a schematic diagram of the method of the present invention;

[0059] Figure 2 This is a schematic diagram of the data processing of the present invention;

[0060] Figure 3 This is a functional block diagram of the present invention. Detailed Implementation

[0061] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0062] This application provides a spatial intelligent environment perception and decision-making method based on consistency verification. The executing entity of the spatial intelligent environment perception and decision-making method based on consistency verification includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the spatial intelligent environment perception and decision-making method based on consistency verification can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0063] Reference Figure 1 The diagram shown is a flowchart illustrating a spatial intelligent environment perception and decision-making method based on consistency verification according to an embodiment of the present invention. In this embodiment, the spatial intelligent environment perception and decision-making method based on consistency verification includes:

[0064] Step S1: Collect data on the spatial environment to obtain spatial perception data, and segment the spatial environment using the spatial perception data to obtain spatial units;

[0065] Step S11: Simultaneously collect environmental information (including images, depth, point clouds, radar signals, inertial data, and semantic tags) from multiple sources such as visual sensors, LiDAR, depth cameras, IMU, infrared and voice sensors to obtain spatial perception data. Based on the spatial perception data, perform semantic segmentation on the spatial environment to obtain spatial units.

[0066] Step S2: Extract the spatial sensing data corresponding to the spatial unit to obtain local sensing data, count the local sensing data at multiple time points, and classify them according to their data sources to obtain a time series data list; construct a spatial evidence consistency matrix from the time series data list, verify the data of the spatial unit, and determine the accuracy of the data from different sensors;

[0067] Step S21: Extract the spatial sensing data corresponding to each spatial unit to obtain local sensing data, collect the data source of the local sensing data, and denote the type of data source as as. Classify and count the local sensing data jbs(a) according to the data source to construct a local sensing data list [jbs(1) to jbs(as)]; perform time-series processing on the local sensing data list to obtain a time-series data list [jbs(1,t) to jbs(as,t)];

[0068] Step S22: Analyze the local sensing data based on the time-series data list, construct a spatial evidence consistency matrix, verify the consistency of the local sensing data using the spatial evidence consistency matrix, and analyze the accuracy of the sensor data based on the verification results.

[0069] Please see Figure 2 Step S221: Extract local sensing data from different data sources under the same time series based on the time series data list to obtain a heterogeneous synchronous data list ytb(t), ytb(t) = [jbs(1,t) to jbs(as,t)]; Traverse the heterogeneous synchronous data list and collect the spatial geometric data kjh(a,t) from each local sensing data.

[0070] It should be noted that spatial geometric data is three-dimensional data with the spatial environment as the main body; local perception data without spatial geometric data is assigned an empty value and is not processed.

[0071] The spatial geometric data is traversed to extract a data point as the standard data bzs; the standard data is then used to calculate the benchmark judgment value jzp(a) with other spatial geometric data.

[0072] ;

[0073] Extract the standard data corresponding to the smallest benchmark judgment value jzp(a) as the spatial geometric standard kbz; calculate the spatial position deviation using the spatial geometric standard and spatial geometric data to obtain kjp(a,t) = kjh(a,t) - kbz;

[0074] Step S222: Traverse the heterogeneous synchronization data list and collect the spatial semantic data kyy(a,t) from each local sensing data.

[0075] It should be noted that spatial semantic data refers to the semantic features of spatial units; local perceptual data without spatial semantic data is assigned an empty value and is not processed.

[0076] The semantic data categories are statistically analyzed to obtain the semantic quantity ys; the proportion of each semantic data category to all semantic data is extracted to obtain the semantic confidence; the semantic confidence under different data sources is statistically analyzed based on the spatial semantic data kyy(a,t) to obtain the semantic confidence yyz(a,t).

[0077] Step S223: Traverse the time series of the heterogeneous synchronization data list to obtain the spatial location deviation and semantic confidence under different time series. Judge the time series stability of the data source by the spatial location deviation and semantic confidence under different time series to obtain the time series stability value sxw(a).

[0078] ;

[0079] It should be noted that kjp(a,t)≤kjp(a,t-1)? 1:0; is a ternary expression, that is, when kjp(a,t)≤kjp(a,t-1), the value is 1, otherwise the value is 0; by traversing the spatial position deviation of multiple time nodes through the ternary expression, when the spatial position deviation decreases, the surface stability is higher, and the total time node with the decrease is recorded. Similarly, the confidence of spatial semantic data is traversed, and the two are combined to obtain the temporal stability of spatial nodes from different temporal sources.

[0080] Step S224: Construct a spatial evidence consistency matrix using spatial semantic data kyy(a,t), semantic confidence yyz(a,t), and temporal stable value sxw(a); Based on the spatial evidence consistency matrix, combine the spatial semantic data kyy(a,t), semantic confidence yyz(a,t), and temporal stable value sxw(a) to obtain the data accuracy zqd(a) = kyy(a,t) × yyz(a,t) × sxw(a);

[0081] It should be noted that in the spatial evidence consistency matrix, t in the spatial semantic data kyy(a,t) and semantic confidence yyz(a,t) represents the current time node.

[0082] Step S3: Analyze the data accuracy of the sensors using a multi-objective optimization algorithm, calculate the optimal weight combination of each sensing source in real time, and obtain a weight list of the sensors; perform weighted processing on the data collected by each sensor based on the weight list of the sensors, and integrate them to obtain optimized spatial data;

[0083] Step S31: Real-time acquisition of data accuracy zqd(a) from different sensors; Calculation of weights for different sensors based on their data accuracy zqd(a) to obtain weight values ​​qzz(a);

[0084] ;

[0085] The weight values ​​are statistically analyzed to obtain the sensor weight list qzl = [qzz(1) to qzz(as)];

[0086] Step S32: Obtain the collected data cjs(a) of each sensor, and perform weighted processing on the collected data by combining the sensor weight list qzl = [qzz(1) to qzz(as)] to obtain the weighted data jqs(a);

[0087] ;

[0088] The weighted data is fused, the range of fused data is recorded, and the data is traversed based on the range of fused data. The fused data is denoted as rhs. The data collected by each sensor, cjs(a), is combined with the sensor weights to calculate the fused data judgment value rpd.

[0089] ;

[0090] The fused data corresponding to the smallest fused data judgment value is statistically analyzed to obtain optimized spatial data.

[0091] Step S4: Perform spatial modeling based on optimized spatial data to obtain a navigation world model; extract spatial nodes from the navigation world model, traverse the spatial nodes, analyze the relationships between different spatial nodes, statistically analyze the spatial nodes and their relationships, and construct a spatial semantic graph.

[0092] Step S41: Construct a 3D spatial model. Import the optimized spatial data into the 3D spatial model to construct a navigation world model. Traverse the navigation world model and extract spatial nodes. Count the spatial nodes to obtain the number of spatial nodes.

[0093] Step S42: Extract a spatial node as the initial node, connect the initial node with other spatial nodes, obtain the local perception data of the connected spatial nodes, extract the spatial geometric data and spatial semantic data of the above spatial nodes, and judge the connection relationship of the spatial nodes based on the spatial geometric data and spatial semantic data. If the spatial geometric data of the above spatial nodes is continuous, it indicates that the above spatial nodes have topological edges. If the spatial semantic data of the above spatial nodes have logical connection relationships, it indicates that the above spatial nodes have semantic edges. Connect the spatial nodes with topological edges and semantic edges and record them.

[0094] Step S43: Based on the number of spatial nodes, traverse the initial node and the connected nodes of the initial node to obtain the connection relationship between all nodes; count the connection relationship between spatial nodes to construct a spatial semantic graph.

[0095] Step S5: Collect optimized spatial data in real time and transmit it to the navigation world model for real-time updates; obtain task information, extract routes from the spatial semantic map based on the task information to obtain the task route, and simulate the task route through the navigation world model to obtain the simulation results;

[0096] Step S51: Real-time acquisition of monitoring data from different sensors, real-time judgment and verification of optimized spatial data based on real-time monitoring data from different sensors, obtaining real-time optimized spatial data, transmitting real-time optimized spatial data to the navigation world model, and updating the navigation world model in real time;

[0097] Step S52: Based on the task information, traverse the navigation world model, extract the starting position and ending position of the task information on the navigation world model, extract the corresponding spatial nodes from the starting position and ending position, and obtain the initial node and the termination node.

[0098] Step S53: Traverse the initial node and the termination node on the spatial semantic graph. Based on the connection relationship between spatial nodes on the spatial semantic graph, count the connected spatial nodes, perform route planning, extract the spatial node routes that are consistent with the spatial geometric relationship and logical relationship, and obtain the task route.

[0099] Step S54: Simulate the navigation world model according to the mission route, and record the spatial passability and logical passability of the spatial nodes passed through during the route simulation; statistically analyze the spatial passability and logical passability of all spatial nodes on the mission route to obtain the simulation results.

[0100] Step S6: Feedback is provided to the navigation world model based on the simulation results, abnormal nodes are recorded, the causes of abnormal nodes are analyzed, the causes of abnormalities are fed back to the sensors, the navigation world model is updated, and the mission route is updated and simulated in conjunction with the spatial semantic map.

[0101] Step S61: Obtain the spatial passability and logical passability of all spatial nodes based on the simulation nodes. If the spatial passability and logical passability of a spatial node are abnormal, record the spatial node as an abnormal node; record the cause of the abnormality of the abnormal node and upload it to the staff.

[0102] Step S62: Extract abnormal nodes, treat them as unconnectable nodes, re-traverse the spatial state graph, update the task route, and repeatedly simulate and judge the task route. When the spatial passability and logical passability of all spatial nodes in the simulation results are normal, the simulation process is completed, and the task route is output.

[0103] Compared to the problems described in the background technology, this invention improves the data accuracy of sensors through multi-sensor fusion. It analyzes the spatial geometric and semantic data of each sensor to determine the accuracy of each sensor's data, and establishes dynamic weights for the sensors based on this accuracy. These weights are then used to calculate data stability and usability. Furthermore, this invention analyzes the spatial geometric and semantic data of the environment to determine the geometric and logical relationships between different spatial nodes. This dual analysis of geometric and logical relationships ensures consistency between the physical logic and spatial causality of the spatial nodes. Finally, this invention updates the model in real-time through data acquisition to ensure model accuracy. The model is used to simulate and plan routes, and the feasibility and stability of the routes are analyzed in real-time. Feedback is provided to the model based on the analysis results to achieve adaptive route optimization. Therefore, the spatial intelligent environmental perception and decision-making method based on consistency verification provided by this invention can improve the accuracy of spatial perception.

[0104] like Figure 3 The diagram shown is a functional block diagram of the spatial intelligent environment perception and decision-making system based on consistency verification according to the present invention.

[0105] The spatial intelligent environment perception and decision-making system based on consistency verification described in this invention can be installed in an electronic device. Depending on the functions implemented, the spatial intelligent environment perception and decision-making system based on consistency verification may include a data acquisition module, a sensor judgment module, a data integration module, a semantic graph construction module, a route simulation module, and a feedback analysis module. The modules described in this invention can also be referred to as units, which are a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, stored in the memory of the electronic device.

[0106] In this embodiment of the invention, the functions of each module / unit are as follows:

[0107] The data acquisition module collects data from the spatial environment to obtain spatial perception data, and then segments the spatial environment using the spatial perception data to obtain spatial units.

[0108] The sensor judgment module is used to extract the spatial sensing data corresponding to the spatial unit to obtain local sensing data, count the local sensing data at multiple time points, and classify them according to their data sources to obtain a time-series data list; construct a spatial evidence consistency matrix from the time-series data list, perform data verification on the spatial unit, and judge the data accuracy of different sensors.

[0109] The data integration module is used to analyze the data accuracy of the sensors, calculate the optimal weight combination of each sensing source in real time, and obtain a weight list of the sensors; based on the weight list of the sensors, the collected data of each sensor is weighted and integrated to obtain optimized spatial data.

[0110] The semantic graph construction module is used to perform spatial modeling based on optimized spatial data to obtain a navigation world model; extract spatial nodes from the navigation world model, traverse the spatial nodes, analyze the relationships between different spatial nodes, statistically analyze the spatial nodes and their relationships, and construct a spatial semantic graph.

[0111] The route simulation module is used to collect optimized spatial data in real time, transmit it to the navigation world model for real-time updates, obtain task information, extract routes from the spatial semantic map based on the task information to obtain the task route, and simulate the task route through the navigation world model to obtain the simulation results.

[0112] The feedback analysis module is used to provide feedback to the navigation world model based on the simulation results, record abnormal nodes, analyze the causes of abnormal nodes, provide feedback to the sensors based on the causes of abnormalities, update the navigation world model, update the mission route and simulate it in conjunction with the spatial semantic graph.

[0113] In detail, the modules in the spatial intelligent environment perception and decision-making system based on consistency verification described in this embodiment of the invention adopt the same approach as described above. Figure 1 The method used is the same as the spatial intelligent environment perception and decision-making method based on consistency verification described in the article, and can produce the same technical effect, so it will not be elaborated here.

[0114] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0115] Finally, it should be noted that deleting any one of the above embodiments does not affect the technical solutions of other embodiments. The above embodiments are only used to illustrate the technical solutions of the present invention and not 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.

Claims

1. A spatial intelligent environment perception and decision-making method based on consistency verification, characterized in that, include: Step S1: Collect data on the spatial environment to obtain spatial unit and spatial sensing data; Step S2: Extract the spatial sensing data corresponding to the spatial unit to obtain local sensing data, classify the local sensing data to obtain a time series data list; use the time series data list to verify the data of the spatial unit and determine the data accuracy of different sensors; The specific steps of step S2 are as follows: Step S21: Extract the spatial sensing data corresponding to each spatial unit to obtain local sensing data, collect the data source of the local sensing data, denote the type of data source as as, classify and statistically analyze the local sensing data jbs(a) according to the data source, and construct a local sensing data list. The local sensing data list is processed to obtain a time-series data list; Step S22: Analyze the local sensing data based on the time-series data list, construct a spatial evidence consistency matrix, verify the consistency of the local sensing data using the spatial evidence consistency matrix, and analyze the accuracy of the sensor data based on the verification results; The specific steps of step S22 are as follows: Step S221: Extract local sensing data from different data sources under the same time series based on the time series data list to obtain the heterogeneous synchronous data list ytb(t); traverse the heterogeneous synchronous data list and collect the spatial geometric data kjh(a,t) from each local sensing data. The spatial geometric data is traversed, and a single data point is extracted as the standard data bzs. The benchmark judgment value jzp(a) is obtained by calculating the standard data with other spatial geometric data. Extract the standard data corresponding to the smallest benchmark judgment value jzp(a) as the spatial geometric standard kbz; calculate the spatial position deviation using the spatial geometric standard and spatial geometric data to obtain kjp(a,t). Step S222: Traverse the heterogeneous synchronization data list and collect the spatial semantic data kyy(a,t) from each local sensing data. The semantic data categories are statistically analyzed to obtain the semantic quantity ys; the proportion of each semantic data category to all semantic data is extracted to obtain the semantic confidence; the semantic confidence under different data sources is statistically analyzed based on the spatial semantic data kyy(a,t) to obtain the semantic confidence yyz(a,t). Step S3: Analyze the data accuracy of the sensors to obtain a weight list of the sensors; integrate the collected data of each sensor based on the weight list to obtain optimized spatial data; Step S4: Perform spatial modeling based on optimized spatial data to obtain a navigation world model; Spatial nodes are extracted from the navigation world model, the relationships between different spatial nodes are analyzed, and the spatial nodes and their relationships are statistically analyzed to construct a spatial semantic graph. Step S5: Collect optimized spatial data in real time and transmit it to the navigation world model for real-time updates; obtain task information, extract routes from the spatial semantic map to obtain the task route, and simulate the task route through the navigation world model to obtain simulation results; Step S6: Feedback is provided to the navigation world model based on the simulation results, and the mission route is updated and simulated in conjunction with the spatial semantic map.

2. The spatial intelligent environment perception and decision-making method based on consistency verification according to claim 1, characterized in that, The subsequent steps of step S222 are as follows: Step S223: Traverse the time series of the heterogeneous synchronization data list to obtain the spatial location deviation and semantic confidence under different time series. Judge the time series stability of the data source by the spatial location deviation and semantic confidence under different time series to obtain the time series stability value sxw(a). Step S224: Construct a spatial evidence consistency matrix from spatial semantic data kyy(a,t), semantic confidence yyz(a,t), and time-series stable value sxw(a); Based on the spatial evidence consistency matrix, combine the spatial semantic data kyy(a,t), semantic confidence yyz(a,t), and time-series stable value sxw(a) to obtain the data accuracy zqd(a).

3. The spatial intelligent environment perception and decision-making method based on consistency verification according to claim 1, characterized in that, The specific steps of step S3 are as follows: Step S31: Real-time acquisition of data accuracy zqd(a) from different sensors; Calculation of weights for different sensors based on their data accuracy zqd(a) to obtain weight values ​​qzz(a); The weight values ​​are statistically analyzed to obtain the sensor weight list qzl = [qzz(1) to qzz(as)]; Step S32: Obtain the collected data cjs(a) of each sensor, and perform weighted processing on the collected data by combining the sensor weight list qzl = [qzz(1) to qzz(as)] to obtain the weighted data jqs(a); The weighted data is merged, the range of the merged data is recorded, the data is traversed based on the range of the merged data, and the merged data is denoted as rhs; The fused data judgment value rpd is obtained by combining the data cjs(a) collected by each sensor with the sensor weights. The fused data corresponding to the smallest fused data judgment value is statistically analyzed to obtain optimized spatial data.

4. The spatial intelligent environment perception and decision-making method based on consistency verification according to claim 1, characterized in that, The specific steps of step S4 are as follows: Step S41: Construct a 3D spatial model by importing optimized spatial data into the 3D spatial model to build a navigation world model; traverse the navigation world model and extract spatial nodes. The number of spatial nodes is obtained by counting the spatial nodes; Step S42: Extract a spatial node as the initial node, connect the initial node with other spatial nodes, obtain the local perception data of the connected spatial nodes, extract the spatial geometric data and spatial semantic data of the above spatial nodes, and judge the connection relationship of the spatial nodes based on the spatial geometric data and spatial semantic data. If the spatial geometric data of the spatial nodes is continuous, it indicates that the spatial nodes have topological edges. If the spatial semantic data of the spatial nodes have logical connection relationships, it indicates that the spatial nodes have semantic edges. Connect the spatial nodes with topological edges and semantic edges and record them. Step S43: Based on the number of spatial nodes, traverse the initial node and the nodes connected to the initial node to obtain the connection relationships between all nodes; By statistically analyzing the connections between spatial nodes, a spatial semantic graph is constructed.

5. The spatial intelligent environment perception and decision-making method based on consistency verification according to claim 1, characterized in that, The specific steps of step S5 are as follows: Step S51: Real-time acquisition of monitoring data from different sensors, real-time judgment and verification of optimized spatial data based on real-time monitoring data from different sensors, obtaining real-time optimized spatial data, transmitting real-time optimized spatial data to the navigation world model, and updating the navigation world model in real time; Step S52: Based on the task information, traverse the navigation world model, extract the starting and ending positions of the task information on the navigation world model, extract the corresponding spatial nodes from the starting and ending positions, and obtain the initial node and the termination node.

6. The spatial intelligent environment perception and decision-making method based on consistency verification according to claim 5, characterized in that, The subsequent steps of step S52 are as follows: Step S53: Traverse the initial node and the termination node on the spatial semantic graph. Based on the connection relationship between spatial nodes on the spatial semantic graph, count the connected spatial nodes, perform route planning, extract the spatial node routes that are consistent with the spatial geometric relationship and logical relationship, and obtain the task route. Step S54: Simulate the navigation world model according to the mission route, and record the spatial passability and logical passability of the spatial nodes traversed during the route simulation. The spatial and logical passability of all spatial nodes along the mission route is statistically analyzed to obtain simulation results.

7. The spatial intelligent environment perception and decision-making method based on consistency verification according to claim 1, characterized in that, The specific steps of step S6 are as follows: Step S61: Obtain the spatial passability and logical passability of all spatial nodes based on the simulation nodes. If the spatial passability and logical passability of a spatial node are abnormal, record the spatial node as an abnormal node; record the cause of the abnormality of the abnormal node and upload it to the staff. Step S62: Extract abnormal nodes, treat them as unconnectable nodes, re-traverse the spatial state graph, update the task route, and repeatedly simulate and judge the task route. When the spatial passability and logical passability of all spatial nodes in the simulation results are normal, the simulation process is completed, and the task route is output.

8. A spatial intelligent environment perception and decision-making system based on consistency verification, applicable to the spatial intelligent environment perception and decision-making method based on consistency verification as described in any one of claims 1-7, characterized in that, The decision-making system includes: Data acquisition module: Collects data from the space environment to obtain spatial unit and spatial sensing data; Sensor judgment module: Extracts spatial sensing data corresponding to spatial units to obtain local sensing data, classifies local sensing data to obtain a time-series data list; uses the time-series data list to verify the data of spatial units and judge the data accuracy of different sensors; Data integration module: Analyzes the accuracy of sensor data to obtain a weight list of sensors; integrates the data collected by each sensor based on the weight list to obtain optimized spatial data; Semantic graph construction module: Based on optimized spatial data, spatial modeling is performed to obtain a navigation world model; spatial nodes are extracted from the navigation world model, the relationships between different spatial nodes are analyzed, and the spatial nodes and their relationships are statistically analyzed to construct a spatial semantic graph; Route simulation module: Real-time acquisition of optimized spatial data, transmission to the navigation world model for real-time updates; acquisition of task information, extraction of routes from the spatial semantic map to obtain the task route, simulation of the task route through the navigation world model to obtain simulation results; Feedback Analysis Module: Provides feedback to the navigation world model based on simulation results, updates the mission route and conducts simulations in conjunction with the spatial semantic map.

Images