A micro-ros mobile robot embedded control system

Abstract

This invention discloses an embedded control system for a micro-ROS mobile robot, belonging to the field of robot control system technology. The control system includes a real-time priority communication scheduling module. This module dynamically adjusts the scheduling polling cycle and token generation rate based on a combination of the token bucket algorithm and a priority queue to solve the latency jitter problem in multi-node concurrent communication. This invention, through real-time priority communication scheduling, achieves hierarchical allocation of bandwidth and priority according to message type, ensuring low latency for hard real-time control commands. It solves the latency jitter caused by contention between high and low priority messages in traditional CAN / UART buses, achieving end-to-end low latency, predictable, and reliable control response.

Description

Technical Field

[0001] This invention relates to the field of robot control system technology, specifically to an embedded control system for a micro-ROS mobile robot. Background Technology

[0002] Traditional mobile robot control systems are mostly based on PCs or high-performance embedded platforms, running ROS2 or self-developed middleware, and achieving inter-node communication via Ethernet or CANBus. However, as robots develop towards miniaturization, low cost, and low power consumption, resource-constrained MCU platforms (such as ARM Cortex-M) struggle to simultaneously meet the requirements of real-time performance, reliability, and battery life. On the one hand, the centralized communication architecture of ROS2 is prone to causing critical message congestion on low-bandwidth buses. On the other hand, there is a lack of unified and dynamic solutions for issues such as node overload, fault switching, and power management, resulting in significant performance degradation of the system in complex field environments.

[0003] Patent CN111813138B discloses a real-time embedded control system for a legged mobile robot. The patent enables centralized control of the robot and enhances its stability.

[0004] The aforementioned patent enables real-time acquisition of joint and torso information of a legged mobile robot, and calculates the robot's motion based on the remote control input, achieving centralized control of the robot and enhancing its stability. However, it also causes latency jitter when multiple nodes communicate concurrently, making it impossible to effectively allocate bandwidth and priority, resulting in high latency of real-time control commands.

[0005] To address this, this application proposes an embedded control system for micro-ROS-based mobile robots that can hierarchically allocate bandwidth and priority according to message type. Summary of the Invention

[0006] The purpose of this invention is to provide an embedded control system for micro-ROS mobile robots to solve the technical problem mentioned in the background art of difficulty in simultaneously meeting the requirements of real-time performance, reliability and battery life.

[0007] To achieve the above objectives, the present invention provides the following technical solution: an embedded control system for a micro-ROS mobile robot, the control system comprising:

[0008] The main control unit of a resource-constrained embedded real-time operating system integrates a micro-ROS Master node;

[0009] Multiple micro-ROS Client nodes are deployed in the sensor acquisition unit, motion drive unit, and power management unit, respectively.

[0010] Each Client node is equipped with a real-time priority communication scheduling module, which is used to dynamically allocate bandwidth and priority according to the preset Topic QoS requirements to ensure that the end-to-end latency does not exceed 5ms.

[0011] The Client node transmits messages via CAN Bus and UART Bus, and has a built-in CRC-16 data verification and retransmission mechanism;

[0012] The real-time priority communication scheduling module is based on a combination of the token bucket algorithm and the priority queue, and dynamically adjusts the scheduling polling period and the token generation rate to solve the problem of latency jitter in multi-node concurrent communication.

[0013] Preferably, the micro-ROS Client node further includes an adaptive load balancing module, which is used for:

[0014] Real-time monitoring of node CPU utilization, memory usage, and message publishing and subscription frequency for each Topic;

[0015] Based on the online transfer learning algorithm Dyna-Q, short-term load trends are predicted, and computing and communication tasks are dynamically redistributed among client nodes in a strategy iteration manner.

[0016] By coordinating load transfer timing through heartbeat synchronization messages between nodes, the overall resource utilization rate is guaranteed to be ≥80%, and the end-to-end latency fluctuation does not exceed ±1ms.

[0017] Preferably, the real-time priority communication scheduling module specifically includes:

[0018] The scheduling submodule based on the EDF earliest deadline first algorithm is used to perform preemptive scheduling on high-priority control topics.

[0019] A soft real-time data submodule based on polling priority (WRR) is used to periodically send low-priority data.

[0020] The dynamic QoS adjustment unit modifies the reliability parameters and historical maximum transmission interval of the Topic in real time based on network congestion.

[0021] Preferably, in the adaptive load balancing module, the online learning algorithm further includes:

[0022] The status indicator describes the node load in the form of triples, which include: CPU utilization, memory usage, and message queue length.

[0023] The reward function is designed as a weighted sum of the system's average response time and resource utilization. The maximum positive reward is given when the response time is ≤5ms and the utilization rate is ≥80%.

[0024] The action set includes operations such as "execute locally", "migrate to an adjacent node", and "reduce load and enter hibernation" to ensure both node load balancing and system real-time performance.

[0025] Preferably, the micro-ROS Client node further includes a fault detection and redundancy switching module, used for:

[0026] Send heartbeat packets at a time interval of T=50ms;

[0027] If no response is received for N=3 consecutive heartbeat cycles or the CRC check fails, the node is considered faulty.

[0028] The standby node maintains the mirror subscription status according to the topic, takes over the functions of the failed node, and triggers the Master node to update the Topic routing table.

[0029] Preferably, the fault detection and redundancy switching module further includes:

[0030] A fault log circular buffer records M=100 past fault events in timestamp format;

[0031] The JSON-based reporting unit sends messages to the main control unit via a secure MQTT channel to support remote online diagnostics.

[0032] Preferably, the micro-ROS Client node further includes an energy-aware communication management module, which is used for:

[0033] Real-time acquisition of battery voltage, current and power consumption data, and acquisition of measurement values ​​with an accuracy of 0.1% through ADC interface;

[0034] The message sending frequency is automatically adjusted based on the remaining battery threshold and the priority of each Topic.

[0035] When the remaining battery power is less than 20%, the system enters a low-power mode, retaining only high-priority communication, while other topics adopt a periodic wake-up strategy to ensure continuous operation for at least 2 hours.

[0036] Preferably, the main control unit adopts an ARM Cortex-M7 architecture MCU with a main frequency of 480MHz, built-in 2MB Flash and 512KB SRAM, and reserves an external SPI Flash interface for storing micro-ROS image and log data.

[0037] Preferably, the control system further includes a security encryption module, which is used for:

[0038] Topic messages transmitted between client nodes are encrypted using AES-128 CTR mode;

[0039] The Master and Client communicate via the ECDH algorithm to negotiate the key and periodically refresh the session key using SHA-256 to ensure secure communication.

[0040] Preferably, the control system further includes an OTA update module, which is used for:

[0041] Pull the differential firmware package from the remote server via HTTPS;

[0042] Perform foreground download and background decompression when the workload is low;

[0043] After the update is complete, the system will automatically switch to the new firmware by restarting the micro-ROS Master node and verifying the new image signature.

[0044] Compared with the prior art, the beneficial effects of the present invention are:

[0045] 1. This invention achieves hierarchical allocation of bandwidth and priority according to message type through real-time priority communication scheduling, ensuring low latency of hard real-time control commands, solving the latency jitter caused by contention between high and low priority messages in traditional CAN / UART buses, and realizing end-to-end low latency, predictable and reliable control response.

[0046] 2. This invention achieves online learning load prediction and task migration based on Dyna-Q through online adaptive load balancing, monitors node resource usage in real time, and dynamically reallocates computing and communication tasks among nodes. This solves the performance bottleneck and real-time control failure caused by node overload under fixed scheduling strategies, improves the overall resource utilization of the system, reduces latency fluctuations, and enhances operational stability.

[0047] 3. This invention achieves automatic detection of node disconnection and data verification failure through fault detection and redundancy switching, solves the control interruption and data loss caused by field environmental interference or hardware failure, improves system robustness, ensures control flow continuity, and reduces downtime and manual intervention.

[0048] 4. This invention achieves automatic adjustment of the publishing frequency of different priority topics and node sleep strategies based on the remaining power through energy consumption sensing communication management, solving the problem of mobile robots struggling to balance low power consumption and real-time requirements. It improves battery life while ensuring critical communication, meeting the needs of long-term field operations. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the control system framework of the present invention. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] Please see Figure 1 The present invention provides an embodiment of an embedded control system for a micro-ROS mobile robot; the control system further includes a security encryption module, which is used for:

[0052] Topic messages transmitted between client nodes are encrypted using AES-128 CTR mode;

[0053] The Master and Client complete key negotiation through the ECDH algorithm and periodically refresh the session key based on SHA-256 to ensure communication security.

[0054] Further, system initialization and Master-Client configuration:

[0055] Hardware power-on self-test: After the Master node MCU is powered on, it first executes the Bootloader to perform a self-test of Flash, SRAM, and peripheral interfaces (CAN, UART, ADC); the Client node synchronously performs a self-test of its local resources and communicates via / hb_<node_id> The topic sends a heartbeat packet to the Master, carrying the node type and capability description (such as the number of CPU cores and available memory).

[0056] Micro-ROS network topology construction: The Master node starts the micro-ROS Agent and listens for registration requests from all Client nodes. The Client nodes call rclc_executor_init() and rclc_node_init() in sequence to send registration messages on the / node_register topic. After receiving the registration, the Master registers the node ID, Topic relationship list and QoS configuration in the internal routing table.

[0057] Key negotiation and secure channel establishment: The Master and each Client exchange public keys using the ECDH algorithm to generate a shared session key; the session key is initially hashed based on SHA-256 and used as the AES-128 CTR mode encryption key; all subsequent Topic messages are sent after being encrypted with AWS-128 CTR to ensure communication privacy and tamper-proof protection.

[0058] Once the system is started, it can form a predictable and secure ROS communication network, laying the foundation for real-time control.

[0059] Please see Figure 1 The present invention provides an embodiment of an embedded control system for a micro-ROS mobile robot; the control system includes:

[0060] The main control unit of a resource-constrained embedded real-time operating system integrates a micro-ROS Master node;

[0061] Multiple micro-ROS Client nodes are deployed in the sensor acquisition unit, motion drive unit, and power management unit, respectively.

[0062] Each Client node is equipped with a real-time priority communication scheduling module, which is used to dynamically allocate bandwidth and priority according to the preset Topic QoS requirements to ensure that the end-to-end latency does not exceed 5ms.

[0063] The Client node transmits messages via CAN Bus and UART Bus, and has a built-in CRC-16 data verification and retransmission mechanism;

[0064] The real-time priority communication scheduling module is based on a combination of the token bucket algorithm and the priority queue, and dynamically adjusts the scheduling polling period and the token generation rate to solve the problem of latency jitter in multi-node concurrent communication.

[0065] The real-time priority communication scheduling module specifically includes:

[0066] The scheduling submodule based on the EDF earliest deadline first algorithm is used to perform preemptive scheduling on high-priority control topics.

[0067] A soft real-time data submodule based on polling priority (WRR) is used to periodically send low-priority data.

[0068] The dynamic QoS adjustment unit modifies the reliability parameters and historical maximum transmission interval of the Topic in real time based on network congestion.

[0069] Topic hierarchy and scheduling strategy:

[0070] High priority: / emergency_stop, / cmd_vel, with end-to-end latency ≤5ms for both hardware and software real-time requirements;

[0071] Medium priority: / odom, / sensor_data, latency ≤10ms;

[0072] Low priority: / status_report, latency ≤100ms;

[0073] Token bucket + EDF (Earliest Deadline First) + WRR hybrid scheduling:

[0074] In each communication cycle T=1ms, a fixed number of tokens are allocated to high-priority topics. If there are not enough tokens, they are immediately preempted. The remaining bandwidth is used to run the EDF earliest deadline priority submodule to schedule medium-priority data before the deadline. The WRR submodule sends low-priority messages in the idle window. The scheduling parameters (token generation rate, EDF earliest deadline priority emergency threshold) are dynamically adjusted according to the error frame count on the bus.

[0075] Real-time process example:

[0076] t=0ms: The Master issues the / cmd_vel command;

[0077] t=0.1ms: The sensor node receives the command and publishes / cmd_vel_ack;

[0078] t=0.5ms: The motion drive node converts the control signal into a PWM waveform and responds immediately;

[0079] The overall end-to-end latency is controlled within the range of 2-4ms;

[0080] Achieve predictable low latency from end to end, avoiding delays or loss of critical control commands.

[0081] Please see Figure 1 One embodiment of the present invention provides: an embedded control system for a micro-ROS mobile robot; the micro-ROS Client node further includes an adaptive load balancing module, which is used for:

[0082] Real-time monitoring of node CPU utilization, memory usage, and message publishing and subscription frequency for each Topic;

[0083] Based on the online transfer learning algorithm Dyna-Q, short-term load trends are predicted, and computing and communication tasks are dynamically redistributed among client nodes in a strategy iteration manner.

[0084] By coordinating load transfer timing through heartbeat synchronization messages between nodes, the overall resource utilization rate is guaranteed to be ≥80%, and the end-to-end latency fluctuation does not exceed ±1ms.

[0085] The online learning algorithm in the adaptive load balancing module further includes:

[0086] The status indicator describes the node load in the form of triples, which include: CPU utilization, memory usage, and message queue length.

[0087] The reward function is designed as a weighted sum of the system's average response time and resource utilization. The maximum positive reward is given when the response time is ≤5ms and the utilization rate is ≥80%.

[0088] The action set includes operations such as "execute locally", "migrate to an adjacent node", and "reduce load and enter hibernation" to ensure both node load balancing and system real-time performance.

[0089] Adaptive load balancing:

[0090] State monitoring and learning model: CPU utilization, memory usage, and message queue length are collected every 10ms; state triples are fed into the Dyna-Q algorithm, and the Q table is updated online in combination with historical data;

[0091] Task migration and reallocation: If a node's CPU usage is greater than 90% and its average latency is greater than 6ms, the "migrate to an adjacent node" action is triggered; the target node is notified to receive the task via the / node_load heartbeat topic, and the micro-ROSService call route is dynamically updated;

[0092] Real-time process example: The sensor node detects that the image preprocessing task has caused the CPU utilization to be 95%. The online learning model recommends migrating some preprocessing algorithms to the driver node. The driver node starts the corresponding service, starts receiving and processing image frames, and the overall system utilization is restored to approximately 85%.

[0093] Ensure the real-time performance of critical tasks and the balanced utilization of system resources under high load scenarios.

[0094] Please see Figure 1 One embodiment of the present invention provides: an embedded control system for a micro-ROS mobile robot; the micro-ROS Client node further includes a fault detection and redundancy switching module, used for:

[0095] Heartbeat packets are sent at a period of T=50ms;

[0096] If no response is received for N=3 consecutive heartbeat cycles or the CRC check fails, the node is considered faulty.

[0097] The standby node maintains the mirror subscription status according to the topic, takes over the functions of the failed node, and triggers the Master node to update the Topic routing table;

[0098] The fault detection and redundancy switching module further includes:

[0099] A fault log circular buffer records M=100 past fault events in timestamp format;

[0100] The JSON-formatted reporting unit sends messages to the main control unit via a secure MQTT channel to support remote online diagnostics.

[0101] Fault detection and redundancy switching:

[0102] Periodic heartbeat detection: A fault is determined when there is no response or CRC check failure after 3 consecutive heartbeats every 50ms.

[0103] Redundant node takeover: The backup node listens to the / hb_backup topic for a long time to maintain the functional image; after a failure occurs, it immediately subscribes to all topics of the failed node and replaces the node ID in the Master routing table; at the same time, it writes the failure event (timestamp, node ID, failure type) to the ring cache and retains the most recent 100 records.

[0104] Real-time process example: The motion drive node fails due to external impact and loses its heartbeat; the backup drive node completes the switchover within 150ms, the control flow continues seamlessly, and motion commands are not lost; the system log records the switchover process for easy post-event diagnosis.

[0105] Significantly improves the robustness and reliability of robots in complex environments.

[0106] Please see Figure 1 One embodiment of the present invention provides: an embedded control system for a micro-ROS mobile robot; the micro-ROS Client node further includes an energy-sensing communication management module, which is used for:

[0107] Real-time acquisition of battery voltage, current and power consumption data, and acquisition of measurement values ​​with an accuracy of 0.1% through ADC interface;

[0108] The message sending frequency is automatically adjusted based on the remaining battery threshold and the priority of each Topic.

[0109] When the remaining battery power is less than 20%, the system enters a low-power mode, retaining only high-priority communication, while other topics adopt a periodic wake-up strategy to ensure continuous operation for at least 2 hours.

[0110] Furthermore, power acquisition and policy thresholds: The ADC acquires battery voltage and current in real time with an accuracy of 0.1%; power thresholds can be set at 80%, 50%, and 20%.

[0111] Dynamic communication adjustment: When the battery level is ≥50%, maintain the normal frequency of all topics; when the battery level is 20% ≤ 50%, reduce the frequency of low-priority topics to 50%; when the battery level is <20%, only high-priority topics are retained, and low-priority topics adopt a 1-second wake-up strategy.

[0112] Real-time process example: When the battery level drops from 55% to 45%, the system automatically lowers the / status_report frequency; the battery life is extended from 1.5 hours to 2.1 hours.

[0113] While ensuring critical control, extend the robot's battery life to balance real-time performance and battery life requirements.

[0114] The following supplementary explanation is provided regarding the implementation method:

[0115] Main control unit: Processor: ARM Cortex-M7, 480MHz, 2MB Flash, 512KB SRAM; Peripherals: CAN transceiver (ISO 11898), UART adapter module, SPI Flash (for storing logs and firmware), ADC (for acquiring battery voltage and current).

[0116] Client Node: Sensor Acquisition Unit: Includes peripherals such as IMU and LiDAR, connected to the MCU via SPI / I2C; Motion Drive Unit: PWM drive motor controller, outputting PWM via GPIO and TIM timer; Power Management Unit: Power detection circuit (voltage divider + current sampling resistor), ADC interface acquisition.

[0117] Communication bus: ACN Bus: 500kbps, using CRC-15 check; UART: 1Mbps, including CRC-16 data check and retransmission mechanism.

[0118] System integration and testing:

[0119] Simulation verification: Load the robot model in Gazebo and use the micro-ROS plugin to simulate network latency and packet loss; adjust the scheduling parameters so that high-priority commands can still arrive within ≤5ms in the simulation environment.

[0120] Hardware testing: Monitor CAN message timing using a logic analyzer; perform path tracing and emergency stop tests in a real-world environment to verify the effectiveness of fault switching and battery life management.

[0121] Performance metrics: Control latency: End-to-end for high-priority commands ≤5ms; Resource utilization: System load maintained between 80% and 90%; Fault switching time: ≤150ms; Improved battery life: 1.5h in normal mode - ≥2h in low-power mode.

[0122] Optional implementation methods and equivalent variations:

[0123] In scheduling algorithms, the token bucket can be replaced by a leaky bucket;

[0124] Load balancing learning algorithms can include SARSA and deep Q-learning;

[0125] The fault switching heartbeat cycle and threshold can be adjusted according to the application scenario;

[0126] The communication bus can also be expanded to CAN FD or Ethernet TSN.

[0127] Working principle: This system consists of a main control unit integrating a micro-ROS Master node and several Client nodes running on a resource-constrained embedded RTOS. The Master node is responsible for network topology management, Topic routing and security key negotiation; the Client nodes carry functional modules such as sensor acquisition, motion driving, and power management, and communicate with the Master and other Client nodes on CAN Bus / UART through the micro-ROS middleware.

[0128] The client node has a built-in real-time priority communication scheduling engine, which adopts a token bucket + EDF earliest deadline priority + WRR hybrid strategy to schedule command control, high-speed feedback and status reporting according to QoS level, ensuring that the end-to-end latency of high-priority messages is controllable within 5ms, and to cope with bus congestion through dynamic QoS parameter adjustment.

[0129] During system operation, Client nodes report load, heartbeat, and power information in real time. Master nodes make load balancing decisions based on online learning algorithms and trigger redundancy switching or task migration when nodes fail or are overloaded. Meanwhile, the energy consumption sensing module dynamically adjusts the communication frequency and sleep strategy based on the remaining power to extend battery life while ensuring real-time performance.

[0130] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. An embedded control system for a micro-ROS mobile robot, characterized in that: The control system includes: The main control unit of a resource-constrained embedded real-time operating system integrates a micro-ROS Master node; Multiple micro-ROS Client nodes are deployed in the sensor acquisition unit, motion drive unit, and power management unit, respectively. Each Client node is equipped with a real-time priority communication scheduling module, which is used to dynamically allocate bandwidth and priority according to the preset Topic QoS requirements to ensure that the end-to-end latency does not exceed 5ms. The Client node transmits messages via CAN Bus and UART Bus, and has a built-in CRC-16 data verification and retransmission mechanism; The real-time priority communication scheduling module is based on a combination of the token bucket algorithm and the priority queue, and dynamically adjusts the scheduling polling period and the token generation rate to solve the problem of latency jitter in multi-node concurrent communication. The micro-ROS Client node further includes an adaptive load balancing module, which is used for: Real-time monitoring of node CPU utilization, memory usage, and message publishing and subscription frequency for each Topic; Based on the online transfer learning algorithm Dyna-Q, short-term load trends are predicted, and computing and communication tasks are dynamically redistributed among client nodes in a strategy iteration manner. By coordinating load transfer timing through heartbeat synchronization messages between nodes, the overall resource utilization rate is guaranteed to be ≥80%, and the end-to-end latency fluctuation does not exceed ±1ms. The online learning algorithm in the adaptive load balancing module further includes: The status indicator describes the node load in the form of triples, which include: CPU utilization, memory usage, and message queue length. The reward function is designed as a weighted sum of the system's average response time and resource utilization. The maximum positive reward is given when the response time is ≤5ms and the utilization rate is ≥80%. The action set includes "local execution", "migrate to adjacent node", and "load reduction and enter hibernation" operations to ensure both node load balancing and system real-time performance. The micro-ROS Client node also includes an energy-aware communication management module, which is used for: Real-time acquisition of battery voltage, current and power consumption data, and acquisition of measurement values ​​with an accuracy of 0.1% through ADC interface; The message sending frequency is automatically adjusted based on the remaining battery threshold and the priority of each Topic. When the remaining battery power is less than 20%, the system enters a low-power mode, retaining only high-priority communication, while other topics adopt a periodic wake-up strategy to ensure continuous operation for at least 2 hours. The control system further includes a security encryption module, which is used for: Topic messages transmitted between client nodes are encrypted using AES-128 CTR mode; The Master and Client communicate via the ECDH algorithm to negotiate the key and periodically refresh the session key using SHA-256 to ensure secure communication.

2. The embedded control system for a micro-ROS mobile robot according to claim 1, characterized in that: The real-time priority communication scheduling module specifically includes: The scheduling submodule based on the EDF earliest deadline first algorithm is used to perform preemptive scheduling on high-priority control topics. A soft real-time data submodule based on polling priority (WRR) is used to periodically send low-priority data. The dynamic QoS adjustment unit modifies the reliability parameters and historical maximum transmission interval of the Topic in real time based on network congestion.

3. The embedded control system for a micro-ROS mobile robot according to claim 1, characterized in that: The micro-ROS Client node also includes a fault detection and redundancy switching module, used for: Send heartbeat packets at a time interval of T=50ms; If no response is received for N=3 consecutive heartbeat cycles or the CRC check fails, the node is considered faulty. The standby node maintains the mirror subscription status according to the topic, takes over the functions of the failed node, and triggers the Master node to update the Topic routing table.

4. The embedded control system for a micro-ROS mobile robot according to claim 3, characterized in that: The fault detection and redundancy switching module further includes: A fault log circular buffer records M=100 past fault events in timestamp format; The JSON-based reporting unit sends messages to the main control unit via a secure MQTT channel to support remote online diagnostics.

5. The embedded control system for a micro-ROS mobile robot according to claim 1, characterized in that: The main control unit adopts an ARM Cortex-M7 architecture MCU with a main frequency of 480MHz, built-in 2MB Flash and 512KB SRAM, and reserves an external SPI Flash interface for storing micro-ROS image and log data.

6. The embedded control system for a micro-ROS mobile robot according to claim 1, characterized in that: The control system further includes an OTA update module, which is used for: Pull the differential firmware package from the remote server via HTTPS; Perform foreground download and background decompression when the workload is low; After the update is complete, the system will automatically switch to the new firmware by restarting the micro-ROS Master node and verifying the new image signature.

Images