Multi-source information fusion underground tunneling inertial navigation system and method

Abstract

The invention discloses an underground tunneling inertial navigation system and method for multi-source information fusion, and relates to the technical field of underground tunneling navigation, and the system comprises a plurality of functional units such as a shape memory alloy self-calibration inertial module and a magnetorheological fluid intelligent vibration reduction platform. Multi-source data fusion processing is realized through technical means of hardware self-calibration, full-band vibration suppression, dynamic weight distribution and the like. In the aspect of inertial navigation, the eight-hour positioning error is smaller than 5 cm and is remarkably improved compared with a traditional system; the positioning precision under complex working conditions reaches + / -10cm / 100m, which is obviously improved compared with a fixed weight fusion method. And the delay time is less than 50ms during multi-machine collaborative operation, and the supporting efficiency is improved by 35%. Meanwhile, geological anomalies can be responded in real time, the fault passing efficiency is improved by 60%, the coal rock cutting error rate is reduced by 70%, and navigation continuity is guaranteed when the sensor fails. According to the invention, the problem of navigation of underground tunneling is effectively solved, and the safety, high efficiency and intelligent level of underground tunneling operation are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of underground tunneling navigation, and specifically to an underground tunneling inertial navigation system and method for multi-source information fusion. Background Art

[0002] In underground tunneling operations, accurate navigation and positioning are crucial for ensuring construction safety, improving tunneling efficiency, and guaranteeing project quality. Traditional underground navigation systems mainly rely on single sensors or simple combinations, and have many limitations in complex underground environments. For example, although the inertial navigation system can work independently, its positioning error will continuously accumulate over time; the measurement accuracy of lidar is severely affected in harsh environments such as dust and water vapor; the performance of vision sensors drops significantly under low light conditions. Moreover, existing systems are difficult to perceive geological structure changes in real time, unable to adjust the tunneling path in a timely manner, easily leading to over-excavation and under-excavation phenomena, resulting in resource waste and safety hazards. In addition, the communication delay is large and the synchronization accuracy is low during multi-machine collaborative operations, severely restricting the intelligent development of underground construction. Therefore, it is urgent to develop an underground tunneling inertial navigation system and method for multi-source information fusion that can adapt to complex underground environments, with high precision and intelligence.

[0003] In view of this, the present application is specifically proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide an underground tunneling inertial navigation system and method for multi-source information fusion to solve the problems mentioned in the above background art.

[0005] To solve the above technical problems, an underground tunneling inertial navigation system for multi-source information fusion provided by the present invention includes: Shape memory alloy self-calibration inertial module: integrated with a nickel-titanium alloy cantilever beam, triggering mechanical deformation recovery through a heating power of 1.5W to correct the zero-offset errors of the MEMS accelerometer (accuracy ±0.1g) and gyroscope (angle random walk 0.01 degrees per square root hour). The self-calibration time is ≤80 ms / time, the operating temperature range is -40°C to 60°C, and the pressure resistance is ≥50 MPa; through hardware-level self-calibration technology, the accelerometer zero drift is reduced from 0.2g to 0.03g, and the gyroscope zero drift is reduced from 0.5° / h to 0.08° / h, suppressing the accumulation of inertial device errors from the physical level, achieving a positioning error of <5 cm in 8 hours, and the improvement amplitude compared with the traditional system far exceeds the original set value; Magnetorheological Fluid Intelligent Vibration Damping Platform: It includes 3 Lord RD-1005-3 type shock absorbers filled with MRF-132DG magnetorheological fluid, with a response time ≤ 0.5 ms, the viscosity is adjusted by a current ranging from 0 A to 1 A (both 0 A and 1 A are included), the vibration attenuation rate ≥ 92% (20 - 300 Hz), the inertial navigation output noise density is reduced to ≤ 5 micrograms per square root hertz, and the stress is measured in real time to dynamically adjust the vibration damping parameters; The full-frequency vibration suppression technology combines stress measurement, reducing the noise density by 90%, increasing the stability of inertial data output by 5 times, and at the same time optimizing the vibration damping effect through stress feedback to ensure the navigation accuracy in complex vibration environments; Variational Bayesian Network Processing Unit: It fuses 100 Hz inertial navigation data, 10 Hz lidar point cloud, 50 Hz ultrasonic radar ranging, and 30 fps visual images in real time, dynamically allocates sensor weights through posterior probability, and when the dust concentration > 1500 mg / m³, the weight of the ultrasonic radar ≥ 0.7; Based on the dynamic weight allocation algorithm of environmental parameters, it automatically switches the optimal data source under complex working conditions such as dust and low light, and the positioning accuracy is significantly improved compared with the fixed weight fusion method, achieving a dynamic positioning accuracy of ±10 cm / 100 m; 5G URLLC Communication Module: It supports the 2.6 GHz frequency band, with a communication delay < 8 ms, a reliability ≥ 99.999%, realizes multi-machine clock synchronization through the time-sensitive network, with a deviation < 500 ns, and the single-device bandwidth ≥ 50 Mbps; The ultra-low latency communication and high-precision clock synchronization technology enable the lag time of multi-machine collaborative operation < 50 ms, the bolt positioning accuracy reaches ±10 cm, and the support efficiency is increased by 35%, realizing the precise coordination of "heading - support" parallel operation; Digital Twin Geological Modeling Unit: Based on lidar with an accuracy of ±3 cm and ground penetrating radar with a detection depth of 15 m, it constructs a three-dimensional point cloud model, and combines reinforcement learning to output cutting path adjustment instructions with a step size of 0.5°, and a response time < 150 ms; The real-time geological modeling and path optimization technology enable the response time of the cutting head angle adjustment < 180 ms, the fault crossing efficiency is increased by 60%, the coal and rock mis-cutting rate is reduced by 70%, and the coal ash content is reduced by 4% compared with the traditional scheme; Conditional Generative Adversarial Network cGAN Data Filling Module: It inputs timestamps and known sensor data, generates inertial navigation data including acceleration and angular velocity, and the training data volume ≥ 1.2×10 6 , and the mean square error (MSE) of the generated data ≤ 0.01 g 2 ; The data-driven missing data filling technology generates inertial navigation data with an accuracy of ±0.01 g when the sensor fails, maintaining the system positioning error within ±10 cm / 100 m, which is significantly improved compared with the traditional interpolation method, ensuring the navigation continuity under complex working conditions; Tactile Sensing Unit: It integrates 128 piezoelectric film tactile sensors with a resolution of 0.1 N / mm2 , the force is measured in real time to sense the distribution of cutting resistance, and the load of the cutting head is evaluated by measuring the torque; the high-precision tactile perception technology monitors the cutting resistance and torque in real time, providing direct load feedback for the optimization of the cutting path. The number of times of mis-cutting the roof and floor in thin coal seam cutting is reduced from 5 times per hour to 0.5 times per hour, improving the purity of coal quality.

[0006] Furthermore, in the shape memory alloy self-calibrating inertial module, the surface of the cantilever beam is coated with an alumina thermal insulation layer with a thickness of 50μm and a thermal conductivity of 0.15W / (m·K), and is connected to the edge computing node through a cable with a temperature resistance of 80℃, and the computing power ≥ 200TOPS; the design of the alumina thermal insulation layer and the high-temperature resistant cable enables the inertial module to work stably in a high-temperature environment of 60℃, broadening the system's environmental adaptability range, reducing the influence of high temperature on the accuracy of inertial devices, and ensuring reliable operation in high-humidity and high-temperature environments such as deep mines.

[0007] Furthermore, in the magnetorheological fluid intelligent vibration damping platform, it is externally wrapped with an aerogel thermal insulation layer with a thermal conductivity of 0.013W / (m·K), vibration data at 1000Hz is collected in real time, the current of the shock absorber is dynamically adjusted through the PWM module to achieve full-frequency vibration suppression, and at the same time, the fluid pressure is measured to optimize the working parameters of the magnetorheological fluid; the combination of the aerogel thermal insulation layer and the measurement of fluid pressure ensures that the shock absorber can achieve a vibration attenuation rate of more than 92% within the temperature range of -40℃ to 60℃, improving the robustness of the system in extreme temperatures and extending the service life of the equipment.

[0008] Furthermore, in the variational Bayesian network processing unit, a built-in fault detection module automatically switches to the fusion positioning of the ultrasonic radar and the magnetic gradient tensor instrument when the number of valid point clouds of the lidar is <5000 for 5 consecutive frames, and the positioning accuracy is ±10cm / 100m; the intelligent fault detection and redundant positioning technology automatically switches to the fusion mode of the ultrasonic radar + magnetic gradient tensor instrument when the lidar fails due to dust interference, and the positioning accuracy still remains ±10cm / 100m, improving the fault tolerance ability by 90% compared with the problem of the error explosion growth after the failure of traditional single sensors.

[0009] Furthermore, in the digital twin geological modeling unit, a geological radar with a resolution of 0.1m is integrated, the accuracy of identifying the dielectric constant difference of the coal-rock interface is ±0.5, combined with a convolutional neural network to detect faults, and the identification accuracy is ≥95%, and the cutting energy consumption is evaluated by measuring the mechanical power; the high-precision geological identification technology combined with the measurement of mechanical power warns of fault activities 2 hours in advance, and at the same time optimizes the cutting power distribution, reducing the equipment energy consumption by 30% and improving the tunneling efficiency and safety.

[0010] Furthermore, in the conditional generative adversarial network (cGAN) data filling module, a 5-layer fully connected generator and a 4-layer discriminator network are adopted. The loss function is equal to the Wasserstein distance plus gradient penalty, and the peak signal-to-noise ratio (PSNR) of the generated data is ≥ 32 dB. The advanced loss function and deep network architecture improve the generalization ability of the cGAN model by 40%, enabling high-precision data filling under 20 complex working conditions and ensuring the reliability of the system in extreme environments.

[0011] Furthermore, the tactile perception unit is linked with the cutting head manipulator of the tunneling equipment. By measuring force and torque, the cutting head rotation speed and feed speed are dynamically adjusted, and the mechanical efficiency is synchronously measured to optimize the cutting parameters. The load sensing and parameter adaptive technology improve the load uniformity of the cutting head by 40% and the mechanical efficiency by 25%, reduce tool wear, and extend the equipment maintenance cycle.

[0012] A usage method of an underground tunneling inertial navigation system with multi-source information fusion includes the following steps: Data acquisition: Data is collected through a shape memory alloy self-calibrating inertial module (100 Hz), lidar (10 Hz), ground penetrating radar (once every 5 seconds), and tactile sensors (5 ms response), including position, attitude, geological structure, and cutting resistance information. The multi-source high-frequency data acquisition technology ensures the real-time acquisition of environmental information and equipment status, providing comprehensive data support for intelligent decision-making.

[0013] Preprocessing: Magnetorheological fluid damping and adaptive notch filtering are performed on the inertial data, and the noise density is reduced to less than or equal to 5 micrograms per square root hertz. Force-electric signal conversion is performed on the tactile data to measure force, with a resolution of 0.1 N / mm². The multi-stage filtering and signal conversion technology improve the data quality, reduce the inertial data noise density by 90%, and shorten the tactile signal response time to 5 ms, laying a foundation for subsequent fusion algorithms.

[0014] Dynamic error suppression: The SMA heating circuit is triggered at 1.5 W, 10 ms per time for hardware self-calibration. Combining with the variational Bayesian network to online evaluate the random walk parameters of inertial devices, the cumulative error of the inertial navigation is calibrated by ±10 cm every 10 seconds. The combination of hardware self-calibration and algorithm compensation corrects the cumulative error of the inertial navigation in real time, avoids the infinite growth of errors over time, and achieves a positioning error of < 5 cm in 8 hours, meeting the millimeter-level accuracy requirements of unmanned tunneling.

[0015] Multi-source fusion: The variational Bayesian network dynamically allocates sensor weights according to environmental parameters (dust concentration, illuminance). In a dusty environment, ultrasonic radar is preferred with a weight ≥ 0.7, and in a normal environment, lidar is preferred with a weight ≥ 0.6. The environment-adaptive fusion algorithm automatically selects the optimal data source under complex working conditions, significantly improving the positioning accuracy, reducing manual intervention, and enhancing the system robustness.

[0016] Multi - machine collaboration: Transmit multi - machine position information (accuracy ±5 cm) through a 5G URLLC network (latency <8 ms), optimize the actions of support robots using a federated learning - blockchain architecture, and achieve action synchronization (deviation <1 μs) with a time - sensitive network; low - latency communication and distributed optimization technologies enable multi - machine sub - meter - level collaborative operations, increasing the support efficiency by 30% and promoting the unmanned and intelligent underground construction.

[0017] Geological adaptability: The digital twin model updates the three - dimensional geological structure in real - time. When encountering a fault with a dielectric constant mutation >5, reinforcement learning generates an obstacle - avoidance path, the step size of the cutting head angle adjustment is 0.5°, and the response time <180 ms; the combination of digital twin and reinforcement learning realizes the automatic identification and rapid response to geological anomalies, increasing the efficiency of crossing faults by 60% and reducing material waste caused by over - excavation and under - excavation by 30%.

[0018] Data filling: When the sensor fails, the conditional generative adversarial network cGAN data filling module inputs the timestamp and known data to generate 60 - second continuous inertial navigation data, and the positioning error after filling ≤±10 cm / 100 m; the data filling technology based on deep learning maintains the continuous operation of the system when the sensor fails, significantly improving compared with the traditional interpolation method and ensuring the continuity of operations.

[0019] Energy efficiency optimization: Dynamically adjust the equipment operation parameters by measuring mechanical power and fluid pressure, reducing the system energy consumption by more than 30%; the energy efficiency optimization technology combines real - time power and pressure monitoring to achieve precise matching of equipment load and energy consumption, reducing energy waste and lowering the underground operation cost.

[0020] Furthermore, in the data acquisition, the tactile sensor covers a distance of 10 meters from the lidar, the lidar point cloud density ≥0.1 point / m², the geological radar detection depth ≥15 m, and the tactile sensor covers ≥80% of the cutting head surface area; high - density data acquisition and comprehensive tactile coverage ensure the precise perception of geological structure and cutting load, providing a reliable basis for optimizing the cutting path.

[0021] Furthermore, in the pre - processing, the adaptive notch filter algorithm suppresses 50Hz power frequency noise, and the magnetorheological fluid vibration damping platform attenuates vibration energy of 20 - 300Hz by ≥92%; targeted noise suppression technology improves the purity of inertial data and ensures the stability of the navigation system in a strong vibration and high electromagnetic interference environment.

[0022] Furthermore, in the dynamic error suppression, the update frequency of the variational Bayesian network is synchronized with the inertial navigation at 100 Hz, the gradient compression rate of the federated learning model is ≥80%, and the accuracy of the blockchain timestamp is 1 μs; high-frequency algorithm updates and efficient data transmission ensure real-time optimization of the error model, non-tamperable data archiving, and improved system error suppression efficiency and data credibility.

[0023] Furthermore, in the multi-source fusion step, in a low-light environment with illuminance <10 lux, the inertial navigation weight is automatically increased to ≥0.8, combined with magnetic gradient tensor instrument-assisted positioning with an accuracy of ±10 cm / 100 m; multi-sensor redundant fusion technology maintains high-precision positioning in low-light environments, broadens the applicable scenarios of the system, and improves the adaptability to complex environments.

[0024] Furthermore, in the multi-machine collaboration, the uplink throughput of the 5G URLLC network is ≥1.2 Gbps, supporting ≥20 devices to be online simultaneously, the bolting robot has an anchor positioning accuracy of ±10 cm, and the latency is <50 ms; high-bandwidth communication and multi-device collaboration technology enable real-time linkage of large-scale devices, improve the efficiency and safety of underground operations, and promote the upgrade of intelligent construction.

[0025] Furthermore, in the geological adaptation, the digital twin model updates the roadway point cloud every 2 seconds, and the reinforcement learning algorithm takes the coal-rock mis-cutting rate (deduction of 10 points for every 1 cm of mis-cutting) and path smoothness (curvature >10 m -1 deduction of 5 points) as the optimization objectives; high-frequency model updates and refined optimization objectives ensure the real-time and accuracy of the cutting path, reduce the coal-rock mis-cutting rate by 70%, and improve the quality of coal mining.

[0026] Furthermore, in the data filling, the training data of the conditional generative adversarial network cGAN data filling module covers ≥20 working conditions (dust, faults, vibrations, etc.), and the peak signal-to-noise ratio between the generated data and the real data is ≥32 dB, and the mean square error is ≤0.008 g 2 ; large-scale training data and strict performance indicators enable the cGAN model to achieve high-precision data filling under extreme working conditions, ensuring the reliability and robustness of the navigation system.

[0027] Compared with the prior art, the beneficial effects of the present invention are: 1. High-precision positioning and error suppression: With the help of a shape memory alloy self-calibrating inertial module, a heating power of 1.5 W is used to trigger the mechanical deformation recovery of the nickel-titanium alloy cantilever beam, which can quickly correct the zero-offset errors of the MEMS accelerometer and gyroscope. The self-calibration time is short, and the working temperature and pressure resistance range are wide, effectively suppressing the error accumulation of inertial devices from a physical level. The positioning error within 8 hours is less than 5 cm, and the improvement amplitude compared with the traditional system far exceeds the original set value, providing a high-precision and long-term stable positioning basis for underground tunneling operations.

[0028] 2. Enhanced adaptability to vibration environment: The magnetorheological fluid intelligent vibration reduction platform adopts 3 Lord RD-1005-3 shock absorbers filled with MRF-132DG magnetorheological fluid, which has a rapid response and can adjust viscosity through current. The vibration attenuation rate in the frequency band of 20 - 300Hz exceeds 92%. By combining stress measurement to dynamically adjust the vibration reduction parameters, the inertial navigation output noise density is reduced by 90%, and the stability of inertial data output is increased by 5 times, greatly ensuring the navigation accuracy in complex vibration environments and ensuring the reliable operation of the system under harsh working conditions.

[0029] 3. Precise positioning under complex working conditions: The variational Bayesian network processing unit can fuse multi-source data in real time and dynamically allocate sensor weights according to environmental parameters (such as dust concentration, illuminance). When the dust concentration is greater than 1500mg / m 3 ³, the weight of the ultrasonic radar ≥ 0.7; in a normal environment, the weight of the lidar ≥ 0.6. This technology automatically switches the optimal data source under complex working conditions, and the positioning accuracy is improved more significantly than the fixed-weight fusion method, achieving a dynamic positioning accuracy of ±10cm / 100m, effectively solving the positioning problem in the complex underground environment.

[0030] 4. Efficient multi-machine collaborative operation: The 5G URLLC communication module supports the 2.6GHz frequency band, with a communication delay of less than 8ms and a reliability of up to 99.999%. It realizes multi-machine clock synchronization (deviation less than 500ns) through the time-sensitive network. The bandwidth of a single device is sufficient, making the lag time of multi-machine collaborative operation < 50ms, the bolt positioning accuracy reaching ±10cm, and the support efficiency increasing by 35%, effectively promoting the precise coordination of "heading - support" parallel operation and improving the underground construction efficiency and safety.

[0031] 5. Geological adaptability and efficient tunneling: The digital twin geological modeling unit constructs a three-dimensional point cloud model based on lidar and ground-penetrating radar, and outputs cutting path adjustment instructions by combining reinforcement learning. The step size is accurate to 0.5°, and the response time is less than 150ms, which can quickly respond to geological anomalies. In practical applications, the fault-crossing efficiency is increased by 60%, the coal-rock mis-cutting rate is reduced by 70%, and the coal ash content is reduced by 4% compared with the traditional scheme, significantly improving the tunneling efficiency, reducing resource waste, and improving the quality of coal mining.

[0032] 6. Ensure navigation continuity: The conditional generative adversarial network cGAN data filling module generates inertial navigation data by inputting the time stamp and known sensor data when the sensor fails. The training data volume is large and the generated data accuracy is high, with the mean square error ≤ 0.01g². The system positioning error is maintained at ±10cm / 100m, which is significantly improved compared with the traditional interpolation method, ensuring navigation continuity under complex working conditions and avoiding operation interruption caused by sensor failures.

[0033] 7. Optimization of cutting parameters and improvement of coal quality: The tactile perception unit integrates 128 high-resolution piezoelectric film tactile sensors to measure the cutting resistance distribution in real time and evaluate the load of the cutting head. It is linked with the cutting head manipulator of the tunneling equipment to dynamically adjust the cutting head speed and feed speed, optimizing the cutting parameters. The number of times of mis-cutting the roof and floor in thin coal seam cutting is reduced from 5 times per hour to 0.5 times per hour, improving the purity of coal quality, reducing equipment wear, and extending the equipment maintenance cycle. Description of the Drawings

[0034] Figure 1 It is a principle block diagram of an underground tunneling inertial navigation system with multi-source information fusion; Figure 2 It is a flowchart of the usage method of an underground tunneling inertial navigation system with multi-source information fusion. Specific Embodiments

[0035] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0036] Please refer to Figure 1-2 , the present invention provides a technical solution: An underground tunneling inertial navigation system and method with multi-source information fusion, including: I. System hardware construction (taking a certain roadheader used in underground tunneling as an example): 1. Configuration of the inertial navigation module: Shape Memory Alloy (SMA) Self-Calibrating Inertial Measurement Unit (IMU): The IMU is installed on top of the roadheader cab and includes a nickel-titanium alloy cantilever beam (phase transition temperature 40°C, geometric dimensions 50 mm × 2 mm × 0.5 mm, resistance value 1.2 Ω), with an alumina thermal insulation layer plated on its surface, thickness 50 μm, thermal conductivity 0.15 W / (m·K), to isolate environmental temperature interference. MEMS inertial devices are integrated at both ends of the cantilever beam: the accelerometer selects ADXL345, range ±16 g, noise density 3.5 mg per square root hertz, zero drift coefficient <0.05 g / °C, meeting the measurement accuracy of ±0.1 g; the gyroscope selects MPU-6050, range ±2000° / s, angular random walk 0.01° per square root hour, zero drift <0.1° / h. It is connected to the edge computing node (ECU, model NVIDIA Jetson AGX Orin, computing power 200 TOPS) through a shielded cable (characteristic impedance 50 Ω) with a temperature resistance of 80°C. The SMA heating circuit is powered by the 5V power supply of the ECU, heating power 1.5W (voltage 12V, current 1.25A), single self-calibration period 80 ms (including cooling time), adapting to a temperature range of -40°C to 60°C and a pressure-resistant environment of 50 MPa.

[0037] Example: When the IMU detects that the temperature rises to 60°C, the SMA heating circuit automatically triggers calibration, heats up to the phase transition temperature within 10 ms, and after 80 ms, the zero drift correction is completed. The zero drift of the accelerometer drops from 0.2 g to 0.03 g, and the zero drift of the gyroscope drops from 0.5° / h to 0.08° / h.

[0038] Magnetorheological Fluid (MRF) Intelligent Vibration Damping Platform: Integrate 3 Lord RD-1005-3 type shock absorbers between the IMU and the roadheader frame, filled with MRF-132DG magnetorheological fluid (density 2.8 g / cm³, yield strength 100 kPa at 1A current), and externally wrapped with an aerogel thermal insulation layer, thickness 10 mm, thermal conductivity 0.013 W / (m·K). The response time of the shock absorber ≤0.5 ms, adjusts the viscosity through a current ranging from 0A to 1A, including 0A and 1A, achieving a vibration attenuation rate of 92% in the 20 - 300 Hz frequency band, suppressing the inertial navigation output noise density from 50 μg per square root hertz to 5 μg per square root hertz (100 Hz bandwidth). The ECU real-time collects 1000 Hz vibration data through the PWM module (frequency 10 kHz) and dynamically adjusts the shock absorption parameters.

[0039] Example: When the vibration amplitude of the cutting arm reaches 5 g, the ECU adjusts the current of the shock absorber to 1A, the viscosity of the MRF rises to 10 Pa·s, the 200 Hz vibration component decays by 95%, and the inertial navigation output noise density drops suddenly from 50 μg per square root hertz to 5 μg per square root hertz, ensuring the stability of inertial data.

[0040] 2. Multi-sensor integration solution: Environmental perception sensor group: Vision and ranging unit: A 16-line lidar (Velodyne VLP-16) is installed at the front end, with a scanning frequency of 10 Hz, a ranging accuracy of ±3 cm, and a point cloud density of 0.1 points / m² at a distance of 10 m; a panoramic camera (Basler acA2500-14gm) with a resolution of 2592×1944, configured with an 850 nm near-infrared filter, and a frame rate of 30 fps; an ultrasonic phased array radar (operating frequency 40 kHz), with a scanning range of 360°×120°, a ranging accuracy of ±3 cm, and suitable for a dust environment of 2000 mg / m³.

[0041] Geological detection unit: A MALA ProEx geological radar is integrated into the cutting arm, with a center frequency of 200 MHz, a detection depth of 15 m, a layer thickness resolution of 0.1 m, and a dielectric constant measurement accuracy of ±0.5; 128-channel Tekscan piezoelectric film tactile sensors are deployed on the surface of the cutting head, with a resolution of 0.1 N / mm² and a response time of 5 ms, to real-time sense the distribution of cutting resistance.

[0042] Example: When the cutting head touches sandy mudstone, the tactile sensor array detects an average resistance of 15 N / mm 2 , and the ECU adjusts the rotational speed of the cutting head to 1500 rpm accordingly, with an efficiency improvement of 40% compared to the traditional empirical adjustment.

[0043] II. System software architecture and algorithm implementation: 1. Variational Bayesian network (VBN) dynamic fusion algorithm: The VBN processing unit receives IMU data at a frequency of 100 Hz, lidar point cloud at 10 Hz, ultrasonic radar ranging at 50 Hz, and visual images at 30 fps in real time, and realizes dynamic allocation of sensor weights through the following steps: 1.1 Noise modeling: Establish a random walk model for the zero-bias drift of inertial devices and a Gaussian mixture model for the ranging error of lidar; 1.2 Weight calculation: Estimate the dust concentration based on the echo intensity of the lidar. When >1500 mg / m³, the weight of the ultrasonic radar is assigned 0.7, the weight of the inertial navigation is 0.2, and the weight of the vision is 0.1; in a normal environment (illuminance >10 lux, dust <500 mg / m³), the weight of the lidar is 0.6, the weight of the vision is 0.3, and the weight of the inertial navigation is 0.1; 1.3 Fault detection: When the number of valid point clouds of the lidar is <5000 for 5 consecutive frames (spot scattering rate >30%), automatically switch to the fusion positioning of the ultrasonic radar and the magnetic gradient tensor instrument, with a positioning accuracy of ±10 cm / 100 m.

[0044] Example: When entering a roadway with a dust concentration of 2000 mg / m³, VBN increases the weight of the ultrasonic radar to 0.7. After fusion, the positioning accuracy is improved from ±20 cm / 100 m of the traditional solution to ±10 cm / 100 m, effectively solving the positioning failure problem in a dusty environment.

[0045] 2. Digital Twin Geological Modeling and Reinforcement Learning Control: Three-dimensional Geological Modeling: The lidar point cloud constructs a global coordinate system through the ICP algorithm (initial accuracy ±5 cm). The geological radar scans the 10 m area ahead every 5 seconds to identify the coal-rock interface (dielectric constant difference >5) and faults (reflection coefficient >0.3), and updates the digital twin model in real time.

[0046] Path Optimization Algorithm: The PPO reinforcement learning algorithm is adopted. The state space includes the cutting head attitude (3-axis angles, accuracy ±0.1°), geological parameters (interface distance ±0.1 m, hardness ±5 MPa), and inertial navigation error (position deviation ±5 cm, angle deviation ±0.2°), a total of 12 dimensions; the action space is the adjustment of the cutting head angle (step size 0.5°). The coal-rock mis-cutting rate (deduct 10 points for every 1 cm of mis-cutting) and path smoothness (curvature >10 m -1 deduct 5 points) are used as the reward function, and the response time is <150 ms after training convergence.

[0047] Example: When the digital twin model identifies a normal fault (dip angle 60°, throw 1.5 m) 5 m ahead, the reinforcement learning algorithm outputs an adjustment instruction for the cutting head horizontal angle +3° and pitch angle -2°, and completes the angle correction within 180 ms, with the efficiency improved by 20 times compared with manual intervention.

[0048] III. Specific Implementation Steps: 1. System Initialization and Self-Calibration (0 - 5 minutes): 1.1 Hardware Self-Calibration: The ECU sends a 100 Hz square wave signal to the SMA heating circuit, triggers 3 times of heating (1.5 W, 10 ms / time), and completes the zero-bias correction of the IMU: the zero-drift of the accelerometer drops from 0.2 g to 0.03 g; the zero-drift of the gyroscope drops from 0.5° / h to 0.08° / h.

[0049] 1.2 Initial Environment Modeling: The lidar completes a 360° scan to generate an initial point cloud map of the roadway with 100,000 points, with a size of 5.2 m × 3.5 m; the geological radar scans 15 m ahead to identify a coal seam thickness of 1.2 m, and the roof lithology is sandy mudstone (density 2.3 g / cm 3 , dielectric constant 8).

[0050] Example: During the initialization process, the SMA self-calibration function completed 3 cycles within 5 minutes, and finally the zero-drift parameters of the IMU stabilized at 0.03g for the accelerometer and 0.08° / h for the gyroscope, providing high-precision initial data for subsequent navigation.

[0051] 2. Real-time navigation and error suppression (5 minutes - 7 hours 55 minutes): Data acquisition and preprocessing: The IMU outputs raw data at 100Hz. After being dampened by magnetorheological fluid, the peak vibration acceleration drops from ±5g to ±0.5g. Adaptive notch filtering is used to suppress 50Hz power frequency noise; the vision sensor detects fluorescent markers at 50m intervals through the YOLOv8 algorithm, and combines with lidar point clouds to calculate the absolute position, and calibrates the cumulative error of the inertial navigation every 10 seconds (average ±10cm).

[0052] Dynamic weight switching: When entering an area with a dust concentration of 2000mg / m³, the VBN automatically increases the weight of the ultrasonic radar to 0.7, and the positioning accuracy after fusion is ±10cm / 100m; in low-light environments (illuminance < 10lux), the weight of the inertial navigation is increased to 0.8, and an auxiliary positioning with an accuracy of ±10cm / 100m is achieved in combination with the magnetic gradient tensor instrument.

[0053] Example: In a low-light roadway (illuminance 5lux), the weight of the inertial navigation is increased to 0.8. Combining with the magnetic anomalies of the anchor bolts detected by the magnetic gradient tensor instrument, the system positioning accuracy is stabilized at ±10cm / 100m, which is 50% higher than the traditional single inertial navigation scheme.

[0054] 3. Geological anomaly response (taking a normal fault as an example, triggered at the 4th hour): Anomaly detection: The geological radar detected that the dielectric constant in front increased suddenly from 8 to 18 at 5m, combined with a sudden 25% increase in the resistance of the cutting head tactile sensor. The CNN algorithm determined it as a normal fault (dip angle 60°, throw 1.5m) with a 95% confidence level.

[0055] Control response: The digital twin model generates an obstacle avoidance path with a right deviation of 1.2m and a slope of +5°. The reinforcement learning outputs an instruction to adjust the angle of the cutting head (horizontal +3°, pitch -2°), and the action is completed within 180ms; instructions are sent to the support robot through the 5G network to reduce the bolt spacing in the fault area from 1.2m to 0.8m and increase the diameter of the cable bolts to 25mm, and strengthen the support within 5 minutes.

[0056] Example: During cutting in the fault area, after adjusting the angle of the cutting head, the mis-cutting rate of coal and rock dropped from 15% before adjustment to 3%, and the ash content of the coal quality dropped from 12% to 8%, improving the coal quality per single shift.

[0057] 4. Sensor failure tolerance (simulating a 60-second disconnection of the lidar): In the conditional generative adversarial network cGAN data filling module, after detecting the interruption of lidar data, the current timestamp, ultrasonic radar, and visual data are input to generate three-axis acceleration (accuracy ±0.01g) and angular velocity (accuracy ±0.1° / s). During the filling period, the heading angle error is maintained within ±0.3°, and the system positioning error is ±10cm / 100m, which is significantly improved compared with the traditional interpolation method.

[0058] Example: During the lidar disconnection, the MSE of the data generated by cGAN is 0.007g², which is significantly lower than 0.1g² of the traditional linear interpolation method, ensuring the stable operation of the navigation system when the sensor fails.

[0059] 5. Job end processing (7 hours 55 minutes - 8 hours): Zero-speed filtering is triggered after the roadheader has been stationary for 10 seconds to correct the cumulative positioning error to <5cm / 8 hours; Navigation data, geological models, and equipment status are stored on the blockchain with a timestamp accuracy of 1μs for subsequent fault tracing and model iteration.

[0060] As shown in Table 1 below:

[0061] Table 1: Comparison table of performance parameters between the technical solution of the present invention and the traditional technical solution.

[0062] In summary, in the present invention, through the hardware self-calibration of the shape memory alloy self-calibrating inertial module, the zero-bias error of the inertial device is corrected, error accumulation is suppressed, and high-precision long-term positioning is achieved. With the help of the magnetorheological fluid intelligent vibration damping platform for full-frequency vibration suppression and stress measurement, the output noise density of the inertial navigation is reduced, and the stability of inertial data and navigation accuracy are improved. The variational Bayesian network processing unit is used to dynamically allocate sensor weights, automatically select the optimal data source under complex working conditions, and improve the positioning accuracy. Relying on the low-latency communication and high-precision clock synchronization of the 5G URLLC communication module, multi-machine collaborative precise operation is realized, and the support efficiency is improved. The digital twin geological modeling unit is used for real-time geological modeling and path optimization, quickly responds to geological anomalies, reduces the coal and rock mis-cutting rate, and improves the coal quality. The cGAN data filling module is used to generate inertial navigation data when the sensor fails, maintain the system positioning accuracy, and ensure the continuity of navigation.

Claims

1. An underground tunneling inertial navigation system with multi-source information fusion, characterized by: include: Shape memory alloy self-calibration inertial module: integrated nickel-titanium alloy cantilever beam, triggers mechanical deformation recovery through 1.5W heating power, corrects the zero bias error of MEMS accelerometer and gyroscope, self-calibration time ≤80ms / time, operating temperature range -40℃~60℃, pressure resistance ≥50MPa; Magnetorheological fluid intelligent vibration reduction platform: contains 3 vibration dampers, filled with MRF-132DG magnetorheological fluid, response time ≤ 0.5ms, viscosity is adjusted by current ranging from 0A to 1A, where 0A and 1A are both included, vibration attenuation rate ≥ 92%, reduces the noise density of inertial navigation output to less than or equal to 5 micrograms per root of hertz, and measures stress in real time to dynamically adjust vibration reduction parameters; Variational Bayesian network processing unit: real-time fusion of 100Hz inertial navigation data, 10Hz lidar point cloud, 50Hz ultrasonic radar ranging and 30fps visual image, dynamic allocation of sensor weights through posterior probability, ultrasonic radar weight ≥ 0.7 when dust concentration > 1500mg / m³; 5G URLLC communication module: supports 2.6GHz frequency band, communication delay <8ms, multi-machine clock synchronization through time-sensitive network, deviation <500ns, single device bandwidth ≥50Mbps; Digital twin geological modeling unit: Based on laser radar, accuracy ±3cm, and geological radar, detection depth 15m, build a three-dimensional point cloud model, combine reinforcement learning to output cutting path adjustment instructions, step length 0.5°, response time <150ms; Conditional Generative Adversarial Network (cGAN) data filling module: input timestamp and known sensor data to generate inertial navigation data including acceleration and angular velocity. The amount of training data is ≥ 1.2×10 6 , generate data mean square error MSE≤0.01g 2 ; Tactile sensing unit: integrated 128-channel piezoelectric film tactile sensor, resolution 0.1N / mm 2 , measure the force in real time to sense the cutting resistance distribution, and evaluate the cutter head load by measuring the torque.

2. The underground tunneling inertial navigation system with multi-source information fusion as claimed in claim 1, characterized in that: In the shape memory alloy self-calibration inertial module, the surface of the cantilever beam is plated with an aluminum oxide insulation layer with a thickness of 50 μm and a thermal conductivity of 0.15 W / (m·K). It is connected to the edge computing node through a temperature-resistant cable of 80°C, and the computing power is ≥200TOPS.

3. The underground tunneling inertial navigation system with multi-source information fusion as claimed in claim 1, characterized in that: The magnetorheological fluid intelligent vibration reduction platform is wrapped with an aerogel insulation layer on the outside, with a thermal conductivity of 0.013W / (m·K). It collects 1000Hz vibration data in real time, dynamically adjusts the shock absorber current through the PWM module to achieve full-band vibration suppression, and measures the fluid pressure to optimize the magnetorheological fluid working parameters.

4. The underground tunneling inertial navigation system with multi-source information fusion as claimed in claim 1, characterized in that: The variational Bayesian network processing unit has a built-in fault detection module. When the number of valid point clouds of the laser radar for 5 consecutive frames is less than 5000, it automatically switches to the fusion positioning of the ultrasonic radar and the magnetic gradient tensor meter, with a positioning accuracy of ±10cm / 100m.

5. The underground tunneling inertial navigation system with multi-source information fusion as claimed in claim 1, characterized in that: In the digital twin geological modeling unit, a geological radar is integrated with a resolution of 0.1m and an accuracy of ±0.5 for identifying the difference in dielectric constants at the coal-rock interface. A convolutional neural network is combined to detect faults with an identification accuracy of ≥95%, and the cutting energy consumption is evaluated by measuring mechanical power.

6. The underground tunneling inertial navigation system with multi-source information fusion as claimed in claim 1, characterized in that: In the conditional generative adversarial network cGAN data filling module, a 5-layer fully connected generator and a 4-layer discriminator network are used, the loss function is equal to Wasserstein distance + gradient penalty, and the peak signal-to-noise ratio of the generated data is ≥32dB.

7. The underground tunneling inertial navigation system with multi-source information fusion as claimed in claim 1, characterized in that: The tactile sensing unit is linked to the cutting head mechanical arm of the tunneling equipment, dynamically adjusts the cutting head rotation speed and feed speed by measuring force and torque, and synchronously measures mechanical efficiency to optimize cutting parameters.

8. A method for using an underground tunneling inertial navigation system with multi-source information fusion, characterized in that: The following steps are involved: Data collection: Data is collected through shape memory alloy self-calibration inertial modules, laser radar, geological radar and tactile sensors, including position, posture, geological structure and cutting resistance information; Preprocessing: magnetorheological fluid vibration reduction and adaptive notch filtering are performed on the inertial data to reduce the noise density to less than or equal to 5 micrograms per root Hertz, and force-to-electric signal conversion is performed on the tactile data to measure the force with a resolution of 0.1N / mm²; Dynamic error suppression: trigger the SMA heating circuit, 1.5W, 10ms / time, perform hardware self-calibration, combine the variational Bayesian network to evaluate the random walk parameters of the inertial device online, and calibrate the inertial guidance cumulative error of ±10cm every 10 seconds; Multi-source fusion: The variational Bayesian network dynamically allocates sensor weights according to environmental parameters. Ultrasonic radar is preferred in dusty environments with a weight ≥ 0.7, while lidar is preferred in normal environments with a weight ≥ 0.

6. Multi-machine collaboration: The location information of multiple machines is transmitted through the 5G URLLC network, the federated learning-blockchain architecture is used to optimize the support robot movements, and the time-sensitive network is used to achieve movement synchronization with a deviation of <1μs; Geological self-adaptation: The digital twin model updates the three-dimensional geological structure in real time. When encountering a fault and the dielectric constant suddenly changes to more than 5, reinforcement learning generates an obstacle avoidance path, the cutting head angle is adjusted in steps of 0.5°, and the response time is less than 180ms. Data filling: When the sensor fails, the conditional generative adversarial network cGAN data filling module inputs the timestamp and known data to generate 60 seconds of continuous inertial navigation data. The positioning error after filling is ≤±10cm / 100m; Energy efficiency optimization: Dynamically adjust equipment operating parameters by measuring mechanical power and fluid pressure, reducing system energy consumption by more than 30%.

9. The method for using a multi-source information fusion underground tunneling inertial navigation system according to claim 8, characterized in that: In the data collection, the detection depth of the geological radar is ≥15m, the tactile sensor covers a distance of 10 meters from the laser radar, the laser radar point cloud density is ≥0.1 point / m², and the cutting head surface is ≥80% of the area.

10. The method for using the underground tunneling inertial navigation system with multi-source information fusion as claimed in claim 8, characterized in that: In the pre-processing, the adaptive notch filter algorithm suppresses 50 Hz power frequency noise, and the magnetorheological fluid vibration reduction platform attenuates 20-300 Hz vibration energy by ≥92%.

11. The method for using the underground tunneling inertial navigation system with multi-source information fusion as claimed in claim 8, characterized in that: In the dynamic error suppression, the update frequency of the variational Bayesian network is synchronized with the inertial navigation at 100 Hz, and the blockchain evidence timestamp accuracy is 1 μs.

12. The method for using the underground tunneling inertial navigation system with multi-source information fusion as claimed in claim 8, characterized in that: In the multi-source fusion, in a weak light environment with an illumination of <10lux, the lower inertial navigation weight is automatically increased to ≥0.8, and combined with the magnetic gradient tensor meter to assist in positioning, the accuracy is ±10cm / 100m.

13. The method for using the underground tunneling inertial navigation system with multi-source information fusion as claimed in claim 8, characterized in that: In the multi-machine collaboration, the uplink throughput of the 5G URLLC network is ≥1.2Gbps, supporting ≥20 devices to be online at the same time. The anchor positioning accuracy of the support robot is ±10cm, and the lag time is <50ms.

14. The method for using the underground tunneling inertial navigation system with multi-source information fusion as claimed in claim 8, characterized in that: In the geological adaptation, the digital twin model updates the tunnel point cloud every 2 seconds, and the reinforcement learning algorithm takes the coal-rock miscutting rate and path smoothness as optimization targets.

15. The method for using the underground tunneling inertial navigation system with multi-source information fusion as claimed in claim 8, characterized in that: In the data filling, the training data of the conditional generative adversarial network cGAN data filling module covers ≥20 working conditions, the peak signal-to-noise ratio of the generated data and the real data is ≥32dB, and the mean square error is ≤0.008g 2 .

Images