An explosion-proof humanoid robot 5G cross-domain low-latency remote control system and method

Abstract

The present application belongs to the technical field of remote control of robots, and particularly relates to a 5G cross-domain low-latency remote control system and method for an explosion-proof humanoid robot. The system comprises a master end teleoperation intelligent cockpit and a slave end remote explosion-proof humanoid robot, and the two are connected through a 5G network for information interaction. The system transmits video streams through H.264 encoding and WebRTC protocol, packs sensor data through JSON, and independently processes information for humanoid control and chassis control. The present application can realize accurate remote force feedback and control, with a humanoid control latency within 100 ms and a chassis control latency within 150 ms, effectively solving the stability and latency problems of remote cross-domain control and ensuring the personal safety of operators in dangerous operating environments.

Description

Technical Field

[0001] This invention pertains to robot remote control technology, specifically a 5G cross-domain low-latency remote control system and method for an explosion-proof humanoid robot. Background Technology

[0002] 5G technology, especially its Ultra Reliable Low Latency Communication (URLLC) capabilities, is paving new paths for cross-domain low-latency remote control. Currently, this technology is being piloted and applied in core scenarios such as industrial automation (e.g., remote control of port gantry cranes and mining trucks), remote surgery, precise load control of power grids, and intelligent transportation vehicle-road cooperation. Its core lies in leveraging the millisecond-level (theoretically down to 1ms) end-to-end latency, ultra-high reliability, and wide-area coverage of 5G networks to efficiently connect control terminals (e.g., operator cabins) and execution terminals (e.g., robotic arms, drones, and vehicles) deployed in different regions (potentially spanning cities, provinces, and even national borders), enabling real-time, precise human-machine interaction or automatic control of remote equipment. This represents a significant evolution in remote operation technology from traditional dedicated wired networks or limited local wireless networks to wide-area, flexible deployment based on public mobile communications.

[0003] Currently, the level of intelligence in robots remains limited and cannot match human intelligence. Therefore, in hazardous work areas (such as bomb disposal, aerospace, and nuclear power), exploring how to integrate human intelligence into remote robot control is of great significance. This human-machine collaborative model can effectively ensure personnel safety while fully leveraging human real-time decision-making capabilities, demonstrating significant application value. While traditional wired remote operation connections are stable, network cables restrict equipment flexibility and deployment is relatively cumbersome. Adopting 5G technology to replace wired connections can significantly improve operational flexibility and deployment convenience.

[0004] Existing technologies have the following problems: traditional wired remote operation methods are inflexible and cumbersome to deploy, and cannot meet the needs of cross-domain remote control; existing wireless remote operating systems have high latency, making it difficult to achieve real-time and accurate control; network instability in cross-domain remote control leads to serious latency jitter and packet loss problems; there is a lack of effective force feedback mechanisms, and operators cannot obtain real tactile feedback; system stability and reliability need to be improved, and it is difficult to adapt to harsh communication environments. Summary of the Invention

[0005] Purpose of the invention: The purpose of this invention is to address the shortcomings of existing technologies by providing a 5G cross-domain low-latency remote control system and method for explosion-proof humanoid robots, so as to solve the problems of high latency, poor stability, and lack of effective force feedback in existing technologies, and to achieve safe and reliable remote control in hazardous working environments.

[0006] Technical Solution: To achieve the above objectives, the present invention provides a 5G cross-domain low-latency remote control system for explosion-proof humanoid robots, comprising:

[0007] The master-side remote-controlled smart cockpit is used by the operator to remotely send control commands and receive feedback information.

[0008] A remote explosion-proof humanoid robot, comprising a mobile chassis and a humanoid body, is used to execute control commands sent by the master remote-operated intelligent cockpit.

[0009] A 5G-based communication system is used to enable 5G communication between the master-end remotely operated intelligent cockpit and the slave-end remote explosion-proof humanoid robot.

[0010] The humanoid body and the mobile chassis are designed with a separate hardware architecture, and their control command information is transmitted through independent 5G communication channels. The communication system employs protocol optimization and a machine learning-based algorithm to combat weak network conditions. The protocol optimization includes using a custom lightweight UDP-based protocol to replace TCP, implementing a retransmission mechanism based on sequence numbers and selective acknowledgment for critical data at the application layer, compressing IP / UDP packet headers using the ROHC algorithm, and embedding CRC32 checksums in application layer data packets for end-to-end verification. The algorithm to combat weak network conditions uses machine learning to predict network congestion and channel status, and dynamically adjusts the modulation scheme, power, and frequency band.

[0011] The 5G cross-domain low-latency remote control method for explosion-proof humanoid robots based on the above system includes the following steps:

[0012] S1: Establish master-slave communication connection and control architecture:

[0013] A remotely operated intelligent cockpit is built on the main end and a main end intelligent terminal is set up. A remote explosion-proof humanoid robot is built on the slave end and a humanoid robot intelligent terminal and a mobile chassis intelligent terminal are set up.

[0014] A communication connection based on a 5G network is established, and a separate control architecture is adopted to separate the control of the humanoid robot from the control of the mobile chassis. The main intelligent terminal and the humanoid robot intelligent terminal communicate through a 5G router, and the main intelligent terminal and the mobile chassis intelligent terminal communicate through a 5G module.

[0015] S2: Execute chassis movement control:

[0016] The main intelligent terminal converts the operation data of the driving control device into linear velocity and angular velocity information in real time, and transmits it to the mobile chassis intelligent terminal through the 5G communication channel. The mobile chassis intelligent terminal then controls the movement and steering of the mobile chassis.

[0017] S3: Perform teleoperation control of the humanoid robot:

[0018] The master intelligent terminal obtains the joint motor position, torque and gripper control information of the remote control device in real time, and transmits it to the slave humanoid robot through the 5G communication channel to realize the synchronous motion control of the two arms and gripper.

[0019] S4: Perform video and sensor data transmission.

[0020] The humanoid robot intelligent terminal collects video streams through multiple cameras and transmits them in real time to the main intelligent terminal for decoding and display; sensor data, including those from the odometer, IMU inertial measurement unit, and lidar, are packaged in JSON format and transmitted in real time to the main intelligent terminal via a 5G communication channel for environmental information analysis.

[0021] S5: Implementation force feedback processing:

[0022] The humanoid robot intelligent terminal acquires the position, torque and six-dimensional force sensor information of the humanoid joint motor in real time, applies a low-pass filter to eliminate high-frequency noise, calculates the force feedback result through a bilateral control algorithm, transmits it to the master intelligent terminal through a 5G communication channel, and provides the force feedback feeling to the operator through the remote control device to realize master-slave force feedback closed-loop operation.

[0023] S6: Optimization of Execution Protocol and Resource Scheduling:

[0024] The transport protocol is optimized at the protocol layer, including using a custom lightweight UDP-based protocol to replace TCP, implementing a retransmission mechanism based on sequence number and selective acknowledgment for critical data at the application layer, using the ROHC algorithm to compress IP / UDP packet headers, and embedding CRC32 checksums in application layer data packets for end-to-end verification to reduce transmission overhead and latency.

[0025] In the resource scheduling and management layer, a machine learning-based anti-weak network algorithm is introduced. By predicting network congestion and channel status, the modulation scheme, power and frequency band are dynamically adjusted to balance latency and energy efficiency.

[0026] Beneficial effects: Compared with the prior art, the advantages of this invention are as follows: This invention achieves true cross-domain low-latency remote control, with humanoid control latency within 100ms and chassis control latency within 150ms; it provides a complete force feedback mechanism, enabling operators to obtain realistic tactile perception; it adopts a split control architecture, improving the stability and reliability of the system; through protocol optimization and intelligent resource scheduling, it significantly reduces communication latency and bandwidth consumption; it is suitable for hazardous operating environments and can effectively ensure personnel safety. Attached Figure Description

[0027] Figure 1A schematic diagram of the overall architecture of a 5G cross-domain low-latency remote control system for explosion-proof humanoid robots.

[0028] Figure 2 This is a diagram of the remote cockpit software interface, showing the buttons and display area of ​​the control panel.

[0029] Figure 3 The flowchart of the 5G cross-domain low-latency remote control protocol for explosion-proof humanoid robots illustrates the data transmission and processing process between the master and slave ends.

[0030] Figure 1 In the Chinese: 10. Remote-operated intelligent cockpit; 20. Remote explosion-proof humanoid robot; 30. 5G base station; 101. Remote driving control cabin; 102. Remote control device; 103. Master intelligent terminal; 104. Master 5G router; 201. Mobile chassis; 202. Humanoid robot; 203. Mobile chassis intelligent terminal; 204. 5G module; 205. Humanoid robot intelligent terminal; 206. Slave 5G router. Detailed Implementation

[0031] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings, but the scope of protection of the present invention is not limited to the embodiments described.

[0032] Example 1: As Figure 1 As shown, the explosion-proof humanoid robot 5G cross-domain low-latency remote control system of the present invention includes a master-end remotely operated intelligent cockpit 10, a slave-end remote explosion-proof humanoid robot 20, and a 5G base station 30.

[0033] The specific configuration of the master-operated remote smart cockpit 10 is as follows:

[0034] Remote driving control cockpit 101: It integrates all control interfaces and human-machine interaction functions, and has an ergonomically designed seat to provide a comfortable operating environment for the operator; the control panel is integrated in front of the legs of the cockpit, and the control panel is equipped with driving control devices, including automotive-grade accelerator pedal, brake pedal and steering wheel; the control panel has a vertical display, which adopts high resolution multi-screen display to display the real-time images of multiple cameras from the slave end;

[0035] Remote control device 102: includes a multi-degree-of-freedom robotic arm structure, capable of accurately simulating human hand movements;

[0036] Main intelligent terminal 103: It adopts an industrial control computer with a dedicated graphics card with hardware code, which has powerful data processing capabilities;

[0037] Main-end 5G Router 104: Adopts an industrial-grade 5G router to ensure communication stability and reliability.

[0038] The specific configuration of the remote-controlled explosion-proof humanoid robot is as follows:

[0039] Mobile Chassis 201: It adopts a four-wheel drive wheeled chassis and has omnidirectional mobility;

[0040] The humanoid robot 202 features: a multi-DOF neck (3 degrees of freedom for flexible head rotation); a waist (2 degrees of freedom for forward / backward bending and left / right rotation); hips connecting the waist to the chassis for cushioning; humanoid arms (7 degrees of freedom per arm, mimicking human arm movement); an end effector for grasping objects of various shapes and sizes; a six-dimensional force sensor for sensing forces and torques in the X, Y, and Z directions; a 360-degree scanning lidar for environmental mapping and obstacle detection; an inertial measurement unit (IMU) providing attitude, acceleration, and angular velocity information; an odometry system based on a wheel encoder for precise position measurement; and multiple cameras, including head, chest, and hand cameras, providing a 360-degree field of view.

[0041] The intelligent terminals include: the humanoid robot intelligent terminal 205 uses NVIDIA Jetson nano, and the mobile chassis intelligent terminal 203 uses Raspberry Pi 4B.

[0042] The 5G communication module includes: a slave 5G router 206 for 5G communication of a humanoid robot, and a 5G module 204 for 5G communication of a mobile chassis.

[0043] Example 2: As Figure 2 As shown, the process of building the communication system for the remotely operated intelligent cockpit is as follows:

[0044] Network connection establishment: The master industrial control computer establishes a communication connection with the slave NVIDIA Jetson nano through an industrial-grade 5G router; the chassis Raspberry Pi establishes an independent communication channel with the master through a 5G module.

[0045] QT Interface Development: Developing a remote cockpit visualization interface based on the QT framework, mainly including:

[0046] Cockpit button panel: Displays various control buttons and status indicator lights;

[0047] Functional interface area: includes real-time display of sensor information status and remote chassis fault clearing function;

[0048] Video window area: Displays real-time footage from multiple cameras on the slave end.

[0049] Client program: Reads information from external sensors, including the status information of menu buttons, accelerator pedal, brake pedal and steering wheel.

[0050] Functional modules: Enable auxiliary functions such as video window refresh, device restart, and fault diagnosis.

[0051] Example 3: A 5G cross-domain low-latency remote control method for an explosion-proof humanoid robot based on the system provided in Example 1, comprising the following steps:

[0052] Step 1: Construct a master-end remotely operated intelligent cockpit and a slave-end remotely operated explosion-proof humanoid robot based on 5G cross-domain low latency:

[0053] The entire control system of this invention consists of a master-end remotely operated intelligent cockpit and a slave-end remotely operated explosion-proof humanoid robot, such as... Figure 1 As shown, the master-end remotely operated intelligent cockpit serves as the control end, including a seat, remote control device, driving control device (accelerator / brake / steering wheel / intelligent cockpit), display, control panel, master-end intelligent terminal, and master-end 5G router, among other equipment. The slave-end remotely operated explosion-proof humanoid robot serves as the controlled end, including a mobile chassis, humanoid body (humanoid multi-degree-of-freedom neck, humanoid waist, humanoid hips), humanoid arms, end effector grippers, six-dimensional force sensor, lidar, IMU, odometer, multiple cameras, slave-end intelligent terminal, 5G module, and slave-end 5G router, among other equipment.

[0054] Step Two: Establish a remote information interaction and communication control system for the remotely operated intelligent cockpit. The intelligent terminal of the remotely operated intelligent cockpit is an industrial control computer with a dedicated graphics card and hardware code, while the intelligent terminal of the humanoid robot is an NVIDIA Jetson Nano. Both communicate with each other via an industrial-grade 5G router. The intelligent terminal of the mobile chassis is a Raspberry Pi, which communicates with the remotely operated intelligent cockpit via a 5G module.

[0055] In the slave-end humanoid robot, the Jetson Nano is responsible for voice, video, and sensor data transmission and reception, such as video streaming, and does not involve upper body joint control; the Raspberry Pi is used for communication with the remote cockpit and video streaming, and there is no business overlap between the Jetson Nano and the Raspberry Pi; the master-end industrial control computer is responsible for real-time calculation and data transmission and reception of teleoperation and force feedback.

[0056] A remote cockpit visualization interface was built using QT, and its functional interface diagram is shown below. Figure 2 As shown, it mainly includes a cockpit button panel, a function interface area, and a video window area. The button panel is based on a QT interface program and reads external sensor information (menu buttons, accelerator, steering wheel) through a client. The function interface includes real-time status display of external sensor information, remote chassis fault clearing, video window refresh, and device restart.

[0057] Step 3: When the main operator is driving in the cockpit, and the cockpit enters the remote control mode, the current speed conversion data is calculated in real time, and the linear velocity / angular velocity information is sent to the Raspberry Pi on the remote mobile chassis. The Raspberry Pi then forwards the information to the chassis controller via the 485 protocol to control the mobile chassis to perform movement and steering control.

[0058] Step 4: The remote explosion-proof humanoid robot acquires video streams from multiple cameras and pushes them to the remotely operated smart cockpit via H.264 encoding and WebRTC streaming protocol. The video streams are automatically decoded and displayed in real time via the web interface. Other sensor data, including those from the odometer, IMU, and LiDAR, are packaged in JSON format and transmitted in real time via a 5G module to the main smart terminal for analysis and judgment of environmental information.

[0059] Prioritizing the delivery of robot joint data and sensor data, with robot body parameter data taking precedence over video streams, the application will assign different Quality of Service (QoS) tags to different data types. For example, the DSCP (Differential Service Code Point) field in the IP header will be used. Adaptive bitrate adjustment will be employed, with real-time monitoring of bandwidth and packet loss rate. If the network is poor, the video stream quality will be automatically reduced to ensure the real-time performance and fidelity of the robot body parameters. Robot parameters and sensor data will be transmitted via UDP, while the video stream will be based on TCP.

[0060] Step 5: When the main operator finishes driving control of the remote robot chassis and begins to remotely operate the telecontrol device, the intelligent terminal obtains the position torque of each joint motor and the gripper control information in real time, and packages the data into JSON and transmits it to the remote humanoid robot for control in real time via a 5G router.

[0061] Step Six: The humanoid robot's intelligent terminal will perform real-time calculations on the position torque of each joint motor and the information from the six-dimensional force sensor, and transmit the force feedback results to the main operator in real time via the 5G module for perception, analysis, judgment, and further real-time control and adjustment of the humanoid robot.

[0062] Master-slave force feedback mechanism: The operator manipulates the master end (7-DOF force feedback input device), and the slave end (a 7-DOF robot) completely replicates the motion of the master end at a remote end; at the same time, the force generated by the interaction between the slave end and the environment is accurately fed back to the master end, giving the operator an immersive tactile experience.

[0063] like Figure 3As shown, the operator moves the master device, and the high-precision sensors on the master end measure the angles of the 7 joints and the pose of the end effector in real time. Through high-bandwidth closed-loop control, the master end's motion commands are mapped to the slave end in real time to reproduce the motion, while the slave end's force sensor information is mapped back to the master end in real time to generate feedback force, forming a position-force coupled bilateral control. Based on model inverse solution and impedance adjustment algorithms, the model differences between the master and slave ends are handled. Through state observers and feedforward dynamic compensation, transmission delays and dynamic interferences are offset, ensuring the transparency and stability of the operation.

[0064] This invention employs a position-force-based bilateral control architecture, specifically implemented as follows:

[0065] Forward path: Master pose Directly used as the desired pose of the slave end ;

[0066] Reverse channel: Force measurement from the end Directly used as the main feedback force ;

[0067] Inverse kinematics (IK) algorithm is used to convert pose commands into kinematics. Real-time calculation of angle commands for each joint at the end. ;

[0068] Impedance adjustment: by adjusting the impedance model Damping in and stiffness The parameters allow the slave end to exhibit different compliance characteristics, thereby addressing the differences in dynamic models between the master and slave ends.

[0069] Using the Romberg observer, its mathematical model is as follows:

[0070]

[0071] in, It is the estimated state vector. This is the derivative of the estimated state, where A, B, and C are the system matrix, input matrix, and output matrix, respectively. It's a delay. The control input, It's a delay. The measurement output, L is the observer gain matrix; the observer utilizes delayed control commands. and sensor data By continuously correcting the predicted state through the gain matrix L, the real-time velocity and contact force of the slave end can be estimated without delay. .

[0072] Feedforward dynamic compensation: The compensation term is ,

[0073] in, It is feedforward compensation torque. It is a joint position vector. It is the joint velocity vector. , These are the desired joint accelerations and velocities. It is the inertia matrix. It is the Coriolis and centrifugal force matrix. It is the gravity vector, and the calculated compensation term Used to compensate for inertial forces, including those of the humanoid robot itself. ,gravity Coriolis force and centrifugal force The dynamic effects, including those of the system, significantly reduce the burden of feedback control and improve system transparency and stability.

[0074] Step 7: To solve the problem of control stability in remote cross-domain applications, the upper body control of the humanoid robot in the control system of this invention is based on the humanoid controller + Raspberry Pi approach, while the chassis is controlled by the chassis controller + Raspberry Pi approach, with hardware isolation to avoid affecting their respective data interaction; the control information of the humanoid robot and the control information of the moving chassis are encoded and decoded separately, and the data interaction and control operation do not affect each other.

[0075] Step 8: To ensure the accuracy of the force feedback results and to avoid affecting the normal operation of the remote operator at the master end, this invention uses low-pass filtering to determine a suitable threshold to eliminate the influence of minute changes, thereby improving the stability and effectiveness of the entire force feedback process.

[0076] Low-pass filtering in force feedback control mainly includes cutoff frequency and filtering algorithm: the core basis for joint module noise is the noise frequency. The cutoff frequency must be higher than the effective frequency of the desired force signal (usually from contact or human operation, generally below 20-30Hz), but much lower than high-frequency electronic noise (usually >100Hz), to strike a balance between fidelity (preserving the true force signal) and noise reduction. The specific value is fine-tuned according to the actual working conditions; the cutoff frequency of the dynamic bandwidth cannot exceed the effective force feedback bandwidth of the master / slave robot. If the frequency of the filtered force signal is higher than the frequency that the robot can respond to, it will lead to system instability (such as oscillation).

[0077] Step Nine: To solve the latency problem of remote cross-domain communication, this invention accelerates and optimizes the protocol and transport layer, trims redundant TCP / IP fields, adopts a lightweight protocol (a custom reliable transport protocol based on UDP), reduces packet encapsulation and decapsulation overhead, and compresses IP / UDP packets.

[0078] Specifically, the present invention makes the following targeted optimizations to the transmission protocol: (1) Protocol selection: a lightweight custom protocol based on UDP is adopted to replace TCP to eliminate the delay caused by handshake, retransmission and congestion control; (2) Custom reliability: a lightweight retransmission mechanism based on sequence number and selective acknowledgment (SACK) is implemented for key data (such as control commands) at the application layer; (3) Packet header compression: the ROHC algorithm is adopted to compress the 40-byte IP / UDP packet header to typically 1-4 bytes in the continuous stream, which greatly reduces the transmission overhead; (4) 5G network adaptation: the high bandwidth and jitter characteristics of 5G network are adapted through forward error correction (FEC) and delay-based congestion control algorithm.

[0079] By employing the aforementioned technical methods, redundant data transmission is reduced, the protocol stack is simplified, and latency is lowered. A standard IP / UDP header is 28 bytes long, and after ROHC compression, it typically becomes 3-4 bytes. Using a conservative estimate of 4 bytes, the overhead reduction is: 28 - 4 = 24 bytes; the overhead reduction percentage is: (24 / 28) × 100% ≈ 85.7%. Considering factors such as connection establishment and non-ideal packet compression, reducing redundant data transmission lowers overhead by 60%, simplifies the protocol stack, and reduces latency.

[0080] Directly removing the TCP checksum would completely eliminate the transport layer's bit error detection capability. In complex networks such as 5G / WAN, although the underlying link reliability is already high, there is still a very low probability of bit flips. Undetected errors can cause robots to receive incorrect commands or status data, potentially leading to safety and stability risks.

[0081] The compensation method is to perform end-to-end verification at the application layer: embedding this verification in the application layer header of all critical data packets (such as motion commands and force feedback values).

[0082] High-efficiency check codes such as RC32; selective retransmission: when the application layer check fails, the receiver will request the sender to retransmit the specific key data packet; data priority differentiation: for extremely high frequency and low priority sensor data, under the premise of confirming that the underlying link quality is extremely high, a small amount of reliability can be sacrificed in exchange for extremely low latency, but the core control commands must always have strong check.

[0083] Step 10: To address the latency issue in remote cross-domain communication, this invention optimizes resource scheduling and management layers, improves traditional weak network resilience algorithms, and introduces machine learning algorithms to dynamically balance latency and energy efficiency. Machine learning is used to predict network congestion and channel conditions, dynamically adjusting modulation schemes, power, and frequency bands to balance latency and energy efficiency.

[0084] The machine learning algorithm used in this application is a combination of online supervised learning and deep reinforcement learning (DRL).

[0085] Online supervised learning: used to build real-time network state prediction models. Its input is a sequence of historical network parameters, and its output is a prediction of the short-term network state (such as latency and bandwidth).

[0086] Deep Reinforcement Learning (DRL): The core algorithm is modeled as a Markov Decision Process (MDP).

[0087] Intelligent Agent: Resource Scheduler;

[0088] State: Real-time and predicted network status (such as latency, packet loss rate, bandwidth, signal strength).

[0089] Action: The communication parameters to be adjusted (such as switching frequency bands, adjusting transmit power, or changing the modulation and coding scheme MCS).

[0090] Reward: The goal is to minimize latency, maintain reliability (reduce packet loss rate), and penalize energy-intensive operations. The algorithm learns by continuously trying different approaches in specific network states.

[0091] The robotic system employs a hybrid approach of "pre-training + online fine-tuning":

[0092] Pre-trained model: The model is pre-trained offline in the dataset and simulation environment to obtain a base model.

[0093] Online learning and fine-tuning: The model is continuously fine-tuned based on real-time network data.

[0094] Fine-tuning the internal parameters of the control algorithm (such as the weights of a neural network) allows its output actions to adapt to real-world dynamics (such as varying friction and inertia) and real-world sensor noise (such as image blurring and force perception delay). This is typically done quickly using a small number of online demos or real-time reward signals.

[0095] Dynamic parameter tuning: Adjusting parameters of the underlying controller, such as PID gain, stiffness, and damping in impedance control. The goal is to make the robot's "feel" better suited to the current task, such as improving sensitivity during delicate operations or increasing stability during heavy assembly.

[0096] Online trajectory correction: Based on real-time sensor feedback (especially force sensing), the predetermined motion trajectory is dynamically adjusted. For example, if a jam is detected during insertion, the position and orientation of the end are fine-tuned to align with the hole; if an increase in contact force is detected, the amplitude of movement is automatically reduced to avoid damaging the object.

[0097] The key parameters input to real-time prediction models mainly include:

[0098] Latency: including one-way delay and round-trip time (RTT), is the core prediction target;

[0099] Packet Loss Rate: A key indicator for measuring channel reliability;

[0100] Bandwidth: An estimate of the available uplink and downlink bandwidth;

[0101] Jitter: The rate of change in latency, affecting the stability of transmission;

[0102] Signal strength (RSRP / RSRQ): In 5G / 4G networks, this is a fundamental physical layer parameter that determines connection quality.

[0103] These parameters are combined into a time series, which serves as input to the prediction model and the reinforcement learning state space.

[0104] Based on the output of Markov decisions, the robot system dynamically adjusts the following communication parameters, and the mechanism is as follows:

[0105] Modulation mechanism: When the predicted network quality is good (high signal-to-noise ratio), high-order modulation (such as 1024-QAM) and efficient coding are selected to improve spectral efficiency, transmit more data in the same amount of time, and thus reduce latency. When the predicted network quality is poor, it automatically degrades to low-order modulation (such as QPSK) and redundant coding, sacrificing rate for reliability and avoiding greater latency caused by retransmission.

[0106] Power mechanism: Power optimization is performed while ensuring connection reliability. When the link quality is expected to be good, the transmit power is intelligently reduced to directly save equipment energy; when signal attenuation or increased interference is predicted, the power is moderately increased in advance to resist quality degradation and avoid the huge latency overhead caused by connection interruption.

[0107] Network frequency band mechanism: Machine learning algorithms learn the historical performance of each frequency band (such as 5G C-band vs. millimeter wave) in different scenarios. For example, when it is predicted that a user will move into a millimeter wave coverage area, the high-speed millimeter wave frequency band will be selected first to minimize latency; when it is predicted that the user will move out of the coverage area, the system will smoothly switch to a lower frequency band with wider coverage in advance to prevent connection interruption and ensure the continuity of the control loop.

[0108] Step 11: The invention of the 5G cross-domain low-latency remote control system and method for explosion-proof humanoid robots can adapt to scenarios such as latency jitter and network packet loss caused by short-term network instability.

[0109] Testing was conducted in real-world 5G / 4G network and internet environments, covering various typical indoor and outdoor application scenarios to account for the impact of network switching and signal fluctuations. Timestamps were applied between the master control command input point and the slave actuator response point (or latency measurement tools were used on the closed-loop video stream) to directly measure the total latency from "brain" to "limbs." Test results showed that the latency for humanoid control and force feedback was within 100ms, and the latency for chassis control and sensors such as cameras was within 150ms. At this latency, the master operator experiences virtually no control delay, maximizing the safety and stability of remote driving and teleoperation.

[0110] This system implements remote master-slave force feedback teleoperation control. The core mechanism is that the operator controls the slave end (including the mobile chassis and humanoid robot) through the master end (including the teleoperation intelligent cockpit). Master-slave synchronization: Motion commands (such as position and posture) of the master end components are collected in real time and sent to the slave end controller, driving the slave end robotic arm to achieve synchronized movement. Slave force feedback: During task execution, the torque / force information generated by the slave end humanoid robot's contact with the environment is sensed and fed back to the master end force feedback device. Operator tactile perception: The feedback torque / force acts on the operator's hand, allowing them to truly perceive the interaction force between the slave end and the environment. Precise closed-loop operation: This force feedback assists the operator in making more precise adjustments, and by continuing to operate the master end control arm, finely controlling the slave end to complete the task.

[0111] A control panel is located in the legroom area of ​​the cockpit. This panel integrates key control interfaces: a mobile chassis control steering wheel, directly connected to the control panel, allowing the operator to steer the mobile chassis; and master / slave control accelerator and brake pedals, located on the control panel. By pressing the accelerator pedal, the operator can control the acceleration and deceleration of the remote humanoid robot chassis to achieve rapid movements over a wide range. Pressing the brake pedal stops the chassis movement. During this stop, the operator can use this brief interval to naturally adjust their sitting or operating posture (e.g., extend their legs, change arm position) to prepare for subsequent remote operations. This design makes human-machine interaction more flexible: the operator can directly control the chassis without getting up or leaving their seat, and can quickly and easily disconnect the master / slave coordination to adjust their posture when needed.

[0112] A display screen in front of the cockpit transmits real-time environmental information from multiple cameras and sensors on the remote humanoid robot back to the main display screen for display. This allows for real-time acquisition of images of the remote, real-world environment and information about the humanoid robot, facilitating precise operation by the operator. The mobile chassis is wheeled, significantly reducing energy consumption during the movement of the remote equipment and enabling rapid movement, thus improving the efficiency of remote operation.

[0113] 5G mobile communication technology, as the mainstream advanced communication standard today, represents a comprehensive iteration and significant leap forward from 4G technology. It aims to build a new generation of communication networks that are ultra-high-speed, extremely responsive, green and efficient, and have massive capacity, supporting the interconnection of massive numbers of devices. Its main advantages include: Ultra-high transmission rates: 5G can provide peak speeds several times, even tens of times, higher than 4G, with theoretical download speeds reaching several Gbps. This allows for near-instantaneous completion of ultra-high-definition video streaming and large file transfers, greatly enhancing the user experience. Ultra-low network latency: Air interface latency can be reduced to the millisecond (ms) level, significantly improving real-time interactivity and providing crucial support for applications with extremely high responsiveness requirements, such as remote surgery, industrial automation control, autonomous driving, and cloud gaming. Significantly optimized energy efficiency: Compared to previous generations, 5G, through more intelligent signaling mechanisms, network slicing technology, and optimized base station design, can significantly reduce device power consumption and extend terminal battery life while providing higher performance, demonstrating stronger environmental sustainability. Massive network bandwidth. By utilizing higher-frequency spectrum resources (such as millimeter waves) and advanced modulation techniques, 5G boasts bandwidth capacity far exceeding that of 4G, enabling it to simultaneously support more high-data-traffic applications and meet the high-bandwidth demands of densely populated areas. Massive concurrent device access is also supported. One of the design goals of 5G networks is to achieve the Internet of Things (IoT), with a dramatic increase in the number of connected terminal devices per cell, reaching a connection density of millions per km². This lays a solid connectivity foundation for building large-scale smart cities, smart homes, sensor networks, and more.

[0114] The Weak Network Resistance Algorithm is a communication optimization technology system designed for unstable network scenarios (high latency, packet loss, jitter). Its core objective is to maintain communication continuity, stability, and low latency through a dynamic adaptive mechanism. Essentially, it builds reliability on unreliable channels. Low-latency control across remote domains places even higher demands on the Weak Network Resistance Algorithm. This invention introduces machine learning algorithms to predict network conditions and intelligently and in real-time adjust key communication parameters (modulation method, power, frequency band) to optimize the balance between latency and energy efficiency under harsh environments. The optimization of the entire Weak Network Resistance Algorithm is completed through steps such as data collection and environmental perception, machine learning predictive modeling, and intelligent resource optimization decision-making.

[0115] This invention simplifies the protocol stack by removing redundant TCP / IP fields. This reduces bandwidth consumption and significantly saves bandwidth overhead in mobile networks. By removing redundant data such as TCP checksums and optional fields, the effective payload ratio is increased, transmission latency is reduced, device resource consumption is optimized, stability in weak network conditions is enhanced, and latency in cross-domain communication is effectively reduced, resulting in a significant improvement in communication efficiency.

[0116] As described above, although the invention has been shown and described with reference to specific preferred embodiments, it should not be construed as limiting the invention itself. Various changes in form and detail may be made without departing from the spirit and scope of the invention as defined in the appended claims.

Claims

1. A 5G cross-domain low-latency remote control system for an explosion-proof humanoid robot, characterized in that, include: The master-side remote-controlled smart cockpit is used by the operator to remotely send control commands and receive feedback information. A remote explosion-proof humanoid robot, including a mobile chassis and a humanoid robot, is used to execute control commands sent by the master remote-operated intelligent cockpit and to provide feedback information. A 5G-based communication system is used to establish a 5G communication channel between the master-end remotely operated intelligent cockpit and the slave-end remote explosion-proof humanoid robot for the transmission of control commands and feedback information. The humanoid robot and the mobile chassis adopt a separate hardware architecture, and their respective control commands and feedback information are transmitted through independent 5G communication channels. The communication system employs protocol optimization and a machine learning-based algorithm to combat weak network conditions. The protocol optimization includes using a custom lightweight UDP-based protocol to replace TCP, implementing a retransmission mechanism based on sequence numbers and selective acknowledgment for critical data at the application layer, compressing IP / UDP packet headers using the ROHC algorithm, and embedding CRC32 checksums in application layer data packets for end-to-end verification. The algorithm to combat weak network conditions uses machine learning to predict network congestion and channel status, and dynamically adjusts the modulation scheme, power, and frequency band. The master-end remote-operated intelligent cockpit includes: a seat for the operator to sit and operate; a remote control device for controlling the upper limb movements of the slave-end remote explosion-proof humanoid robot; a driving control device for controlling the movement and steering of the mobile chassis; a display for displaying environmental information and robot status transmitted by the slave-end remote explosion-proof humanoid robot in real time; a master-end intelligent terminal for data processing; and a master-end 5G router for 5G communication with the slave-end intelligent terminal. The remote-controlled explosion-proof humanoid robot includes: a four-wheel drive wheeled mobile chassis with omnidirectional mobility; a humanoid robot having a humanoid body, humanoid arms, and end grippers installed at the ends of the arms; a sensor assembly including a six-dimensional force sensor, a lidar, an IMU inertial measurement unit, and an odometer based on a wheel encoder installed at the end grippers; multiple cameras installed on the humanoid body and humanoid arms; a slave intelligent terminal including a humanoid robot intelligent terminal and a mobile chassis intelligent terminal; and a 5G communication module including a slave 5G router and a 5G module, wherein the slave 5G router is used for 5G communication between the humanoid robot intelligent terminal and the master intelligent terminal, and the 5G module is used for 5G communication between the mobile chassis intelligent terminal and the master intelligent terminal. The main intelligent terminal responds to the operation of the driving control device, converts the control commands into linear velocity and angular velocity information, and transmits them to the mobile chassis intelligent terminal through the 5G communication channel to control the movement and steering of the mobile chassis; at the same time, the main intelligent terminal obtains the joint motor position torque and gripper control information of the remote control device, and transmits it to the humanoid robot intelligent terminal through the 5G communication channel to drive the humanoid arms and end grippers to achieve synchronous movement; The humanoid robot intelligent terminal acquires force and torque information in real time from the six-dimensional force sensor installed at the end gripper. After low-pass filtering, the transmission delay and dynamic interference are offset by a bilateral control algorithm combined with a state observer and feedforward dynamic compensation. The force feedback results are transmitted to the master intelligent terminal through a 5G communication channel, and the remote control device provides force feedback to the operator to realize master-slave force feedback closed-loop operation. The bilateral control algorithm includes: using the master end pose as the slave end desired pose through the forward channel, using the force measured by the slave end as the master end feedback force through the reverse channel, and using the inverse kinematics algorithm to solve the model inverse solution; The impedance model used in the inverse kinematics algorithm is: , in, For the power of expectation, , , These are the desired inertia matrix, damping matrix, and stiffness matrix, respectively. , , These are position error, velocity error, and acceleration error, respectively; the damping matrix in the impedance model is adjusted. and stiffness matrix This allows the slave humanoid robot to exhibit different compliance characteristics in order to handle the differences in dynamic models between the master and slave ends; The state observer uses a Romberg observer, and its mathematical model is as follows: ， in It is the estimated state vector. This is the derivative of the estimated state, where A, B, and C are the system matrix, input matrix, and output matrix, respectively. It's a delay. The control input, It's a delay. The measurement output, L is the observer gain matrix; The formula for calculating the compensation term of the feedforward dynamic compensation is as follows: ,in, It is feedforward compensation torque. It is a joint position vector. It is the joint velocity vector. , These are the desired joint accelerations and velocities. It is the inertia matrix. It is the Coriolis and centrifugal force matrix. It is the gravity vector, and the calculated compensation term Used to compensate for inertial forces, including those of the humanoid robot itself. ,gravity Coriolis force and centrifugal force The dynamic effects included.

2. The explosion-proof humanoid robot 5G cross-domain low-latency remote control system according to claim 1, characterized in that, In the aforementioned split control architecture, the humanoid robot intelligent terminal uses an NVIDIA Jetson Nano processor to control the humanoid robot, while the mobile chassis intelligent terminal uses a Raspberry Pi processor to control the mobile chassis. Both communicate with the master intelligent terminal through a slave 5G router and a 5G module, respectively.

3. The explosion-proof humanoid robot 5G cross-domain low-latency remote control system according to claim 2, characterized in that, The communication system is used to realize cross-domain information interaction between the master-end remotely operated intelligent cockpit and the slave-end remote explosion-proof humanoid robot, including video stream transmission, sensor data feedback, control command transmission and force feedback; Different quality of service (QoS) flags are set for different data types. Among them, robot joint motor position torque data and sensor data are transmitted via UDP and are given priority over video streams transmitted via TCP.

4. The explosion-proof humanoid robot 5G cross-domain low-latency remote control system according to claim 3, characterized in that, The humanoid robot intelligent terminal transmits multiple camera video streams via H.264 encoding and WebRTC protocol, and sensor data, including those from the odometer, IMU inertial measurement unit, and lidar, are transmitted in real time via JSON packaging.

5. The explosion-proof humanoid robot 5G cross-domain low-latency remote control system according to claim 1, characterized in that, The custom lightweight UDP-based protocol eliminates the latency caused by TCP handshake, retransmission, and congestion control. It implements a retransmission mechanism based on sequence numbers and selective acknowledgment at the application layer. When the receiver detects a lost data packet, it only requests the retransmission of the lost data packet. It also embeds a CRC32 checksum at the application layer for end-to-end verification. If the receiver fails to verify a critical data packet, it triggers selective retransmission.

6. The explosion-proof humanoid robot 5G cross-domain low-latency remote control system according to claim 5, characterized in that, The proposed anti-weak network algorithm combines online supervised learning with deep reinforcement learning, models resource scheduling as a Markov decision process, uses real-time and predicted network states as state inputs, communication parameter adjustments as action outputs, and minimizes latency and maintains reliability as reward objectives. The state inputs include latency, packet loss rate, bandwidth, and signal strength, and the action outputs include modulation scheme adjustment, power control, and frequency band switching.

7. The method for the 5G cross-domain low-latency remote control system of an explosion-proof humanoid robot according to claim 1, characterized in that, Includes the following steps: S1: Establish master-slave communication connection and control architecture: A remotely operated intelligent cockpit is built on the main end and a main end intelligent terminal is set up. A remote explosion-proof humanoid robot is built on the slave end and a humanoid robot intelligent terminal and a mobile chassis intelligent terminal are set up. A communication connection based on a 5G network is established, and a separate control architecture is adopted to separate the control of the humanoid robot from the control of the mobile chassis. The main intelligent terminal and the humanoid robot intelligent terminal communicate through a 5G router, and the main intelligent terminal and the mobile chassis intelligent terminal communicate through a 5G module. S2: Execute chassis movement control: The main intelligent terminal converts the operation data of the driving control device into linear velocity and angular velocity information in real time, and transmits it to the mobile chassis intelligent terminal through the 5G communication channel. The mobile chassis intelligent terminal then controls the movement and steering of the mobile chassis. S3: Perform teleoperation control of the humanoid robot: The master intelligent terminal obtains the joint motor position, torque and gripper control information of the remote control device in real time, and transmits it to the slave humanoid robot through the 5G communication channel to realize the synchronous motion control of the two arms and gripper. S4: Perform video and sensor data transmission. The humanoid robot intelligent terminal collects video streams through multiple cameras and transmits them in real time to the main intelligent terminal for decoding and display; sensor data, including those from the odometer, IMU inertial measurement unit, and lidar, are packaged in JSON format and transmitted in real time to the main intelligent terminal via a 5G communication channel for environmental information analysis. S5: Implementation force feedback processing: The humanoid robot intelligent terminal acquires the position, torque and six-dimensional force sensor information of the humanoid joint motor in real time, applies a low-pass filter to eliminate high-frequency noise, calculates the force feedback result through a bilateral control algorithm, transmits it to the master intelligent terminal through a 5G communication channel, and provides the force feedback feeling to the operator through the remote control device to realize master-slave force feedback closed-loop operation. S6: Optimization of Execution Protocol and Resource Scheduling: The transport protocol is optimized at the protocol layer, including using a custom lightweight UDP-based protocol to replace TCP, implementing a retransmission mechanism based on sequence number and selective acknowledgment for critical data at the application layer, using the ROHC algorithm to compress IP / UDP packet headers, and embedding CRC32 checksums in application layer data packets for end-to-end verification to reduce transmission overhead and latency. In the resource scheduling and management layer, a machine learning-based anti-weak network algorithm is introduced. By predicting network congestion and channel status, the modulation scheme, power and frequency band are dynamically adjusted to balance latency and energy efficiency.

Images