High-precision telemetering terminal data acquisition system and method based on multi-source sensor fusion

Abstract

The invention provides a high-precision telemetering terminal data acquisition system and method based on multi-source sensor fusion. A high-precision telemetering terminal data acquisition system based on multi-source sensor fusion comprises a multi-source sensor module which is composed of a water level sensor, a flow sensor, a water quality sensor, a meteorological sensor and a vibration sensor, and all the sensors are connected with a main control unit through standardized interfaces; the system comprises a data fusion processing unit, a dynamic priority scheduling module, an edge computing unit, a multi-channel communication module and an anti-interference power management module. The high-precision telemetering terminal data acquisition system and method based on multi-source sensor fusion have the advantages that the centimeter-level precision of hydrological data can be achieved through multi-source dynamic fusion and intelligent verification, the availability of communication is improved, and meanwhile the operation and maintenance cost can be reduced.

Description

Technical Field

[0001] The present invention relates to the field of robotics technology, and in particular to a high-precision telemetry terminal data acquisition system and method based on multi-source sensor fusion. Background Art

[0002] Traditional telemetry terminals often use a single sensor to collect hydrological data, resulting in data silos and poor anti-interference capabilities. While existing technologies offer some multi-sensor systems, they lack dynamic fusion mechanisms, leading to high data conflict rates and low reliability in complex environments. Furthermore, existing communication modules often rely on a single transmission mode, making communication prone to interruptions in extreme weather conditions. Furthermore, fixed sensor calibration cycles make them unsuitable for adapting to sudden hydrological events.

[0003] Therefore, it is necessary to provide a high-precision telemetry terminal data acquisition system and method based on multi-source sensor fusion to solve the above technical problems. Summary of the Invention

[0004] The technical problem solved by the present invention is to provide a high-precision telemetry terminal data acquisition system and method based on multi-source sensor fusion, which can achieve centimeter-level accuracy of hydrological data through multi-source dynamic fusion and intelligent verification, improve the availability of communication, and reduce operation and maintenance costs.

[0005] To solve the above technical problems, the present invention provides a high-precision telemetry terminal data acquisition system based on multi-source sensor fusion, including: a multi-source sensor module: composed of a water level sensor, a flow sensor, a water quality sensor, a meteorological sensor and a vibration sensor, each sensor being connected to a main control unit via a standardized interface;

[0006] Data fusion processing unit: Built-in Kalman filter algorithm, adaptive weighted fusion algorithm and spatiotemporal alignment module, used to synchronize timestamps, calibrate spatial positions and compensate errors for heterogeneous data from multi-source sensors;

[0007] Dynamic priority scheduling module: Dynamically allocates sampling frequency based on sensor type and environmental status. Water level and flow sensors are prioritized to millisecond-level sampling during flood season, while water quality sensors use minute-level sampling in steady state.

[0008] Edge computing unit: Integrates an LSTM neural network model to predict sensor data anomalies in real time and trigger redundancy verification mechanisms;

[0009] Multi-channel communication module: supports multi-mode switching of 4G / 5G, LoRa, and Beidou satellite communications, and automatically selects the optimal transmission path according to network conditions;

[0010] Anti-interference power management module: adopts solar-supercapacitor dual power supply system, built-in pulse current suppression circuit and sensor isolation power supply unit.

[0011] Preferably, the data fusion processing unit further includes:

[0012] The time-space alignment module unifies the time-space benchmarks of sensor data through GPS timing and Beidou differential positioning, with a time synchronization accuracy of ≤1ms and a spatial positioning error of ≤0.1m.

[0013] The adaptive weighted fusion algorithm dynamically adjusts the weight according to the historical credibility of the sensor. The weight calculation formula of the dissolved oxygen data of the water quality sensor is:

[0014]

[0015] in, is the sensor variance, t is the time interval since the last calibration, and λ is the attenuation coefficient.

[0016] Preferably, the edge computing unit performs the following operations:

[0017] The prediction interval of the flow sensor data is established through the LSTM model. When the real-time data deviates from the predicted value by ±15%, the redundant sensor group cross-validation is started;

[0018] If the pH value in the water quality sensor conflicts with the conductivity data, the association rules in the historical database are called for logical verification.

[0019] Preferably, the deployment method of the multi-source sensor module includes:

[0020] The water level sensor adopts radar and pressure dual-mode deployment, and the flow sensor adopts ultrasonic Doppler array and electromagnetic complementary deployment;

[0021] The vibration sensor is installed at the key nodes of the pipeline in the form of a triaxial accelerometer, and the sampling frequency varies linearly with the water flow rate.

[0022] The present invention also provides a method for collecting high-precision telemetry terminal data based on multi-source sensor fusion, comprising the following steps:

[0023] S1: Water level, flow, water quality parameters and environmental data are collected synchronously through the multi-source sensor module. The collection frequency is dynamically adjusted by the dynamic priority scheduling module according to the hydrological characteristics;

[0024] S2: Perform spatiotemporal alignment on the raw data, unify the time base using GPS / Beidou differential signals, and compensate for differences in sensor placement using a spatial interpolation algorithm.

[0025] S3: Adopting the improved DS evidence theory to fuse multi-source data, the entropy weight method is introduced to modify the basic probability distribution function for the conflicting evidence set;

[0026] S4: Anomaly detection is performed through the edge computing unit. If the traffic data mutation exceeds the threshold, the dual verification process of the ultrasonic and electromagnetic sensors is initiated;

[0027] S5: Compress the fused data into JSON-LD format and upload it to the cloud platform through the multi-channel communication module, while storing a copy of the original data with a digital signature locally.

[0028] Preferably, the specific process of the improved DS evidence theory in S3 includes:

[0029] Calculate the confidence interval and uncertainty of each sensor evidence body. The confidence level Bel(A) of the water quality sensor is expressed as:

[0030]

[0031] Where m(B) is the evidence quality function;

[0032] The Josselme distance metric is used for conflicting evidence, and when the distance is ≥ 0.5, the entropy weight method is triggered to redistribute the weights.

[0033] Preferably, the specific method of anomaly detection in S4 is:

[0034] Construct the flow-water level joint probability density function f(Q, H), and mark the real-time data as abnormal when it falls outside the 95% confidence region;

[0035] A sliding window CUSUM algorithm is used to detect drift faults in pH sensors, and the window size is adaptively adjusted according to the rate of change of water quality.

[0036] Preferably, the method further includes a data backtracking verification step:

[0037] Establish a blockchain-based sensor data evidence chain in the cloud, and generate a Merkle tree hash value for each collection record;

[0038] When data disputes occur, the smart contract is called to compare the local storage on the terminal with the cloud hash value to verify data integrity.

[0039] Preferably, the calibration method of the multi-source sensor module includes:

[0040] Perform online self-calibration every 24 hours to activate the self-cleaning electrode of the water quality sensor by injecting standard solution;

[0041] The water level sensor uses a laser rangefinder as a temporary reference benchmark, and an alarm is triggered when the calibration error exceeds 2%.

[0042] Compared with related technologies, the high-precision telemetry terminal data acquisition system and method based on multi-source sensor fusion provided by the present invention has the following beneficial effects:

[0043] The present invention provides a high-precision telemetry terminal data acquisition system and method based on multi-source sensor fusion. Through the spatiotemporal alignment and adaptive weighted fusion of multi-source sensors, the water level measurement error can be reduced, thereby improving the confidence of the flow data; through dynamic priority scheduling and edge computing redundancy verification, the continuous operation failure rate of the equipment during the flood season can be reduced; when the 4G signal is lost, the multi-mode switching communication module can automatically switch to the Beidou satellite in ≤3 seconds, ensuring the real-time nature of the data; the online self-calibration function reduces the frequency of manual calibration from once a week to once a quarter, thereby saving annual operation and maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 This is a functional block diagram of the high-precision telemetry terminal data acquisition system based on multi-source sensor fusion provided by the present invention;

[0045] Figure 2 This is a flow chart of the high-precision telemetry terminal data acquisition method based on multi-source sensor fusion provided by the present invention. DETAILED DESCRIPTION

[0046] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0047] Please refer to Figure 1 and Figure 2 ,in, Figure 1 This is a functional block diagram of the high-precision telemetry terminal data acquisition system based on multi-source sensor fusion provided by the present invention; Figure 2 This is a flow chart of the high-precision telemetry terminal data acquisition method based on multi-source sensor fusion provided by the present invention. The high-precision telemetry terminal data acquisition system and method based on multi-source sensor fusion include: a multi-source sensor module: including a water level sensor, a flow sensor, a water quality sensor, a meteorological sensor, and a vibration sensor, each sensor being connected to the main control unit via a standardized interface;

[0048] The transducer spacing of the ultrasonic array of the flow sensor is set to 1 / 8 of the pipe diameter to obtain the best Doppler effect;

[0049] The water quality sensor includes pH value, dissolved oxygen, turbidity, and conductivity detection units, wherein the dissolved oxygen sensor needs to be equipped with a temperature compensation unit with a compensation coefficient α = 0.03 / °C;

[0050] The meteorological sensors include a temperature sensor, a humidity sensor and a wind speed sensor;

[0051] Data fusion processing unit: Built-in Kalman filter algorithm, adaptive weighted fusion algorithm and spatiotemporal alignment module, used to synchronize timestamps, calibrate spatial positions and compensate errors for heterogeneous data from multi-source sensors;

[0052] Among them, in the built-in Kalman filter algorithm, the state transfer matrix Q takes the diagonal matrix [0.01, 0.01, 0.05];

[0053] In the adaptive weighted fusion algorithm, the weight update period Δt = 300s ± 10%;

[0054] Dynamic priority scheduling module: Dynamically allocates sampling frequency based on sensor type and environmental status. Water level and flow sensors are prioritized to millisecond-level sampling during flood season, while water quality sensors use minute-level sampling in steady state.

[0055] Edge computing unit: Integrates an LSTM neural network model to predict sensor data anomalies in real time and trigger redundancy verification mechanisms;

[0056] Multi-channel communication module: supports multi-mode switching of 4G / 5G, LoRa, and Beidou satellite communications, and automatically selects the optimal transmission path according to network conditions;

[0057] Anti-interference power management module: adopts solar-supercapacitor dual power supply system, built-in pulse current suppression circuit and sensor isolation power supply unit.

[0058] The data fusion processing unit further includes:

[0059] The spatiotemporal alignment module unifies the spatiotemporal benchmarks of sensor data through GPS timing and Beidou differential positioning, with a time synchronization accuracy of ≤1ms and a spatial positioning error of ≤0.1m.

[0060] The time-space alignment module synchronizes the clocks of each sensor through GPS second pulses (the synchronization error compensation formula is Δt_c=1.2×10 -3 ×T 2 , T is the number of hops from the master node);

[0061] The adaptive weighted fusion algorithm dynamically adjusts the weight according to the historical credibility of the sensor (such as calibration cycle and failure rate). The weight calculation formula of the dissolved oxygen data of the water quality sensor is:

[0062]

[0063] in, is the sensor variance, t is the time interval since the last calibration, and λ is the attenuation coefficient.

[0064] The edge computing unit performs the following operations:

[0065] The prediction interval of the flow sensor data is established through the LSTM model. When the real-time data deviates from the predicted value by ±15%, the redundant sensor group cross-validation is started;

[0066] The LSTM model input layer normalization uses the min-max method (the water level data normalization range is 0-10m, the flow rate is 0-50m 3 / s);

[0067] If the pH value in the water quality sensor conflicts with the conductivity data, the association rules in the historical database (such as the pH-turbidity correlation model) are called for logical verification.

[0068] The deployment method of the multi-source sensor module includes:

[0069] The water level sensor adopts radar and pressure dual-mode deployment, and the flow sensor adopts ultrasonic Doppler array and electromagnetic complementary deployment;

[0070] The vibration sensor is installed at the key nodes of the pipeline in the form of a three-axis accelerometer. The sampling frequency changes linearly with the water flow rate. The vibration sensor must be installed at a 45° angle to the pipeline axis (tolerance ±2°) to capture both axial and radial vibrations.

[0071] The present invention also provides a method for collecting high-precision telemetry terminal data based on multi-source sensor fusion, comprising the following steps:

[0072] S1: Water level, flow, water quality parameters and environmental data are collected synchronously through the multi-source sensor module. The collection frequency is dynamically adjusted by the dynamic priority scheduling module according to the hydrological characteristics;

[0073] The specific implementation of dynamic priority scheduling is as follows: when the Central Meteorological Observatory issues a yellow warning for heavy rain, the sampling frequency of the water level sensor is increased from 1Hz to 10Hz;

[0074] S2: Perform spatiotemporal alignment on the raw data, unify the time base using GPS / Beidou differential signals, and compensate for differences in sensor placement using a spatial interpolation algorithm.

[0075] S3: Adopting the improved DS evidence theory to fuse multi-source data, the entropy weight method is introduced to modify the basic probability distribution function for the conflicting evidence set;

[0076] S4: Anomaly detection is performed through the edge computing unit. If the traffic data mutation exceeds the threshold, the dual verification process of the ultrasonic and electromagnetic sensors is initiated;

[0077] S5: Compress the fused data into JSON-LD format and upload it to the cloud platform through the multi-channel communication module, while storing a copy of the original data with a digital signature locally.

[0078] The specific process of the improved DS evidence theory in S3 includes:

[0079] Calculate the confidence interval and uncertainty of each sensor evidence body. The confidence level Bel(A) of the water quality sensor is expressed as:

[0080]

[0081] Where m(B) is the evidence quality function;

[0082] The Josselme distance metric is used for conflicting evidence, and when the distance is ≥ 0.5, the entropy weight method is triggered to redistribute the weights.

[0083] The specific method of anomaly detection in S4 is:

[0084] Construct the flow-water level joint probability density function f(Q, H), and mark the real-time data as abnormal when it falls outside the 95% confidence region;

[0085] The specific methods of anomaly detection are:

[0086] The reference value k of the sliding window CUSUM algorithm is 2% of the sensor range, and the decision threshold h = 5σ (σ is the standard deviation of historical data);

[0087] Supplementary note: For pH sensors, three-point verification is required before each test using pH = 4.01 / 7.01 / 9.21 standard solutions;

[0088] A sliding window CUSUM algorithm is used to detect drift faults in pH sensors, and the window size is adaptively adjusted according to the rate of change of water quality.

[0089] It also includes data backtracking verification steps:

[0090] Establish a blockchain-based sensor data evidence chain in the cloud, and generate a Merkle tree hash value for each collection record;

[0091] In this embodiment, the Merkle tree is constructed using the SHA-256 algorithm (the number of leaf nodes is 2 n , n≥8), the block generation interval of the blockchain is set to 10 minutes;

[0092] When data disputes occur, the smart contract is called to compare the local storage on the terminal with the cloud hash value to verify data integrity.

[0093] The calibration method of the multi-source sensor module includes:

[0094] An online self-calibration is performed once every 24 hours, and the self-cleaning electrode of the water quality sensor is activated by injecting a standard solution;

[0095] The standard solution injection flow rate is controlled at 0.5 mL / s±5% (achieved by a micro peristaltic pump), and the bristle pressure for cleaning the electrode is 0.2 N;

[0096] The water level sensor uses a laser range finder as a temporary reference datum, and an alarm is triggered when the calibration error exceeds 2%.

[0097] Compared with the related art, the high-precision telemetry terminal data acquisition system and method based on multi-source sensor fusion provided by the application has the following beneficial effects:

[0098] The high-precision telemetry terminal data acquisition system and method based on multi-source sensor fusion provided by the application can reduce water level measurement errors and thus improve the confidence of flow data through multi-source sensor space-time alignment and adaptive weighted fusion, can reduce the continuous operation failure rate of the equipment in the flood season through dynamic priority scheduling and edge computing redundancy checking, can automatically switch to the Beidou satellite when the 4G signal is lost, with a time consumption of less than 3 seconds, thus ensuring data real-time performance, and can reduce the frequency of manual calibration from once a week to once a quarter through the online self-calibration function, thus being able to save annual operation and maintenance costs.

[0099] The above description is only an embodiment of the application, and does not limit the patent scope of the application, and any equivalent structure or equivalent process transformation using the content of the specification and drawings, or direct or indirect application in other related technical fields, is also included in the patent protection scope of the application.

Claims

1. A high-precision telemetry terminal data acquisition system based on multi-source sensor fusion, characterized in that: include: Multi-source sensor module: consists of a water level sensor, flow sensor, water quality sensor, meteorological sensor and vibration sensor. Each sensor is connected to the main control unit through a standardized interface. Data fusion processing unit: Built-in Kalman filter algorithm, adaptive weighted fusion algorithm and spatiotemporal alignment module, used to synchronize timestamps, calibrate spatial positions and compensate errors for heterogeneous data from multi-source sensors; Dynamic priority scheduling module: Dynamically allocates sampling frequency based on sensor type and environmental status. Water level and flow sensors are prioritized to millisecond-level sampling during flood season, while water quality sensors use minute-level sampling in steady state. Edge computing unit: Integrates an LSTM neural network model to predict sensor data anomalies in real time and trigger redundancy verification mechanisms; Multi-channel communication module: supports multi-mode switching of 4G / 5G, LoRa, and Beidou satellite communications, and automatically selects the optimal transmission path according to network conditions; Anti-interference power management module: adopts solar-supercapacitor dual power supply system, built-in pulse current suppression circuit and sensor isolation power supply unit.

2. The high-precision telemetry terminal data acquisition system based on multi-source sensor fusion according to claim 1 is characterized in that: The data fusion processing unit further includes: The time-space alignment module unifies the time-space benchmarks of sensor data through GPS timing and Beidou differential positioning, with a time synchronization accuracy of ≤1ms and a spatial positioning error of ≤0.1m. The adaptive weighted fusion algorithm dynamically adjusts the weight according to the historical credibility of the sensor. The weight calculation formula of the dissolved oxygen data of the water quality sensor is: in, is the sensor variance, t is the time interval since the last calibration, and λ is the attenuation coefficient.

3. The high-precision telemetry terminal data acquisition system based on multi-source sensor fusion according to claim 1 is characterized in that: The edge computing unit performs the following operations: The prediction interval of the flow sensor data is established through the LSTM model. When the real-time data deviates from the predicted value by ±15%, the redundant sensor group cross-validation is started; If the pH value in the water quality sensor conflicts with the conductivity data, the association rules in the historical database are called for logical verification.

4. The high-precision telemetry terminal data acquisition system based on multi-source sensor fusion according to claim 1 is characterized in that: The deployment method of the multi-source sensor module includes: The water level sensor adopts radar and pressure dual-mode deployment, and the flow sensor adopts ultrasonic Doppler array and electromagnetic complementary deployment; The vibration sensor is installed at the key nodes of the pipeline in the form of a triaxial accelerometer, and the sampling frequency varies linearly with the water flow rate.

5. A data acquisition method using a high-precision telemetry terminal data acquisition system based on multi-source sensor fusion as described in any one of claims 1 to 4, characterized in that: The following steps are involved: S1: Water level, flow, water quality parameters and environmental data are collected synchronously through the multi-source sensor module. The collection frequency is dynamically adjusted by the dynamic priority scheduling module according to the hydrological characteristics; S2: Perform spatiotemporal alignment on the raw data, unify the time base using GPS / Beidou differential signals, and compensate for differences in sensor placement using a spatial interpolation algorithm. S3: Adopting the improved DS evidence theory to fuse multi-source data, the entropy weight method is introduced to modify the basic probability distribution function for the conflicting evidence set; S4: Anomaly detection is performed through the edge computing unit. If the traffic data mutation exceeds the threshold, the dual verification process of the ultrasonic and electromagnetic sensors is initiated; S5: Compress the fused data into JSON-LD format and upload it to the cloud platform through the multi-channel communication module, while storing a copy of the original data with a digital signature locally.

6. The high-precision telemetry terminal data acquisition method based on multi-source sensor fusion according to claim 5 is characterized in that: The specific process of the improved DS evidence theory in S3 includes: Calculate the confidence interval and uncertainty of each sensor evidence body. The confidence level Bel(A) of the water quality sensor is expressed as: Where m(B) is the evidence quality function; The Josselme distance metric is used for conflicting evidence, and when the distance is ≥ 0.5, the entropy weight method is triggered to redistribute the weights.

7. The high-precision telemetry terminal data acquisition system and method based on multi-source sensor fusion according to claim 5 is characterized in that: The specific method of anomaly detection in S4 is: Construct the flow-water level joint probability density function f(Q, H), and mark the real-time data as abnormal when it falls outside the 95% confidence region; A sliding window CUSUM algorithm is used to detect drift faults in pH sensors, and the window size is adaptively adjusted according to the rate of change of water quality.

8. The high-precision telemetry terminal data acquisition system and method based on multi-source sensor fusion according to claim 5 is characterized in that: It also includes data backtracking verification steps: Establish a blockchain-based sensor data evidence chain in the cloud, and generate a Merkle tree hash value for each collection record; When data disputes occur, the smart contract is called to compare the local storage on the terminal with the cloud hash value to verify data integrity.

9. The high-precision telemetry terminal data acquisition system and method based on multi-source sensor fusion according to claim 5 is characterized in that: The calibration method of the multi-source sensor module includes: Perform online self-calibration every 24 hours to activate the self-cleaning electrode of the water quality sensor by injecting standard solution; The water level sensor uses a laser rangefinder as a temporary reference benchmark, and an alarm is triggered when the calibration error exceeds 2%.

Images