Self-adaptive grabbing fault-tolerant control system for chemical laboratory automation

Abstract

The invention discloses a self-adaptive grabbing fault-tolerant control system for chemical laboratory automation, and relates to the technical field of laboratory automation control. The system comprises a hardware execution system and an intelligent control system, the hardware execution system is provided with a self-adaptive grabbing tail end module, a multi-degree-of-freedom motion module and a multi-mode sensing module, and the intelligent control system is provided with a self-adaptive grabbing control unit and a fault-tolerant identification and decision-making unit. The multi-mode sensing module collects the characteristics and environment data of a grabbed target, the self-adaptive grabbing control unit dynamically adjusts grabbing parameters and motion parameters according to fused data, and precise self-adaptive grabbing of targets of different materials, sizes and postures is achieved. And the fault-tolerant identification and decision-making unit monitors the operation state in real time, identifies four faults of grabbing, movement, sensor and environment, and executes graded fault-tolerant processing according to grades. According to the method, the adaptability, the fault handling capacity and the operation safety of the grabbing operation of the laboratory are remarkably improved through self-adaptive adjustment and a grading fault-tolerant mechanism.

Description

Technical Field

[0001] This invention relates to the field of laboratory automation control technology, and more specifically, to an adaptive grasping fault-tolerant control system for chemical laboratory automation. Background Technology

[0002] With the continuous improvement of laboratory automation, grasping and transport operations have become a key hub connecting core experimental processes such as sample processing, reagent addition, and reaction incubation. Laboratory scenarios are characterized by diverse and complex grasping targets and harsh environments, involving various objects such as glass reagent bottles, plastic centrifuge tubes, and irregularly shaped reaction containers. They also often face complex operating conditions such as corrosive reagents, fragile materials, and high and low temperatures, which places extremely high demands on the environmental adaptability and fault response capabilities of the grasping control system.

[0003] Existing laboratory gripping equipment mostly adopts a fixed parameter control mode, relying on preset programs to set gripping force, clamping angle, and motion trajectory, lacking the ability to dynamically adapt to changes in the gripped target and environment. On the one hand, manual adjustment of clamping parameters is required for gripping targets of different sizes, materials, and weights. When dealing with irregularly shaped containers or slightly deformed objects, problems such as unstable gripping and slippage easily occur, indicating insufficient adaptive capability. On the other hand, its fault tolerance mechanism is simplistic, relying only on simple sensors for basic obstacle avoidance, lacking the ability to identify and respond to various faults during the gripping process, such as force overload, container displacement, sensor malfunction, and environmental interference. Once a sudden anomaly occurs, it often shuts down immediately, seriously affecting the continuity of experiments and operational efficiency. In addition, existing equipment still has significant deficiencies in terms of protection against corrosive environments, gripping accuracy control, and real-time response speed, making it difficult to meet the urgent needs of modern laboratories for highly reliable and safe automated operations. Summary of the Invention

[0004] To address the aforementioned technical problems in related technologies, this invention proposes an adaptive grasping fault-tolerant control system for chemical laboratory automation, which can overcome the above-mentioned shortcomings of existing technologies.

[0005] To achieve the above-mentioned technical objectives, the technical solution of the present invention is implemented as follows: An adaptive grasping fault-tolerant control system for chemical laboratory automation; This adaptive grasping fault-tolerant control system for chemical laboratory automation includes: The hardware execution system includes an adaptive grasping end effector module, a multi-degree-of-freedom motion module, and a multi-modal perception module. The intelligent control system includes an adaptive grasping control unit and a fault-tolerant identification and decision-making unit; The multimodal perception module is configured to collect characteristic data and / or environmental data of the target being grasped; The adaptive grasping control unit is communicatively connected to the multimodal sensing module, the adaptive grasping end module, and the multi-degree-of-freedom motion module, and is configured to dynamically adjust the grasping parameters of the adaptive grasping end module and / or the motion parameters of the multi-degree-of-freedom motion module based on the data collected by the multimodal sensing module. The fault-tolerant identification and decision-making unit is configured to identify faults in the grasping process based on feedback data from the multimodal perception module and / or the adaptive grasping control unit, and to perform corresponding graded fault-tolerant handling according to the fault level.

[0006] Furthermore, the adaptive grasping end module includes: The three-jaw linkage clamping structure is provided with three sets of flexible clamping arms, and each set of flexible clamping arms is equipped with a micro force sensor, a displacement sensor and a torque sensor. A quick-release flexible gripper is switchably mounted on the end of the flexible gripping arm and includes a silicone gripper, a fluororubber gripper, and a honeycomb anti-slip rubber gripper. The surface of the quick-release flexible gripper has a plasma-modified coating and has an adaptive deformation range of 0-5mm. A dual-mode clamping force adjustment mechanism is configured to drive the flexible gripper to open and close, supporting switching between constant force clamping mode and constant stroke clamping mode. The multi-degree-of-freedom motion module includes a six-degree-of-freedom robotic arm, which has linear motion degrees of freedom along the X, Y, and Z axes, as well as rotational degrees of freedom around the X, Y, and Z axes. The six-degree-of-freedom robotic arm uses a servo motor and a precision ball screw drive, and is equipped with an absolute encoder and a torque sensor for closed-loop feedback. The multi-degree-of-freedom motion module also has a built-in dynamic anti-shake and inertia compensation algorithm, configured to dynamically adjust the driving torque according to the load inertia.

[0007] Furthermore, the multimodal sensing module includes: A vision sensor configured to identify the position, size, material, orientation, and surface condition of the target being grasped; A laser rangefinder sensor is configured to detect the distance to obstacles and the target capture distance. An environmental sensor is configured to collect data on corrosive gas concentration, temperature, and humidity. An infrared temperature sensor is configured to detect the surface temperature of the target being grasped. The self-cleaning unit is configured to purge the sensor probe with high-pressure dry nitrogen gas; The system also includes a force sensor and a torque sensor disposed within the flexible clamping arm, which together form part of the multimodal sensing module.

[0008] Furthermore, it also includes: Environmental protection module, the environmental protection module includes: Labyrinth-type sealing structure and fluororubber sealing ring are installed at the joints of the robotic arm; The nitrogen purging device linked to the environmental sensor is configured to adaptively adjust the nitrogen flow rate according to the concentration of corrosive gases to form positive pressure protection; An electrical control box, which is equipped with a dehumidification module, an anti-corrosion filter module and a temperature regulation unit, and the surface of the box is coated with polytetrafluoroethylene. The security protection module includes: The emergency braking unit is configured with dual triggering of hardware mechanical braking and software electronic braking; The reagent leakage detection unit is configured with dual detection of electrochemical sensors and visual recognition. The overload protection unit is configured to monitor the load through a force sensor, torque sensor, and motor current sensor working together. An anti-tipping unit, comprising an IMU inertial sensor configured to monitor the container's attitude; An explosion-proof unit, the explosion-proof unit including explosion-proof electrical components.

[0009] Furthermore, the intelligent control system adopts a three-level architecture of host computer-edge node-actuator. The edge node is integrated into the multi-degree-of-freedom motion module and the multi-modal sensing module, and communicates with the fieldbus via dual links through industrial Ethernet.

[0010] Furthermore, the adaptive grasping control unit has a built-in three-dimensional mapping model of target characteristics, clamping parameters, and environmental parameters, and integrates fuzzy PID algorithm and reinforcement learning algorithm. The adaptive grasping control unit is configured to acquire the size, material, weight, posture, surface state, temperature, and ambient gas concentration of the grasping target through the multimodal perception module, call the preset parameter database to match the initial grasping parameters, and dynamically adjust the clamping force, clamping angle, movement speed, and acceleration based on the force-displacement-torque-visual four-dimensional feedback data during the grasping process using the fuzzy PID algorithm. The reinforcement learning algorithm uses grasping success rate, clamping stability, and operation efficiency as reward and punishment objectives, iteratively optimizes the clamping parameters of similar targets, and generates personalized parameter templates.

[0011] Furthermore, the fault-tolerant identification and decision-making unit is configured to execute a fully closed-loop fault-tolerant process, including fault prediction, real-time identification, graded handling, and post-event optimization. The fault prediction uses a trend analysis algorithm to provide early warning of sensor data drift and clamping force attenuation. The fault types identified in real time include at least grasping faults, motion faults, sensor faults, and environmental faults. The tiered treatment includes: Level 1 fault tolerance enables self-healing of minor faults through parameter correction, self-cleaning, or trajectory fine-tuning. Level 2 fault tolerance: For moderate faults, continuous operation is maintained by switching to backup parameter schemes, activating redundant sensors or auxiliary power units. Level 3 fault tolerance triggers emergency braking, activates protective devices, and records fault data in response to major faults.

[0012] Furthermore, the intelligent control system also includes a multi-source data fusion unit, which is configured to use Kalman filtering to remove sensor measurement noise, use an attention mechanism to perform weighted fusion of data from different operation stages, and use an evidence theory model to cross-validate the multi-source data; the data processing delay of the multi-source data fusion unit is ≤5ms, and it supports automatic switching to redundant sensor channels when data is abnormal.

[0013] Furthermore, the intelligent control system also includes a safety monitoring and feedback unit, which is configured to monitor and record capture operation parameters, fault handling details, equipment status and operation videos in real time, and generate operation logs that meet GLP / GMP requirements; the safety monitoring and feedback unit is also equipped with hierarchical alarm, remote control, hierarchical management of operation permissions and equipment health diagnosis functions.

[0014] Furthermore, the intelligent control system also includes a safety monitoring and feedback unit, which is configured to monitor and record capture operation parameters, fault handling details, equipment status and operation videos in real time, and generate operation logs that meet GLP / GMP requirements; the safety monitoring and feedback unit is also equipped with hierarchical alarm, remote control, hierarchical management of operation permissions and equipment health diagnosis functions.

[0015] The beneficial effects of this invention are as follows: Through the coordinated operation of the multimodal perception module and the adaptive grasping control unit, precise parameter adaptation is achieved for grasping targets of different materials, sizes, and postures. Diverse grasping operations can be completed without manual intervention, thus significantly improving the system's environmental adaptability and versatility. A fully closed-loop hierarchical fault-tolerant system is constructed through the fault-tolerant identification and decision-making unit, employing self-healing, adjustment, and emergency response strategies for minor, moderate, and major faults respectively. This effectively avoids operational interruptions caused by single faults, thereby greatly improving the continuity and reliability of grasping operations. The integrated design of the environmental protection module and the safety assurance module forms a multi-dimensional active protection mechanism, effectively reducing safety risks such as corrosion, reagent leakage, and container damage, thereby comprehensively improving the safety level of automated laboratory operations. Detailed Implementation

[0016] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the scope of protection of the present invention.

[0017] An adaptive grasping fault-tolerant control system for chemical laboratory automation according to an embodiment of the present invention includes: The hardware execution system includes an adaptive grasping end effector module, a multi-degree-of-freedom motion module, and a multi-modal perception module. The intelligent control system includes an adaptive grasping control unit and a fault-tolerant identification and decision-making unit; The multimodal perception module is configured to collect characteristic data and / or environmental data of the target being grasped; The adaptive grasping control unit is communicatively connected to the multimodal sensing module, the adaptive grasping end module, and the multi-degree-of-freedom motion module, and is configured to dynamically adjust the grasping parameters of the adaptive grasping end module and / or the motion parameters of the multi-degree-of-freedom motion module based on the data collected by the multimodal sensing module. The fault-tolerant identification and decision-making unit is configured to identify faults in the grasping process based on feedback data from the multimodal perception module and / or the adaptive grasping control unit, and to perform corresponding graded fault-tolerant handling according to the fault level.

[0018] According to an embodiment of the adaptive gripping fault-tolerant control system for chemical laboratory automation, in a specific implementation, the adaptive gripping end module includes: The three-jaw linkage clamping structure is provided with three sets of flexible clamping arms, and each set of flexible clamping arms is equipped with a micro force sensor, a displacement sensor and a torque sensor. A quick-release flexible gripper is switchably mounted on the end of the flexible gripping arm and includes a silicone gripper, a fluororubber gripper, and a honeycomb anti-slip rubber gripper. The surface of the quick-release flexible gripper has a plasma-modified coating and has an adaptive deformation range of 0-5mm. A dual-mode clamping force adjustment mechanism is configured to drive the flexible gripper to open and close, supporting switching between constant force clamping mode and constant stroke clamping mode. The multi-degree-of-freedom motion module includes a six-degree-of-freedom robotic arm, which has linear motion degrees of freedom along the X, Y, and Z axes, as well as rotational degrees of freedom around the X, Y, and Z axes. The six-degree-of-freedom robotic arm uses a servo motor and a precision ball screw drive, and is equipped with an absolute encoder and a torque sensor for closed-loop feedback. The multi-degree-of-freedom motion module also has a built-in dynamic anti-shake and inertia compensation algorithm, configured to dynamically adjust the driving torque according to the load inertia.

[0019] According to an embodiment of the adaptive gripping fault-tolerant control system for chemical laboratory automation, in a specific implementation, the adaptive gripping end module includes: The three-jaw linkage clamping structure is provided with three sets of flexible clamping arms, and each set of flexible clamping arms is equipped with a micro force sensor, a displacement sensor and a torque sensor. A quick-release flexible gripper is switchably mounted on the end of the flexible gripping arm and includes a silicone gripper, a fluororubber gripper, and a honeycomb anti-slip rubber gripper. The surface of the quick-release flexible gripper has a plasma-modified coating and has an adaptive deformation range of 0-5mm. A dual-mode clamping force adjustment mechanism is configured to drive the flexible gripper to open and close, supporting switching between constant force clamping mode and constant stroke clamping mode. The multi-degree-of-freedom motion module includes a six-degree-of-freedom robotic arm, which has linear motion degrees of freedom along the X, Y, and Z axes, as well as rotational degrees of freedom around the X, Y, and Z axes. The six-degree-of-freedom robotic arm uses a servo motor and a precision ball screw drive, and is equipped with an absolute encoder and a torque sensor for closed-loop feedback. The multi-degree-of-freedom motion module also has a built-in dynamic anti-shake and inertia compensation algorithm, configured to dynamically adjust the driving torque according to the load inertia.

[0020] The adaptive grasping fault-tolerant control system for chemical laboratory automation according to an embodiment of the present invention further includes, in a specific embodiment: Environmental protection module, the environmental protection module includes: Labyrinth-type sealing structure and fluororubber sealing ring are installed at the joints of the robotic arm; The nitrogen purging device linked to the environmental sensor is configured to adaptively adjust the nitrogen flow rate according to the concentration of corrosive gases to form positive pressure protection; An electrical control box, which is equipped with a dehumidification module, an anti-corrosion filter module and a temperature regulation unit, and the surface of the box is coated with polytetrafluoroethylene. The security protection module includes: The emergency braking unit is configured with dual triggering of hardware mechanical braking and software electronic braking; The reagent leakage detection unit is configured with dual detection of electrochemical sensors and visual recognition. The overload protection unit is configured to monitor the load through a force sensor, torque sensor, and motor current sensor working together. An anti-tipping unit, comprising an IMU inertial sensor configured to monitor the container's attitude; An explosion-proof unit, the explosion-proof unit including explosion-proof electrical components.

[0021] According to an embodiment of the present invention, the adaptive grasping fault-tolerant control system for chemical laboratory automation adopts a three-level architecture of host computer-edge node-actuator in a certain specific embodiment. The edge node is integrated into the multi-degree-of-freedom motion module and the multi-modal sensing module, and communicates with the fieldbus via dual links through industrial Ethernet.

[0022] According to an embodiment of the present invention, an adaptive grasping fault-tolerant control system for chemical laboratory automation is provided. In a specific implementation, the adaptive grasping control unit has a built-in three-dimensional mapping model of target characteristics, clamping parameters, and environmental parameters, and integrates fuzzy PID algorithm and reinforcement learning algorithm. The adaptive grasping control unit is configured to acquire the size, material, weight, posture, surface state, temperature, and ambient gas concentration of the grasping target through the multimodal perception module, call a preset parameter database to match the initial grasping parameters, and dynamically adjust the clamping force, clamping angle, movement speed, and acceleration based on the force-displacement-torque-visual four-dimensional feedback data during the grasping process using the fuzzy PID algorithm. The reinforcement learning algorithm uses grasping success rate, clamping stability, and operation efficiency as reward and punishment objectives, iteratively optimizes the clamping parameters of similar targets, and generates personalized parameter templates.

[0023] According to an embodiment of the present invention, the adaptive grasping fault-tolerant control system for chemical laboratory automation is configured, in a specific embodiment, to execute a fully closed-loop fault-tolerant process including fault prediction, real-time identification, graded handling, and post-event optimization. The fault prediction uses a trend analysis algorithm to provide early warning of sensor data drift and clamping force attenuation. The fault types identified in real time include at least grasping faults, motion faults, sensor faults, and environmental faults. The tiered treatment includes: Level 1 fault tolerance enables self-healing of minor faults through parameter correction, self-cleaning, or trajectory fine-tuning. Level 2 fault tolerance: For moderate faults, continuous operation is maintained by switching to backup parameter schemes, activating redundant sensors or auxiliary power units. Level 3 fault tolerance triggers emergency braking, activates protective devices, and records fault data in response to major faults.

[0024] According to an embodiment of the adaptive grasping fault-tolerant control system for chemical laboratory automation, in a specific implementation, the intelligent control system further includes a multi-source data fusion unit. This multi-source data fusion unit is configured to use Kalman filtering to remove sensor measurement noise, employ an attention mechanism to weightedly fuse data from different operational stages, and use an evidence theory model to cross-validate the multi-source data. The data processing latency of the multi-source data fusion unit is ≤5ms, and it supports automatic switching to redundant sensor channels when data anomalies occur.

[0025] According to an embodiment of the present invention, the adaptive grasping fault-tolerant control system for chemical laboratory automation further includes, in a specific embodiment, a safety monitoring and feedback unit. This safety monitoring and feedback unit is configured to monitor and record grasping operation parameters, fault handling details, equipment status, and operation videos in real time, generating an operation log that meets GLP / GMP requirements. The safety monitoring and feedback unit is also equipped with tiered alarm functions, remote control, tiered management of operating permissions, and equipment health diagnostics.

[0026] According to an embodiment of the present invention, the adaptive grasping fault-tolerant control system for chemical laboratory automation has, in a specific implementation, more than 300 laboratory container parameters built into the preset parameter database, and supports parameter migration across scenarios.

[0027] To facilitate understanding of the above technical solutions of the present invention, the following detailed description of the above technical solutions of the present invention is provided through specific embodiments and examples.

[0028] In practical applications, the adaptive grasping fault-tolerant control system for chemical laboratory automation according to the present invention includes a hardware execution system and an intelligent control system, which work together to achieve adaptive grasping and full-dimensional fault-tolerant control.

[0029] The hardware execution system adopts a modular design, including an adaptive grasping end module, a multi-degree-of-freedom motion module, a multi-modal perception module, an environmental protection module, and a security module. Each module is linked together through a standardized interface.

[0030] The adaptive gripping end module adopts a switchable three-jaw linkage clamping structure, equipped with three sets of flexible gripping arms. Each set of flexible gripping arms has a built-in micro force sensor, displacement sensor and torque sensor to construct a three-dimensional force-displacement-torque feedback system. The end of the gripping arm adopts quick-release flexible grippers, including silicone grippers, fluororubber grippers and honeycomb anti-slip rubber grippers. The gripper surface is plasma modified and has an adaptive deformation range of 0-5mm. It has a built-in dual-mode gripping force adjustment mechanism, which drives the gripper to open and close through a stepper motor. The gripping range is 5-100mm, supports switching between constant force gripping mode and constant stroke gripping mode, and has a buffer gripping subroutine.

[0031] The multi-degree-of-freedom motion module includes a six-degree-of-freedom robotic arm with linear motion degrees of freedom along the X, Y, and Z axes, as well as rotational degrees of freedom around the X, Y, and Z axes. The robotic arm uses a servo motor and precision ball screw drive, and is equipped with an absolute encoder and torque sensor closed-loop feedback, achieving a positioning accuracy of ±0.008mm and a repeatability of ±0.005mm. The multi-degree-of-freedom motion module incorporates dynamic anti-shake and inertia compensation algorithms, configured to dynamically adjust the driving torque based on load inertia, and supports real-time trajectory correction based on Bézier curves, with a response latency of ≤0.1s.

[0032] The multimodal sensing module includes a visual sensor, a laser rangefinder, an environmental sensor, an infrared temperature sensor, a self-cleaning unit, and force and torque sensors disposed within the flexible gripping arm. The visual sensor is configured to identify the position, size, material, posture, and surface condition of the target to be grasped. The laser rangefinder is configured to detect the distance to obstacles and the target grasping distance, supporting dynamic switching of sampling frequency from 10 to 100 Hz. The environmental sensor integrates an electrochemical gas sensor and a temperature and humidity sensor to collect data on corrosive gas concentration and temperature and humidity. The infrared temperature sensor detects the surface temperature of the target to be grasped, with a measurement range of -50℃ to 300℃ and an accuracy of ±0.2℃. The self-cleaning unit uses high-pressure dry nitrogen to periodically purge the sensor probes. All sensor probes are coated with a PTFE anti-corrosion coating and a nano-hydrophobic coating.

[0033] The environmental protection module includes a labyrinth-type sealing structure and fluororubber sealing rings installed at the joints of the robotic arm, a nitrogen purging device linked to environmental sensors, and an electrical control box. The nitrogen purging device is configured to adaptively adjust the nitrogen flow rate from 0.5 to 2 L / min according to the concentration of corrosive gases to form positive pressure protection. The electrical control box has a built-in dehumidification module, an anti-corrosion filter module, and a temperature regulation unit. The surface of the box has a polytetrafluoroethylene coating and an IP65 protection rating.

[0034] The safety protection module includes an emergency braking unit, a reagent leakage detection unit, an overload protection unit, an anti-tipping unit, and an explosion-proof unit. The emergency braking unit employs dual triggering of hardware mechanical braking and software electronic braking. The reagent leakage detection unit uses dual detection of electrochemical sensors and visual recognition. The overload protection unit monitors the overload through the coordinated use of force sensors, torque sensors, and motor current sensors. The anti-tipping unit includes an IMU inertial sensor configured to monitor the container's attitude. The explosion-proof unit includes explosion-proof electrical components.

[0035] The intelligent control system adopts a three-level architecture of host computer-edge node-actuator. The edge node is integrated into the multi-degree-of-freedom motion module and the multi-modal perception module, and communicates through a dual link of gigabit industrial Ethernet and Profinet bus, with a data transmission latency of ≤8ms. The intelligent control system includes an adaptive grasping control unit, a fault-tolerant identification and decision-making unit, a multi-source data fusion unit, and a safety monitoring and feedback unit.

[0036] The adaptive grasping control unit has a built-in three-dimensional mapping model of target characteristics, clamping parameters, and environmental parameters, and integrates fuzzy PID algorithm and reinforcement learning algorithm. The adaptive grasping control unit is configured to acquire the size, material, weight, posture, surface state, temperature, and ambient gas concentration of the grasping target through the multimodal perception module, call the preset parameter database to match the initial grasping parameters, and dynamically adjust the clamping force, clamping angle, movement speed, and acceleration based on the force-displacement-torque-visual four-dimensional feedback data during the grasping process using the fuzzy PID algorithm. The reinforcement learning algorithm uses grasping success rate, clamping stability, and operation efficiency as reward and punishment objectives, iteratively optimizes the clamping parameters of similar targets, and generates personalized parameter templates.

[0037] The fault-tolerant identification and decision-making unit is configured to execute a fully closed-loop fault-tolerant process including fault prediction, real-time identification, graded handling, and post-event optimization; fault prediction uses trend analysis algorithms to provide early warnings of sensor data drift and clamping force attenuation; real-time identification includes fault types such as grasping faults, motion faults, sensor faults, and environmental faults; graded handling includes level one fault tolerance, level two fault tolerance, and level three fault tolerance.

[0038] The multi-source data fusion unit is configured to use Kalman filtering to remove sensor measurement noise, use an attention mechanism to perform weighted fusion of data from different operation stages, and use an evidence theory model to cross-validate the multi-source data. The data processing latency of the multi-source data fusion unit is ≤5ms, and it supports automatic switching to redundant sensor channels when data is abnormal.

[0039] The safety monitoring and feedback unit is configured to monitor and record capture operation parameters, fault handling details, equipment status and operation videos in real time, and generate operation logs that meet GLP / GMP requirements; the safety monitoring and feedback unit is also configured with hierarchical alarm, remote control, hierarchical management of operation permissions and equipment health diagnosis functions.

[0040] The preset parameter database contains more than 300 laboratory container parameters and supports parameter migration across scenarios.

[0041] The present invention will be further described in detail below with reference to specific embodiments, but should not be construed as limiting the present invention in any way.

[0042] I. System Setup Based on the above technical solution, an adaptive gripping fault-tolerant control system for chemical laboratory automation was built. The hardware component uses a corrosion-resistant aerospace-grade aluminum alloy frame, equipped with a six-degree-of-freedom robotic arm. The adaptive gripping end effector is equipped with three sets of flexible gripping arms and switchable quick-release jaws, including silicone, fluororubber, and honeycomb anti-slip rubber jaws. The multimodal sensing module uses a Basler acA2500-15uc corrosion-resistant industrial camera with a telecentric lens, a SickTIM300 laser rangefinder, a miniature high-precision force sensor, a displacement sensor, a torque sensor, an electrochemical gas sensor, a temperature and humidity sensor, and an infrared temperature sensor. The environmental protection module is equipped with labyrinth-type sealing joints, fluororubber sealing rings, a nitrogen purging device, and an IP65-rated electrical control box. The box is coated with PTFE and includes a dehumidification module, a corrosion-resistant filter module, and a temperature control unit. The safety module integrates an emergency braking device, a reagent leakage detector, an overload protection unit, an IMU inertial sensor, and explosion-proof electrical components.

[0043] The control system uses an industrial computer as the host computer and an FPGA+MCU heterogeneous architecture as edge nodes, communicating via a dual-link connection between gigabit industrial Ethernet and Profinet bus. The adaptive capture control unit runs fuzzy PID and reinforcement learning algorithms, the fault-tolerant identification and decision-making unit has four types of fault identification models built in, the multi-source data fusion unit adopts a fusion model of Kalman filtering, attention mechanism and evidence theory, and the safety monitoring and feedback unit is equipped with a 15-inch high-definition touch screen and an 8GB data cache module. The preset parameter database contains more than 300 common laboratory container parameters, including but not limited to 1.5mL centrifuge tubes, 5mL cryovials, 10mL glass test tubes, 25mL volumetric flasks, 50mL reagent bottles, 100mL beakers, 250mL conical flasks and various irregularly shaped reaction containers, and supports cross-scenario parameter migration.

[0044] II. Experimental Procedure Taking laboratory acid-base reagent transfer as an example, the specific steps for using this system to pick up and transfer a 50mL glass sodium hydroxide reagent bottle are as follows: Preprocessing and parameter matching: The user imports the task through the host computer, setting the target to be grasped as a "50mL glass sodium hydroxide reagent bottle" and the transport path as "reagent cabinet - reaction table - waste liquid recovery area". The system initiates equipment self-check, completes laser sensor calibration, force sensor calibration, and robotic arm zeroing; automatically matches fluororubber grippers according to the target material and temperature; the nitrogen purging device is activated to establish positive pressure protection; the multimodal perception module collects initial environmental data of the grasping area, constructs a 1mm precision raster dynamic environmental map and marks dangerous areas; the vision sensor and infrared temperature sensor work together to identify the target's position, size, posture and surface temperature, and call the preset parameter database to match the initial grasping parameters: clamping force 8N, clamping angle 30°, movement speed 200mm / s, acceleration 1.2m / s², buffer time 0.3s, and corrects the movement speed to 180mm / s based on the environmental data.

[0045] Multimodal real-time perception and data fusion: The robotic arm moves to the reagent cabinet area according to the planned path. The perception module collects data collaboratively at a frequency of 20Hz, while the laser rangefinder and IMU collect data at a high frequency of 100Hz. The vision sensor monitors the position and attitude of the reagent bottle in real time, the force sensor and torque sensor are ready to collect clamping feedback data, the infrared temperature sensor continuously monitors the container temperature, and the environmental sensor updates the concentration of corrosive gas synchronously. The multi-source data fusion unit focuses on key data according to the operation stage, removes noise through Kalman filtering, assigns weights through attention mechanism, and cross-validates through evidence theory, outputs accurate target state and environmental state data, and synchronizes them to the adaptive grasping control unit and the fault-tolerant recognition and decision-making unit.

[0046] Adaptive gripping and posture optimization execution: Based on fused data, the adaptive gripping control unit optimizes the gripping parameters through a fuzzy PID algorithm, adjusting the gripping force to 7N and correcting the gripping angle to 35°. The gripper is driven to smoothly grip the reagent bottle in a three-step process of "buffered contact - gradual force gripping - stable gripping". During the gripping process, the force sensor reports that the gripping force is stable at 7N±0.1N, and the torque sensor detects no slippage trend. After gripping, the robotic arm moves along the optimized trajectory at a speed of 180mm / s. During the journey, the acceleration is adaptively adjusted to 1.0m / s² based on the weight estimated by force-torque data fusion. The laser rangefinder dynamically detects obstacles and updates the path in real time.

[0047] Fully closed-loop fault-tolerant monitoring and graded handling: When the device was transferred to the top of the reaction table, the laser rangefinder detected a temporary obstacle, triggering a level one fault. The fault-tolerant identification and decision-making unit corrected the trajectory within 0.1 seconds, bypassing the obstacle, and the clamping stability was not affected. During the operation, the environmental sensor detected a slight fluctuation in gas concentration and activated the sensor self-cleaning unit to purge with nitrogen. After 0.3 seconds, the data returned to normal, and the operation continued without triggering any interruption.

[0048] Data recording, self-learning, and parameter optimization: After the transfer is completed, the safety monitoring and feedback unit records the grabbing time, clamping parameters, fault handling, equipment status, and operation video, and generates a standardized operation report; the system analyzes the grabbing data through reinforcement learning algorithms, using grabbing success rate, clamping stability, and operation efficiency as reward and penalty targets, iteratively optimizes the optimal clamping force of the sodium hydroxide reagent bottle to 6.8N, updates the parameter database, and generates a personalized parameter template; at the same time, it assesses the health status of the equipment and generates maintenance suggestions.

[0049] This embodiment completed 100 continuous gripping and transfer operations with a gripping success rate of 99.5%. Only once was there a slight slippage caused by oil on the container surface. The system recovered after strengthening the gripping force by 15% through a two-level fault tolerance strategy. There were no reagent leaks or container damages. The average single operation time was shortened by 30% compared to traditional equipment, and the fault response time was ≤0.2s.

[0050] III. Verification of Abnormal Scenario Response To verify the system's fault tolerance capability, two typical abnormal scenarios were simulated: Container tilting and offset scenario: The reagent bottle is tilted at 10° and its position is offset by 2mm. The vision sensor quickly identifies the posture deviation, and the adaptive gripping control unit corrects the rotation angle and clamping angle of the robotic arm within 0.15s to ensure that the deviation between the clamping center and the container's center of gravity is ≤0.5mm. The force sensor dynamically adjusts the clamping force to 6.5N, successfully and stably clamping the container. There is no slippage during the transfer process, and the operation proceeds normally.

[0051] Sensor failure scenario: The vision sensor experienced signal fluctuations due to interference from corrosive gases. The multi-source data fusion unit cross-verified the data from the force sensor and the laser rangefinder, determining that the sensor probe was contaminated. The self-cleaning unit was activated to purge with nitrogen. The vision signal returned to normal after 0.3 seconds. During this period, the robotic arm maintained its original gripping state, and the grasping and transfer were stable and uninterrupted.

[0052] In summary, by utilizing the technical solutions described above, the multimodal perception module and the adaptive grasping control unit work together to achieve precise parameter adaptation for grasping targets of different materials, sizes, and postures. This enables diverse grasping operations to be completed without manual intervention, significantly improving the system's environmental adaptability and versatility. Furthermore, by constructing a fully closed-loop hierarchical fault-tolerant system through fault-tolerant identification and decision-making units, self-healing, adjustment, and emergency response strategies are adopted for minor, moderate, and major faults, respectively. This effectively avoids operational interruptions caused by single faults, thereby greatly improving the continuity and reliability of grasping operations. Finally, the integrated design of the environmental protection module and the safety assurance module forms a multi-dimensional active protection mechanism, effectively reducing safety risks such as corrosion, reagent leakage, and container damage, thus comprehensively improving the safety level of automated laboratory operations.

[0053] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An adaptive grasping fault-tolerant control system for chemical laboratory automation, characterized in that, include: The hardware execution system includes an adaptive grasping end effector module, a multi-degree-of-freedom motion module, and a multi-modal perception module. The intelligent control system includes an adaptive grasping control unit and a fault-tolerant identification and decision-making unit; The multimodal perception module is configured to collect characteristic data and / or environmental data of the target being grasped; The adaptive grasping control unit is communicatively connected to the multimodal sensing module, the adaptive grasping end module, and the multi-degree-of-freedom motion module, and is configured to dynamically adjust the grasping parameters of the adaptive grasping end module and / or the motion parameters of the multi-degree-of-freedom motion module based on the data collected by the multimodal sensing module. The fault-tolerant identification and decision-making unit is configured to identify faults in the grasping process based on feedback data from the multimodal perception module and / or the adaptive grasping control unit, and to perform corresponding graded fault-tolerant handling according to the fault level.

2. The adaptive grasping fault-tolerant control system for chemical laboratory automation according to claim 1, characterized in that, The adaptive crawling end module includes: The three-jaw linkage clamping structure is provided with three sets of flexible clamping arms, and each set of flexible clamping arms is equipped with a micro force sensor, a displacement sensor and a torque sensor. A quick-release flexible gripper is switchably mounted on the end of the flexible gripping arm and includes a silicone gripper, a fluororubber gripper, and a honeycomb anti-slip rubber gripper. The surface of the quick-release flexible gripper has a plasma-modified coating and has an adaptive deformation range of 0-5mm. A dual-mode clamping force adjustment mechanism is configured to drive the flexible gripper to open and close, supporting switching between constant force clamping mode and constant stroke clamping mode. The multi-degree-of-freedom motion module includes a six-degree-of-freedom robotic arm, which has linear motion degrees of freedom along the X, Y, and Z axes, as well as rotational degrees of freedom around the X, Y, and Z axes. The six-degree-of-freedom robotic arm uses a servo motor and a precision ball screw drive, and is equipped with an absolute encoder and a torque sensor for closed-loop feedback. The multi-degree-of-freedom motion module also has a built-in dynamic anti-shake and inertia compensation algorithm, configured to dynamically adjust the driving torque according to the load inertia.

3. The adaptive grasping fault-tolerant control system for chemical laboratory automation according to claim 2, characterized in that, The multimodal sensing module includes: A vision sensor configured to identify the position, size, material, orientation, and surface condition of the target being grasped; A laser rangefinder sensor is configured to detect the distance to obstacles and the target capture distance. An environmental sensor is configured to collect data on corrosive gas concentration, temperature, and humidity. An infrared temperature sensor is configured to detect the surface temperature of the target being grasped. The self-cleaning unit is configured to purge the sensor probe with high-pressure dry nitrogen gas; The system also includes a force sensor and a torque sensor disposed within the flexible clamping arm, which together form part of the multimodal sensing module.

4. The adaptive grasping fault-tolerant control system for chemical laboratory automation according to claim 3, characterized in that, Also includes: Environmental protection module, the environmental protection module includes: Labyrinth-type sealing structure and fluororubber sealing ring are installed at the joints of the robotic arm; The nitrogen purging device linked to the environmental sensor is configured to adaptively adjust the nitrogen flow rate according to the concentration of corrosive gases to form positive pressure protection; An electrical control box, which is equipped with a dehumidification module, an anti-corrosion filter module and a temperature regulation unit, and the surface of the box is coated with polytetrafluoroethylene. The security protection module includes: The emergency braking unit is configured with dual triggering of hardware mechanical braking and software electronic braking; The reagent leakage detection unit is configured with dual detection of electrochemical sensors and visual recognition. The overload protection unit is configured to monitor the load through a force sensor, torque sensor, and motor current sensor working together. An anti-tipping unit, comprising an IMU inertial sensor configured to monitor the container's attitude; An explosion-proof unit, the explosion-proof unit including explosion-proof electrical components.

5. The adaptive grasping fault-tolerant control system for chemical laboratory automation according to claim 1, characterized in that, The intelligent control system adopts a three-level architecture of host computer-edge node-actuator. The edge node is integrated into the multi-degree-of-freedom motion module and the multi-modal sensing module, and communicates with the fieldbus via dual links through industrial Ethernet.

6. The adaptive grasping fault-tolerant control system for chemical laboratory automation according to claim 1, characterized in that, The adaptive grasping control unit has a built-in three-dimensional mapping model of target characteristics, clamping parameters, and environmental parameters, and integrates fuzzy PID algorithm and reinforcement learning algorithm. The adaptive grasping control unit is configured to acquire the size, material, weight, posture, surface state, temperature, and ambient gas concentration of the grasping target through the multimodal perception module, call the preset parameter database to match the initial grasping parameters, and dynamically adjust the clamping force, clamping angle, movement speed, and acceleration based on the force-displacement-torque-visual four-dimensional feedback data during the grasping process using the fuzzy PID algorithm. The reinforcement learning algorithm uses grasping success rate, clamping stability, and operation efficiency as reward and punishment objectives, iteratively optimizes the clamping parameters of similar targets, and generates personalized parameter templates.

7. The adaptive grasping fault-tolerant control system for chemical laboratory automation according to claim 1, characterized in that, The fault-tolerant identification and decision-making unit is configured to execute a fully closed-loop fault-tolerant process, including fault prediction, real-time identification, graded handling, and post-event optimization. The fault prediction uses a trend analysis algorithm to provide early warning of sensor data drift and clamping force attenuation. The fault types identified in real time include at least grasping faults, motion faults, sensor faults, and environmental faults. The tiered treatment includes: Level 1 fault tolerance enables self-healing of minor faults through parameter correction, self-cleaning, or trajectory fine-tuning. Level 2 fault tolerance: For moderate faults, continuous operation is maintained by switching to backup parameter schemes, activating redundant sensors or auxiliary power units. Level 3 fault tolerance triggers emergency braking, activates protective devices, and records fault data in response to major faults.

8. The adaptive grasping fault-tolerant control system for chemical laboratory automation according to claim 1, characterized in that, The intelligent control system also includes a multi-source data fusion unit, which is configured to use Kalman filtering to remove sensor measurement noise, use an attention mechanism to perform weighted fusion of data from different operation stages, and use an evidence theory model to cross-validate the multi-source data; the data processing delay of the multi-source data fusion unit is ≤5ms, and it supports automatic switching to redundant sensor channels when data is abnormal.

9. The adaptive grasping fault-tolerant control system for chemical laboratory automation according to claim 1, characterized in that, The intelligent control system also includes a safety monitoring and feedback unit, which is configured to monitor and record capture operation parameters, fault handling details, equipment status and operation videos in real time, and generate operation logs that meet GLP / GMP requirements. The safety monitoring and feedback unit is also equipped with hierarchical alarm, remote control, hierarchical management of operation permissions and equipment health diagnosis functions.

10. The adaptive grasping fault-tolerant control system for chemical laboratory automation according to any one of claims 1 to 9, characterized in that, The preset parameter database contains more than 300 laboratory container parameters and supports parameter migration across scenarios.