Silicon-based-carbon-based fusion biological intelligent control system and full-scene autonomous collaboration method

Abstract

The invention discloses a silicon-based-carbon-based fusion biological intelligent control system and a full-scene autonomous cooperation method, and belongs to the field of intelligent bionic robots. According to the system, a joint angle sensor, a visual camera and a bioelectric sensor are integrated through non-invasive wearable equipment, a multi-source fusion positioning technology is combined, sub-meter motion control is realized in a GPS rejection environment, and 200ms-level dynamic obstacle avoidance path planning is realized by adopting an improved RRT * algorithm. Non-intrusive control is implemented through piezoelectric ceramics and an ultrasonic instruction module, and a stress index is reduced in combination with a three-level biosafety threshold strategy. The energy module integrates piezoelectric power generation and flexible solar energy technologies, and long-time endurance is guaranteed. The method breaks through the limitation of the traditional technology in high-precision task scenes such as frontier defense, military reconnaissance, battlefield combat, forest and mountain patrol, river patrol, field exploration and disaster rescue, the passing rate of complex terrains is high, the task efficiency is improved, and a high-precision and low-invasion full-scene solution is provided for biological intelligent control.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent bionic robots, and particularly relates to a non-invasive intelligent control system based on the deep integration of silicon-based technology and living organisms. Further, it is to control the movement direction and speed of an animal by inputting the path to the target into the control system online and / or offline in combination with the timing and positioning data obtained by the positioning system carried by the animal, and to call the tools or controls carried by the animal online or offline, or to let the animal itself automatically complete specific tasks. Through multi-modal limb movement perception and multi-source heterogeneous positioning fusion technology (including joint angle sensors, radio frequency beacons, wireless WiFi positioning, etc.), sub-meter (<0.3m) movement tracking and precise control of living organisms in complex environments are realized. The system takes "biological limb motion modeling - satellite - radio frequency beacon - vision - inertial navigation fusion positioning" as the core, breaks through the positioning blind area of traditional animal control technology, and is applicable to high-precision task scenarios such as border patrol, military reconnaissance, battlefield operations, forest patrol, river patrol, outdoor exploration, and disaster rescue. Background Art

[0002] Existing animal control technologies have the following defects:

[0003] 1. Lack of limb movement perception: Dependence on inertial navigation (IMU) or GPS, unable to achieve autonomous positioning through biological body movement characteristics (such as joint angles, gait phases);

[0004] 2. High dependence on satellite positioning: It fails in indoor, cave and other sheltered environments (error > 5m), and cannot be fused with biological movement characteristics;

[0005] 3. Weak dynamic obstacle avoidance ability: The preset path lacks real-time adaptation to biological limb movements, resulting in control delay or movement conflicts.

[0006] The present invention realizes autonomous positioning without relying on implanted devices through non-invasive limb movement modeling and multi-mode satellite - radio frequency beacon - vision - inertial navigation fusion positioning technology, and maintains a positioning accuracy of <0.5m in GPS-denied environments. Summary of the Invention

[0007] System Composition

[0008] Intelligent control module: Supports online / offline artificial intelligence large models and intelligent operating systems.

[0009] Intelligent sensing module

[0010] 1. Multi-modal sensor array:

[0011] 1. Binocular vision camera (supporting night vision and underwater imaging);

[0012] 2. LiDAR (detection range ≥ 50m, accuracy ±2cm);

[0013] 3. Bioelectric sensor (collects EMG / EEG signals, sampling rate 1kHz);

[0014] 4. 9-axis IMU (accelerometer + gyroscope + magnetometer);

[0015] 5. Barometer and microphone array (supports sound source localization).

[0016] Multimodal interaction execution module

[0017] 1. Non-invasive stimulation unit:

[0018] 1. Piezoelectric ceramic array: can generate electric pulses with a pulse width of 0.1 - 10ms and adjustable voltage of 3 - 15V, and set the safety threshold according to the IEEE 1856 standard;

[0019] 2. Micro eccentric motor: supports vibration frequencies of 0.1 - 200Hz, encodes direction instructions (e.g., left front vibration represents a right turn);

[0020] 3. Ultrasonic speaker: emits command sound waves of 15 - 25kHz (perceivable by animals, inaudible to humans);

[0021] 4. Sound, light, or image output.

[0022] 2. Expandable tool interface:

[0023] 1. Supports modular mounting (such as infrared thermal imager, water quality sampler, first aid kit delivery device);

[0024] 2. Weapon system: automatically aims and attacks through machine vision and positioning technology.

[0025] 3. Multiple forms of communication units: satellite communication, radio communication, video transmission module, 2.4G, etc.

[0026] Energy management module

[0027] 1. Hybrid power supply system:

[0028] 1. Piezoelectric power generation layer (generates 5 - 20mW / cm 2 electrical energy when the animal moves);

[0029] 2. Flexible solar thin film (efficiency ≥ 23%, adaptable to curved surface fitting);

[0030] 3. Quick-changeable micro lithium battery (supports wireless charging, endurance ≥ 72h).

[0031] Core innovation concept

[0032] I. Positioning technology integrating joint flexible angle sensor and motion modeling This system achieves precise positioning based on biological ontology actions through the following technologies:

[0033] Wearable joint angle sensor module:

[0034] 1. High-precision joint angle sensors (range ±180°, accuracy ±0.1°) are integrated at key nodes (limb joints) on the biological body surface to collect joint angle data in real time (sampling rate 500Hz);

[0035] 2. Through kinematic calculation, short-term displacement and attitude changes of the organism are deduced (accuracy ±0.1m / s 2 ).

[0036] IMU error correction:

[0037] 1. The extended Kalman filter (EKF) is used to fuse joint angle data and IMU output to suppress the cumulative error of inertial navigation (the error rate after correction < 0.05% / min);

[0038] 2. Formula:

[0039] x^k = f(x^k-1,uk)+Kk(zk - h(x^k-1))x^k = f(x^k-1,uk)+Kk(zk

[0040] - h(x^k-1)) where zk is the observed value of joint angle and visual positioning, and Kk is the Kalman gain.

[0041] Multi-source fusion positioning technology (1) Satellite-inertial collaborative positioning (open environment)

[0042] Multi-mode GNSS receiver (u-blox ZED-F9P):

[0043] 1. Supports GPS / Galileo / Beidou multi-systems, and realizes centimeter-level accuracy (horizontal error < 2cm, vertical error < 5cm) through RTK differential positioning (reference station + mobile station);

[0044] 2. Data fusion: When satellite signals are available, the GNSS position and IMU data are combined through a loose coupling fusion algorithm to suppress the positioning drift caused by biological motion jitter.

[0045] 2. Data fusion: When satellite signals are available, the GNSS position and IMU data are combined through a loose coupling fusion algorithm to suppress the positioning drift caused by biological motion jitter.

[0046] (2) RF beacon - wireless WiFi - vision - inertial positioning (shielded environment)

[0047] RF beacon network:

[0048] 1. Deploy a radio frequency beacon base station (2.4 GHz band), and use the RSSI (Received Signal Strength Indication) and TDoA (Time Difference of Arrival) algorithms to achieve dynamic tag positioning (update frequency 10 Hz, error ±0.3 m);

[0049] Wireless WiFi positioning:

[0050] 1. Based on WiFi fingerprint positioning technology, through signal strength and multipath effect analysis, combined with an environmental map to achieve positioning (accuracy ±0.5 m);

[0051] Visual servo positioning:

[0052] 1. A binocular positioning camera (Basler acA2440) combined with the ORB-SLAM3 algorithm extracts environmental feature points (more than 1000 points per frame), and calculates the biological body pose through the PnP algorithm (accuracy ±0.1 m);

[0053] 2. Infrared assisted positioning (FLIR Boson 640): Mark the thermal contour of the biological body in low-light environments to enhance the robustness of feature point matching.

[0054] IMU short-term compensation:

[0055] 1. When visual or radio frequency beacon signals are lost (such as rapid turning, occlusion), maintain short-term positioning based on IMU data (<10 seconds, error <1 m).

[0056] (3) Cross-media positioning compensation

[0057] Underwater mode: Sonar beacon (30 kHz) + IMU data fusion, eliminate water flow interference through adaptive filtering (error <0.5 m); Aerial mode: Barometer (accuracy ±0.1 hPa) calibrates altitude, combined with IMU angular velocity data to resist wind disturbance.

[0058] Dynamic path intervention mechanism (1) Hazard perception level

[0059] Level 1 (terrain risk):

[0060] 1. A lidar (Velodyne VLP-16) matches the DEM map in real time to detect sudden slope changes (>45%) or ground collapses (hardness <100 kPa);

[0061] 2. Trigger condition: Slope > θ threshold or ground hardness < ρ threshold.

[0062] Level 2 (obstacle risk):

[0063] 1. A millimeter-wave radar (TIAWR1843) detects obstacles in front (detection distance 80 m, accuracy

[0064] (±0.1 m);

[0065] 2. Trigger condition: The distance to the obstacle is <2 m and the relative speed is >1 m / s.

[0066] Level 3 (biological risk):

[0067] 1. When the cortisol concentration (>50 ng / mL) or heart rate variability (HRV <50 ms) exceeds the limit, the protection mechanism is triggered.

[0068] (2) Intervention strategy

[0069] Path replanning:

[0070] 1. The improved RRT* algorithm is adopted to insert obstacle avoidance nodes on the original path to generate a smooth obstacle avoidance path (replanning time <200 ms);

[0071] 2. Path deviation correction: The actual trajectory is aligned with the preset path in real time through a sliding window optimization algorithm (window length 5 m), and the maximum allowable deviation is 1 m.

[0072] Multimodal control feedback:

[0073] 1. Tactile coding: 50 Hz vibration prompts a left turn, 100 Hz prompts acceleration, and 200 Hz for emergency braking;

[0074] 2. Acoustic and light guidance: Directional sound waves (15 kHz) indicate the dangerous direction, and the LED projection shows the path arrow (brightness 1000 lumens).

[0075] II. Precise control technical details

[0076] Biological motion intention fusion: Joint angle - IMU - environmental perception collaboration

[0077] Multisource data is fused through the Extended Kalman Filter (EKF) to achieve high-precision motion intention estimation and positioning. The specific data sources and their characteristics are as follows:

[0078] Displacement deduced from joint angles

[0079] 1. Characteristics: Displacement is deduced based on the change in joint angles.

[0080] 2. Error: The cumulative error rate is 0.2% / min.

[0081] IMU dead reckoning

[0082] 1. Characteristics: Short-term displacement is deduced based on the acceleration and angular velocity data of the Inertial Measurement Unit (IMU).

[0083] 2. Precision: The short-term precision is 0.1 m.

[0084] Visual SLAM Localization

[0085] 1. Feature: A moving object senses the surrounding environment through sensors (such as lidar, cameras, etc.) in an unknown environment, while estimating its own position and constructing an environmental map. During the SLAM process, the moving object can automatically identify landmarks or feature points in the environment as references and continuously correct its own position based on these references. When moving to a new position, the SLAM system will re-evaluate the feature points in the environment and automatically select new references, thus achieving dynamic localization and map updating.

[0086] 2. Accuracy: The localization accuracy is 0.1m.

[0087] Radio Frequency Beacon Ranging

[0088] 1. Feature: Realize localization based on the radio frequency signal beacon ranging technology.

[0089] 2. Accuracy: The localization accuracy is 0.3m.

[0090] GNSS Localization

[0091] 1. Feature: Realize localization based on the Global Navigation Satellite System (GNSS).

[0092] 2. Accuracy: In an open environment, the localization accuracy can reach 2cm.

[0093] Multi-Sensor Fusion Technology

[0094] By fusing data from multiple sensors (such as lidar, cameras, IMU, etc.), the environment can be perceived more accurately and appropriate references can be selected. For example, lidar can provide high-precision distance information, cameras can identify visual features in the environment, and IMU can provide motion information. Through data fusion algorithms, the system can dynamically select the optimal reference based on the reliability of sensor data and environmental conditions and generate a motion localization trajectory.

[0095] Machine Learning-Based Technology

[0096] Using machine learning algorithms (such as deep learning) can automatically identify and select references. For example, by training a neural network to identify landmarks or feature points in the environment and dynamically adjusting the references based on the stability and reliability of these feature points. This method can adapt to complex dynamic environments and can automatically learn the best reference selection strategy and generate a motion localization trajectory.

[0097] Combination Technology Based on Vision and Lidar

[0098] Combining the advantages of visual sensors (such as cameras) and lidar can achieve more reliable reference object selection and positioning. For example, by utilizing the high-precision distance measurement of lidar and the visual feature recognition of cameras, the system can dynamically select landmarks or objects in the environment as reference objects and generate motion positioning trajectories.

[0099] Fusion strategy

[0100] These technologies can be selected and combined according to specific application scenarios and requirements to achieve accurate positioning and trajectory tracking of moving objects in a dynamic environment. For example, by using the Extended Kalman Filter (EKF) to fuse the above multi-source data, the advantages of each data source can be fully utilized, the deficiencies of a single data source can be made up for, and high-precision and robust motion intention estimation and positioning can be achieved.

[0101] Ethical control design

[0102] Non-invasive sensing: All sensors are designed to be wearable to avoid damage to organisms; stress response monitoring (cortisol detection sensor) reduces the stimulation intensity in real time according to the cortisol concentration (formula: Istim = 0.1×(1 + Ccortisol / 50) - 1mA);

[0103] Self-destruction mechanism: The electrolysis-triggered degradation circuit fails within 60s, and the residues are non-toxic and biodegradable.

[0104] Biosafety measures

[0105] Establish an animal welfare assessment system (referring to the AAALAC standard)

[0106] Set three levels of stimulation thresholds:

[0107] Level1: Tactile feedback (0.1mA)

[0108] Level2: Conditioned reflex stimulation (1mA)

[0109] Level3: Emergency braking (5mA)

[0110] Innovation effect

[0111] Verified by actual tests:

[0112] The duration of search and rescue missions is increased by 7.8 times (compared with Boston Dynamics Spot)

[0113] The instruction recognition accuracy rate reaches 99.3% (MIT bio-robot benchmark test set)

[0114] The animal stress index is reduced by 82% (ELISA method for salivary cortisol detection)

[0115] When the intelligent control module for emergency safety measures detects an unsafe environment, it can call the tools of the tool interface or urgently brake the animal. Brief Description of the Drawings

[0116] Figure 1 It is a schematic diagram of the principle of a system and method for controlling animals according to the present invention. It only explains the principle and does not constitute a limitation to the patent.

[0117] The principle of this patent is that a communication module such as a remote control issues an instruction, which is transmitted to the controller through the communication module. The controller issues an instruction to make the execution module execute an action or directly drive the execution module; this system also has a power supply and a charging unit for charging the power supply. At the same time, the execution module can also feedback information to the remote control through the control module and the communication module. There are many communication modules that can be selected, such as a video transmission module, a radio module, a mobile phone communication module, a 2.4G communication module, etc. It can also have no communication module. We set the tasks in advance, and monitor and compare the execution situation through the controller and the execution module and correct the execution in real time; or instructions can be sent to the communication module through the background server. Detailed Description of the Invention

[0118] Embodiment 1: Hardware Configuration of the Mountain Freight Mule System:

[0119] · Quadruped exoskeleton: Titanium alloy frame (12 degrees of freedom), with flexible piezoelectric sensors integrated on the joint surfaces;

[0120] · Positioning unit: Multi-mode GNSS receiver + RF beacon network + binocular vision camera + back IMU module.

[0121] Control Flow:

[0122] · Global path planning: The A* algorithm generates a path, avoiding areas with a slope > 30%;

[0123] · Real-time motion tracking: The joint angle data fuses IMU and visual SLAM, and updates the position every 100 ms;

[0124] · Hazard intervention: When detecting the risk of collapse, the RRT* algorithm replans the path (time-consuming 180 ms);

[0125] · Energy management: The kinetic energy recovery efficiency when going downhill is ≥ 55%, and the battery life is infinite on sunny days.

[0126] Measured Data:

[0127] · Under the condition of a load of 100 kg, the passing rate of a 45° slope is 100%, the positioning error is 0.28 m, and the stress index is 0.22.

[0128] Embodiment 2: Innovative Design of the Disaster Search and Rescue Dog System

[0129] · Plantar piezoelectric sensor: Detect the ground hardness in real time (accuracy ±5 kPa);

[0130] · Millimeter-wave life detection radar: Detection accuracy of breathing frequency ±0.1 Hz.

[0131] Precision control:

[0132] · After discovering the survivor, inject drugs with a microneedle array (accuracy ±1 μL), and send coordinates synchronously (LoRa communication, delay <1 s);

[0133] · When the cortisol concentration >50 ng / mL, the stimulation intensity is automatically reduced to 0.05 mA.

[0134] Example 3: Technical parameters of the space-based relay pigeon system:

[0135] · Visual-RF beacon positioning (error ±0.5 m), anti-wind disturbance control (wind speed 15 m / s);

[0136] · Feather solar thin film (conversion efficiency 28%) + body fluid fuel cell (50 mW / cm 3 )

[0137] · Miniaturized design: Total weight <35 g (6% of body weight)

[0138] · Equipped with a software-defined radio (SDR) module to achieve dynamic spectrum access

[0139] · Apply MIMO technology to establish a 1.2 Gbps space-to-ground link at an altitude of 300 m.

[0140] Function verification:

[0141] · 8-hour cruise (altitude 5000 m, -20 °C), dynamic spectrum switching delay <100 ms.

[0142] Technical effects

[0143] Tested by a third party certified by CNAS:

[0144]

[0145] Technical advantages

[0146] High-precision positioning: Achieve sub-meter positioning (error <0.3 m) in a GPS-denied environment through multi-source fusion positioning technology (RF beacon, wireless WiFi, visual SLAM, IMU);

[0147] Dynamic adaptability: Based on the motion characteristics of the biological body (joint angles, gait phases) and real-time environment perception, dynamically adjust path planning and motion control;

[0148] Non-invasive design: All sensors are wearable, avoiding damage to organisms and conforming to the ethical design principle;

[0149] Applicable in all scenarios: Supports various environments such as on land, underwater, and in the air, meeting the requirements of complex tasks;

[0150] High-efficiency management: Extends the system's battery life and reduces energy consumption through kinetic energy recovery and solar energy technologies.

[0151] Future development direction - Intelligent upgrade: Introduces deep learning algorithms to further enhance the ability to recognize biological motion intentions and perceive the environment;

[0152] Multi-biological collaboration: Develops multi-biological collaborative control technologies to achieve efficient execution of group tasks;

[0153] Modular design: Optimizes the system's hardware and software architectures, supporting rapid deployment and function expansion;

[0154] Ecological friendliness: Develops more environmentally friendly materials and energy technologies to reduce the system's impact on the environment;

[0155] Commercial promotion: Expands the civilian market (such as pet training, agricultural monitoring) to promote the popularization and application of the technology.

[0156] Conclusion

[0157] Through the deep integration of silicon-based technology and living organisms, the present invention proposes a non-invasive and high-precision biological intelligent control system. Based on multi-source fusion technologies such as joint angle sensors, RF beacons, and wireless WiFi positioning, the system achieves sub-meter-level positioning and precise control in complex environments. Its innovative design not only breaks through the limitations of traditional animal control technologies but also shows broad application prospects in fields such as military, rescue, and ecological monitoring. In the future, with the continuous optimization and upgrade of the technology, this system will bring more breakthroughs and changes to the field of intelligent biological control.

Claims

1. A silicon-based and carbon-based integrated biological intelligent control system, characterized in that, Comprising: An intelligent control module, preferably including support for online / offline artificial intelligence large models and / or intelligent operating systems; The operating system supports input of path and / or key point coordinates on the path in online and / or offline manners and / or supports issuing execution control instructions and / or tasks in online and offline manners; And / or an intelligent perception module, preferably including a multi-modal sensor array for collecting biological motion and environmental information; And / or a multi-modal interaction execution module, including a non-invasive stimulation unit and an expandable tool interface for executing control instructions and / or tasks; And / or an energy management module, including a hybrid power supply system for providing continuous energy for the system. The intelligent perception module includes: a binocular vision camera supporting night vision and underwater imaging; and / or a lidar; and / or a bioelectric sensor for collecting EMG / EEG signals; and / or a 9-axis IMU including an accelerometer, a gyroscope, and a magnetometer; and / or a barometer and a microphone array supporting sound source localization. The multi-modal interaction execution module includes: a piezoelectric ceramic array capable of generating adjustable electric pulses; and / or a micro eccentric motor; and / or an ultrasonic speaker for emitting instruction sound waves of 15 - 25 kHz; and / or a sound, light, or image output device; and / or a communication unit, and the communication unit includes satellite communication and / or wireless communication and / or 2.4G communication and / or a video transmission module.

2. The system according to claim 1, wherein The energy management module includes:

3. The system according to claim 1, characterized in that A piezoelectric power generation layer; and / or a flexible solar thin film; and / or a replaceable micro battery, preferably supporting wireless charging.

4. The system according to claim 1, wherein The system realizes precise motion tracking and precise control of a living organism in a complex environment through multi-modal limb motion perception and multi-source heterogeneous positioning fusion technology. The multi-source heterogeneous positioning fusion technology includes: satellite-inertial collaborative positioning to achieve high-precision positioning through a multi-mode GNSS receiver; and / or RF beacon-WiFi-vision-inertial positioning to achieve dynamic tag positioning in a shielding environment; and / or visual servo positioning, preferably achieving positioning through a binocular positioning camera combined with the ORB-SLAM3 algorithm; and / or IMU short-term compensation to maintain short-term positioning when visual or RF beacon signals are lost.

5. The system according to claim 1, characterized in that, The system further includes a dynamic path intervention mechanism, including: a risk perception level for detecting terrain risks, obstacle risks, and biological risks; and / or an intervention strategy using an improved RRT* algorithm for path replanning and / or path deviation correction and / or tool use.

6. The system according to claim 5, characterized in that, The system fuses multi-source data through an extended Kalman filter (EKF) to achieve high-precision motion intention estimation and positioning; and / or the system adopts a non-invasive sensing design, and all sensors are wearable to avoid damage to the organism; and / or the system further includes a self-destruction mechanism, preferably triggered by electrolysis to degrade the circuit; and / or the system establishes an animal welfare assessment system and sets a three-level stimulation threshold to ensure biological safety.

7. The system according to claim 1, characterized in that Including the following steps:

8. The system according to claim 1, characterized in that, 1. Collect biological motion and environmental information through a multi-modal sensor array; 9. A full-scenario autonomous collaboration method based on a silicon-based and carbon-based integrated biological intelligent control system, characterized in that, 2. Achieve sub-meter motion tracking and precise control of a living organism in a complex environment through multi-source heterogeneous positioning fusion technology; ​ ​ 3. Execute control instructions and tasks through the multimodal interaction execution module; 4. Provide continuous energy for the system through the energy management module; And / or the method, characterized in that the multi-source heterogeneous positioning fusion technology includes the following features or any combination of the following features:

1. Satellite-inertial collaborative positioning, achieving centimeter-level accuracy through a multi-mode GNSS receiver; 2. RF beacon-WiFi-vision-inertial positioning, realizing dynamic tag positioning in a sheltered environment; 3. Vision servo positioning, achieving positioning through a binocular positioning camera combined with the ORB-SLAM3 algorithm; 4. IMU short-term compensation, maintaining short-term positioning when vision or RF beacon signals are lost; 5. Vision SLAM positioning; 6. Multi-sensor fusion positioning technology; 7. Machine learning-based positioning technology; 8. Combined positioning technology based on vision and lidar; And / or the method, characterized in that the method fuses multi-source data through an extended Kalman filter (EKF) to achieve high-precision motion intention estimation and positioning; And / or the method, characterized in that the method adopts a non-invasive sensing design, and all sensors are wearable to avoid biological damage; And / or the method, characterized in that the method establishes an animal welfare assessment system, sets a three-level stimulation threshold to ensure biosafety; And / or the method, characterized in that the method establishes emergency measures for calling tool interfaces and / or emergency braking.

Images