Mechanical device control system for grabbing scrap steel

Abstract

The invention relates to a mechanical device control system for grabbing waste steel, and belongs to the technical field of waste steel recovery. In order to solve the problems that an existing mechanical device is unstable in grabbing, complex in operation, insufficient in open-loop control precision and the like, a hardware system comprising an omni-directional AGV mobile platform, a seven-degree-of-freedom cooperative mechanical arm and a dexterous gripper is provided, and a hybrid power supply framework, FPGA / DSP hierarchical control and high-speed LVDS bus communication are combined. According to the technical scheme, a multi-loop control algorithm is adopted, and dynamic torque compensation and high-precision trajectory tracking are achieved through current loop feed-forward compensation, two-degree-of-freedom PID speed adjustment and position feed-forward filtering. The system can adapt to grabbing of waste steel in complex shapes, the operation stability and safety are improved, the manual intervention requirement is reduced, meanwhile, the modular maintenance advantage is achieved, and the waste steel recovery efficiency and the system reliability are effectively improved.

Description

Technical Field

[0001] The invention belongs to the technical field of scrap steel recovery and relates to a mechanical device control system for scrap steel grabbing. Background Art

[0002] Scrap steel recycling is a crucial step in resource recycling, and its efficiency and quality directly impact the sustainable development of the steel industry. During the scrap steel recycling process, the mechanical device control system is responsible for key operations such as gripping and handling, and its performance directly determines operational accuracy and safety. Traditional scrap steel recycling mechanical devices mostly use an open-loop control system design, relying solely on preset parameters to drive the mechanical device's movement. They are unable to perceive changes in the external environment and load conditions in real time, leading to problems such as unstable gripping force and insufficient positioning accuracy. Especially when faced with scrap steel with complex shapes and rough surfaces, existing systems often experience gripping failures or material shedding due to a lack of dynamic adjustment capabilities, seriously affecting recycling efficiency.

[0003] Existing technologies often require complex mechanical devices to operate, requiring specialized personnel to set parameters and conduct real-time intervention. This not only increases labor costs but also poses safety risks due to operational errors. Furthermore, traditional robotic arms, limited by their insufficient degrees of freedom and rigid structural design, struggle to adapt to the diverse scenarios of scrap steel handling. For example, open-loop control systems are unable to compensate for torque fluctuations during the arm's motion in real time, causing the end effector to vibrate or deflect at high speeds, further reducing gripping reliability.

[0004] Although some improvements have attempted to incorporate sensor feedback, limitations in hardware architecture and algorithm design still pose challenges such as high control latency and insufficient communication bandwidth. For example, the processing power of the underlying microcontroller-based controller is limited, making it difficult to achieve coordinated control of multiple joints. Systems powered by a single power supply are susceptible to electromagnetic interference, which can affect signal transmission stability. Furthermore, existing end effectors often utilize electromagnetic or vacuum adsorption, which are less adaptable to irregularly shaped scrap and have high maintenance costs.

[0005] Therefore, there is an urgent need for a scrap steel grabbing control system with high degree of freedom, high precision and real-time dynamic adjustment capability to solve the core problems existing in the existing technology, such as complex operation, poor grabbing stability and insufficient safety. Summary of the Invention

[0006] In view of this, an object of the present invention is to provide a control system for a mechanical device for scrap steel grabbing.

[0007] In order to achieve the above object, the present invention provides the following technical solutions:

[0008] A control system for a mechanical device for scrap steel grabbing, comprising:

[0009] The mobile platform uses an omnidirectional AGV mobile robot with a laser radar and path planning module installed on its top;

[0010] A seven-degree-of-freedom collaborative robot arm, whose base is fixed on the mobile platform, includes a shoulder joint, a wrist joint and a hollow shaft structure, has a working radius of ≥900mm, and a repeatability accuracy of ±1mm;

[0011] The end effector is installed at the end of the robotic arm and adopts a coupled transmission dexterous gripper to drive multi-finger linkage through a steel wire mechanism;

[0012] A power supply system, electrically connected to the mobile platform, the robotic arm, and the end effector, adopting a hybrid power supply architecture combining a switching power supply and a linear power supply;

[0013] The underlying control system is integrated into the joints of the robotic arm and implements motor drive and real-time processing of sensor signals based on FPGA;

[0014] The upper control system communicates with the lower control system via the PCI bus and uses a DSP / FPGA architecture to perform kinematics calculations and trajectory planning;

[0015] A high-speed serial bus connects the bottom layer and the upper layer control system based on low voltage differential signal (LVDS) to transmit joint status data and control instructions;

[0016] The control algorithm module is deployed in the upper and lower control systems, including the current loop controller, the speed loop controller and the position loop controller, wherein the current loop adopts i d =0 control strategy.

[0017] Furthermore, the joint structure of the seven-degree-of-freedom collaborative robot arm includes:

[0018] The shoulder joint is fixed to the upper surface of the mobile platform through a flange connection, and an outer rotor brushless motor and a harmonic reducer are integrated therein, and the output shaft of the motor is rigidly connected coaxially with the wave generator of the harmonic reducer;

[0019] The wrist joint is connected in series with the output end of the harmonic reducer of the shoulder joint through a hollow shaft sleeve, and has a built-in outer rotor brushless motor and a four-point contact bearing. The inner ring of the bearing is fixed to the motor rotor, and the outer ring is connected to the joint housing;

[0020] The dual encoder system includes a relative encoder on the motor side and an absolute encoder on the joint side, where:

[0021] The magnetic ring of the relative encoder is fixed to the rotor end face of the outer rotor brushless motor and rotates coaxially with the motor;

[0022] The reading head of the absolute encoder is installed on the inner side of the joint housing, and its magnetic ring is interference-fitted with the output shaft of the harmonic reducer through an adapter;

[0023] Wires run through the hollow shafts of the shoulder joint and the wrist joint, and the wires are electrically connected to the FPGA signal interface of the underlying control system.

[0024] Furthermore, the dexterous gripper of the end effector comprises:

[0025] The drive unit is composed of a disc motor and a harmonic reducer, wherein the disc motor is fixed inside the base joint of the finger, and its output shaft is connected to the input shaft of the harmonic reducer through a synchronous belt mechanism;

[0026] The multi-finger linkage mechanism includes four sets of parallel steel wire transmission assemblies. One end of each steel wire assembly is wound around the output wheel of the harmonic reducer, and the other end passes through the guide pulley of the finger joint and is fixed to the end knuckle.

[0027] A plastic packaging piece, which is detachably covered on the outside of the finger joint by a snap-fit ​​structure, and has a friction pattern on its inner surface;

[0028] The transmission ratio of the wire transmission assembly and the synchronous belt mechanism is configured to be 1:2 to 1:4, and the closing stroke of the terminal knuckle is controlled by the output angle of the harmonic reducer.

[0029] Furthermore, the power supply system includes:

[0030] A main DC / DC conversion module, whose input end is electrically connected to an external 48V DC power supply, and whose output end includes a first output end and a second output end, wherein the first output end provides a 12V voltage and the second output end provides a 48V voltage;

[0031] The linear voltage regulator module is connected to the 12V output terminal of the main DC / DC conversion module through a copper busbar, and provides a 5V regulated power supply to the FPGA, DSP and sensor interface after passing through a multi-stage filtering circuit;

[0032] The motor drive dedicated isolated power supply adopts a flyback topology. Its primary side winding is connected to the 48V output terminal of the main DC / DC conversion module, and the secondary side winding supplies power to the brushless motor drivers of each joint through a magnetic isolation transformer;

[0033] A common-mode inductor is provided between the linear voltage stabilization module and the motor drive dedicated isolation power supply to suppress high-frequency conducted interference.

[0034] Furthermore, the underlying control system includes:

[0035] An FPGA chip, whose signal input terminal is electrically connected to the magnetic encoder of each joint, is configured to convert the original encoder signal into a quadruple frequency pulse signal through a code disk quadruple frequency module;

[0036] An M / T speed measurement module is integrated into the FPGA chip, receives the quadruple frequency pulse signal and calculates the real-time speed of the joint motor;

[0037] An SPWM generation module generates a three-phase SPWM waveform for driving a brushless motor according to the output of the M / T method speed measurement module and the current loop control instruction;

[0038] The communication interface of the FPGA chip is connected to the DSP processor of the upper control system through a high-speed serial bus to transmit the real-time speed and joint position data;

[0039] The resolution of the code disk quadruple frequency module is adapted to the number of lines of the magnetic encoder, specifically satisfying:

[0040] Assume that the number of lines of the magnetic encoder is N, the equivalent number of pulses after quadruple frequency is 4N, and the reduction ratio of the harmonic reducer is i. Then the angle corresponding to a single pulse of the joint output shaft is

[0041] The number of lines of the magnetic encoder N = 1200, the harmonic reduction ratio i = 250, and the angle corresponding to a single pulse of the joint output shaft is Combined with the robot arm working radius L ≥ 900mm, the end positioning accuracy is L·sin (0.0003°) ≤ ± 0.1mm.

[0042] Furthermore, the upper control system includes:

[0043] a DSP processor, electrically connected to an external control terminal via a PCI bus, configured to receive a grasping instruction and perform an inverse kinematic solution of the manipulator to generate a desired trajectory of joint angles;

[0044] A dual-port RAM memory, wherein the first port thereof is electrically connected to the PCI bus and the second port thereof communicates with the FPGA of the underlying control system via an LVDS bus interface, thereby realizing bidirectional buffering of control instructions and joint status data;

[0045] A feedforward compensation module, integrated into the DSP processor, receives the desired trajectory of the joint angle and calculates gravity compensation and nonlinear friction force feedforward to generate compensation parameters;

[0046] A trajectory smoothing filter is connected between the DSP processor and the feedforward compensation module, and uses a second-order Butterworth filter to perform phase compensation and high-frequency noise suppression on the desired trajectory;

[0047] The compensation parameters and the expected trajectory of the joint angle are sent to the underlying control system via the LVDS bus to drive the motors of each joint to move according to the planned trajectory.

[0048] Furthermore, the communication protocol of the high-speed serial bus includes:

[0049] The data packet structure includes a checksum field and a node identification field; the checksum field is used to verify data integrity, and the node identification field uniquely identifies each joint controller of the robotic arm; the data packet is generated by the joint controller, encoded by the LVDS differential driver chip, and then sent to the FPGA of the upper control system;

[0050] The interrupt trigger mechanism is configured to trigger data transmission immediately via hardware interrupt when joint status data is updated or control instructions arrive;

[0051] Impedance matching circuit, set between the transmitting end and the receiving end of the LVDS bus, suppresses signal reflection to ensure transmission stability;

[0052] The data packets are transmitted via the LVDS bus in a time-division multiplexing manner, and the communication cycle between the upper and lower control systems is adapted to the real-time control requirements of the robotic arm.

[0053] Furthermore, the current loop controller realizes:

[0054] The motor current sampling module has an input end electrically connected to the three-phase current sensor of each joint brushless motor and an output end connected to the Clarke transformation module;

[0055] A vector coordinate transformation module is configured to perform Clarke transformation and Park transformation to convert the three-phase current into i d 、i q Component, where i d The reference value is set to 0; Clarke transformation converts the three-phase current i a ,i b ,i c Converted to the two-phase stationary coordinate system current component i α ,i β , Park transformation is further mapped to the rotor coordinate system i d ,i q The feedforward compensation module calculates the compensation value in real time through the FPGA's built-in DSPSlice. The SPWM generation module adopts a symmetrical regular sampling method with a carrier frequency of 20kHz.

[0056] The feedforward compensation module receives the motor rotor position signal and speed signal and calculates the back electromotive force compensation voltage u dc =p n ωi q and u qc=-p n ωi d -p n ωψ f , superimposed on the output end of the current loop;

[0057] PI adjustment module, for i q The component errors are subjected to proportional-integral operation to generate the q-axis voltage command;

[0058] An SPWM generation module generates a three-phase drive waveform according to the q-axis voltage command and the compensation voltage, and outputs the waveform to the brushless motor via an isolation drive circuit;

[0059] The vector coordinate transformation module, feedforward compensation module and PI adjustment module are all integrated into the FPGA chip of the underlying control system in the form of hardware logic.

[0060] Furthermore, the speed loop controller adopts a two-degree-of-freedom PID structure, including:

[0061] The forward channel regulator receives the speed reference signal from the upper control system at its input, and its transfer function is C1(s)=(1-α)K p +K I ·1 / s+(1-b)K D s, the output end is connected to the command input end of the current loop controller;

[0062] Feedback channel regulator, whose input end is electrically connected to the real-time speed signal of the joint encoder, and the transfer function is C2(s)=aK p +b·K D s, the output end of which is superimposed on the output end of the forward channel regulator;

[0063] The parameter setting module is integrated into the DSP processor of the upper control system, dynamically adjusts the α and b parameters according to the step response curve, and sends the adjusted parameters to the FPGA of the lower control system via the LVDS bus;

[0064] The operation logic of the forward channel regulator and the feedback channel regulator is solidified in the FPGA chip with the hardware description language, and the regulator output is connected to the current loop controller i q The command end is electrically connected.

[0065] Furthermore, the position loop controller includes:

[0066] The proportional regulator receives the position command from the upper control system and the position feedback signal from the joint absolute encoder at its input, and generates the basic position correction at its output;

[0067] A speed feedforward module is electrically connected to the DSP processor of the upper control system, receives the trajectory planning speed instruction signal generated by the DSP processor, and generates a pre-compensation amount through a feedforward coefficient K;

[0068] The trajectory smoothing filter has an input end connected to the output end of the velocity feedforward module, and uses a second-order low-pass filter to perform phase lag compensation on the pre-compensation amount; the trajectory smoothing filter uses a Butterworth second-order low-pass filter with a cutoff frequency f c =50Hz, and compensate 3ms group delay through inverse filter;

[0069] A position correction synthesis module, which superimposes the basic position correction amount and the filtered pre-compensation amount to generate a final position control instruction;

[0070] Position feedback interface, which collects the position data of each joint's absolute encoder in real time through the LVDS bus and uploads it to the DSP processor of the upper control system;

[0071] The final position control instruction is converted into a motor drive signal by the FPGA of the underlying control system to drive the joint motor to move according to the planned trajectory.

[0072] The beneficial effects of the present invention are:

[0073] (1) Through the collaborative design of a seven-degree-of-freedom robotic arm and a dexterous gripper, the system can autonomously adjust its grasping posture to adapt to the needs of grasping scrap steel of different shapes and sizes, overcoming the limitations of traditional rigid structures in complex working conditions.

[0074] (2) Based on the multi-loop control algorithm and real-time feedback mechanism, the system dynamically compensates for the torque fluctuations and load disturbances in the movement of the robotic arm, ensuring the precise positioning and stable grasping of the end effector, and significantly reducing the risk of material falling off.

[0075] (3) A hybrid power supply architecture and modular hardware design are adopted to balance power supply stability and system lightweight, reduce the impact of electromagnetic interference on signal transmission, and reduce maintenance complexity.

[0076] (4) Through the kinematic solution and path planning of the upper control system, combined with the real-time control of the underlying high response speed, automated operation is achieved, the need for manual intervention is reduced, and operational safety hazards are reduced.

[0077] (5) Based on the two-degree-of-freedom PID regulator and back-electromotive force feedforward compensation, the system can still maintain stable operation in high-speed motion scenarios and effectively suppress the impact of external disturbances on control performance.

[0078] (6) The open and dexterous gripper design and modular joint structure facilitate the rapid replacement of actuators or adjustment of grasping strategies, adapting to the needs of diverse industrial scenarios while simplifying the equipment maintenance process.

[0079] Other advantages, objects, and features of the present invention will be described in part in the following description and, in part, will be apparent to those skilled in the art upon examination of the following description or may be learned from practice of the present invention. The objects and other advantages of the present invention may be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below with reference to the accompanying drawings, in which:

[0081] Figure 1 It is a schematic diagram of the overall structure;

[0082] Figure 2 Schematic diagram of a seven-degree-of-freedom robotic arm;

[0083] Figure 3 This is a schematic diagram of hand grasping;

[0084] Figure 4 Schematic diagram of the overall structure of the shoulder joint;

[0085] Figure 5 This is a schematic diagram of the motor extension shaft;

[0086] Figure 6 This is a schematic diagram of the sealing of the harmonic reducer;

[0087] Figure 7 This is a schematic diagram of the power supply system principle;

[0088] Figure 8 It is the underlying controller system based on Xilinx FPGA;

[0089] Figure 9 It is a DSP / FPGA upper control system based on PCI bus;

[0090] Figure 10 It is PP-LVDS serial bus;

[0091] Figure 11 This is the control block diagram of the joint motor servo system;

[0092] Figure 12 This is the speed loop control block diagram;

[0093] Figure 13 This is the block diagram of the current control loop. DETAILED DESCRIPTION

[0094] The following describes the embodiments of the present invention by means of specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and the following embodiments and features in the embodiments can be combined with each other without conflict.

[0095] Among them, the accompanying drawings are only for illustrative purposes and represent only schematic diagrams rather than actual pictures, and should not be understood as limiting the present invention. In order to better illustrate the embodiments of the present invention, some parts of the accompanying drawings may be omitted, enlarged or reduced, and do not represent the dimensions of actual products. For those skilled in the art, it is understandable that some well-known structures and their descriptions may be omitted in the accompanying drawings.

[0096] The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar parts; in the description of the present invention, it should be understood that if there are terms such as "upper", "lower", "left", "right", "front", "back", etc. indicating directions or positional relationships, they are based on the directions or positional relationships shown in the drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operate in a specific direction. Therefore, the terms describing the positional relationship in the drawings are only used for illustrative purposes and cannot be understood as limiting the present invention. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0097] like Figure 1 The entire hardware system of the control system of the mechanical device for scrap steel grabbing of the present invention is composed of a mobile platform, a seven-degree-of-freedom collaborative robot arm, and an end effector.

[0098] The mobile platform described in this article is an AGV (Automated Guided Vehicle) mobile robot, a standard robotic platform for indoor industrial environments. It has a rated load of 300 kg and can achieve omnidirectional movement within the factory. Equipped with two P&F R2000 LiDAR sensors, it provides mapping, path navigation, trajectory planning, and collision avoidance capabilities. Its omnidirectional nature ensures the AGV's flexible and reliable movement, enabling operations in confined areas.

[0099] In order to achieve the grasping of 5Kg of scrap steel, the end effector can adopt the following solutions: electromagnetic adsorption, vacuum adsorption, dexterous grippers, etc.

[0100] Electromagnetic suction is simple and convenient, but it also attracts nearby scrap, making it difficult to operate independently. Vacuum suction is convenient and feasible, but due to the wide variety of shapes (such as flat, cylindrical, and irregular shapes) and the uneven surface, suction and grasping are more difficult. Dexterous grippers can adapt to the grasping of scrap of various shapes, but their structure is complex. Taking all factors into consideration, the weight of the scrap in this project is less than 5 kg, so this project uses a dexterous gripper.

[0101] The finger unit is also driven by a disc motor, located inside the first joint, and a harmonic reducer located at the third joint, J3. The synchronous belt mechanism between the two is similar to that of the base joint. The finger's end joint, J4, does not have independent degrees of freedom, but instead uses a coupled transmission mechanism via a wire mechanism to transmit power to the end joint. Figure 3 This is a structural diagram of the finger joints. The axial width of the fingers of the dexterous hand is the most important dimension to ensure that the size of the dexterous hand is close to that of a human hand.

[0102] Brushless motor: Due to the structural form of the deeply integrated joint, the brushless motor used in this invention can only be of the outer rotor type. Maxon's EC-flat fremeless series of motors has a suitable power range for this robot and stable performance, so it was selected from Maxon's EC-flat fremeless series. According to the motor power obtained by indicator analysis, the EC60flat motor is selected for the shoulder joint and the EC45flat motor is used for the wrist joint. For the sake of parts interchangeability and maintainability, joints 3 and 4 are combined into a group with relatively similar power requirements, that is, joints 1, 2, 3, and 4 use EC60flat motors, while joints 5, 6, and 7 use EC45flat motors. Accordingly, two types of robot integrated joints will be developed.

[0103] The EC60flat motor has a rated torque of 257 mNm and a rated current of 2.3 A. The corresponding harmonic reducer is custom-made based on the specific structural design. For lifespan considerations, the reduction ratio of a cup-type harmonic reducer is generally i≤250. To reduce the burden on the motor, the reduction ratio is calculated as a maximum of 250. Based on the required motor torque:

[0104] Tm1=T1 / 250=166800 / 250=667.2mNm

[0105] This is 2.59 times the rated N = 667.2 / 257. Although it exceeds the rated power, the operating conditions are also extreme conditions, and experience shows that the excess is within the acceptable range.

[0106] The motor has an idling speed of 3970 rpm. At maximum torque, the back EMF is U = 48 - 2.59 x 2.3 = 42.043 V. Therefore, the speed at maximum torque is 42.043 / 48 x 3970 = 3477. After a 250 reduction ratio, the speed is 13.9 rpm, which meets the requirement.

[0107] EC45flat motor, rated torque 134mNm, rated current 0.936A. The reduction ratio is 100, then according to the required motor torque:

[0108] Tm2=T2 / 100=12622 / 100=126mNm

[0109] This is 126 / 134 = 0.94 times the rated value, which is completely acceptable.

[0110] Crossed roller bearings: Calculate the bearing load under extreme operating conditions based on indicator analysis. The extreme operating condition is when the joint is subjected to the maximum overturning moment.

[0111] For the shoulder joint, the maximum overturning moment is when the robotic arm is extended horizontally. The overturning moment is equal to the moment provided by the joint, 167 Nm. In addition, there is a radial load of 280 N due to the weight of the entire arm and the load. The dynamic equivalent load formula is:

[0112]

[0113] in

[0114] PC: Dynamic equivalent radial load (N)

[0115] Fr: Radial load (N)

[0116] Fa: axial load (N)

[0117] M: torque (N·mm)

[0118] X: Dynamic radial coefficient

[0119] Y: Dynamic axial coefficient

[0120] dp: Roller pitch diameter (mm)

[0121] And the empirical formula for classifying different working conditions:

[0122]

[0123] It can be seen that when F a When =0, X=1, Y=0.45.

[0124] Substituting into the equivalent radial load:

[0125] Pc=1×(280+2×166800 / 98)+0=3684N=3.7kN

[0126] Given that the dynamic load and static load safety factors are 4 and 8 respectively, the bearing manufacturer provides a solution with a stick diameter of 4, a bearing cross-sectional area of ​​11mm×11mm, and a pitch circle of 98mm. Dimensions other than the cross-sectional area are designed and customized.

[0127] Motor rotor bearings: The motor rotor is rigidly connected to the wave generator cam, forming a deeply integrated structure. Due to the characteristics of harmonic drive, the cam will be subjected to axial force during the transmission process, which requires rotor bearings to offset it.

[0128] The shafting scheme is configured with one end fixed and the other end free-wheeling. A uniform-section four-point contact bearing or a small cross-roller bearing is used at the fixed end, while a thin-walled deep-groove ball bearing is used at the free-wheeling end.

[0129] The calculation of the shoulder joint axial force refers to the 32 model of the harmonic series. The axial force is calculated as follows:

[0130] F=2×167 / (32×0.00254)×0.07×tan20°=104.7N

[0131] This value is used as a reference for selecting four-point contact bearings.

[0132] The calculation of the wrist joint axial force refers to the 25 model of the harmonic series. The axial force is calculated as follows:

[0133] F=2×126 / (25×0.00254)×0.07×tan20°=101.1N

[0134] Encoder: A relative encoder is used on the motor side for position servo control of the motor, ultimately ensuring the positioning accuracy of the joint angle. An absolute encoder is used on the joint side so that the robot central control can know the current position of the robot every time it is powered on. The magnetic encoder has a line count of 1200PPR, and the equivalent resolution reaches 4800 pulses / rev after quadruple frequency multiplication. Combined with the 250:1 reduction ratio of the harmonic reducer, it achieves an angle resolution of 0.075° at the end, corresponding to a positioning accuracy of ±1mm; the carrier frequency of the SPWM generation module can be configured to 15-30kHz to adapt to different load conditions. The encoder of the shoulder joint uses an RLS product. The RLS series encoders have the same reading head, but different specific models have different magnetic ring diameters. The final selection is determined based on the specific structural design.

[0135] Joint structure design

[0136] Shoulder joint: The shoulder joint has a diameter of 113mm, a length of 73mm, a hollow shaft diameter of 19mm, and a weight of 2.45kg, meeting the lightweight requirements and having a smaller axial dimension than traditional joints.

[0137] The four-point contact bearing used for the motor support is an 8mm uniform cross-section four-point contact bearing from Luoyang Preson:

[0138] The rated axial load is 556kg, which meets the axial force requirements of harmonic drive.

[0139] The cross roller bearing is a custom-made XK-98 from Xinkai Bearing. The inner and outer rings are integrally constructed for superior strength. The rollers are installed from the side and equipped with plungers. The relative encoder uses an RLS MR080R radial magnetic ring. The accuracy can reach ±0.00236°. The overall structure of the shoulder joint is shown in the figure below. Figure 4 shown.

[0140] The wrist joint has a diameter of 102.5mm, a length of 59.2mm, a hollow shaft diameter of 14mm, and a weight of 1.44kg.

[0141] Since the inner diameter of the motor rotor is To provide a shoulder for the deep groove ball bearing, which only bears radial loads, the bearing is selected based on the bore size. The relevant parameters of the 61803 and 61803-2Z bearings are shown in Table 1.

[0142] Table 1

[0143] Bearing name SKFW61803-2Z bearings model W61803-2Z Inner diameterd(mm) 17 Outer diameter D(mm) 26 Thickness B (mm) 5 brand SKF category Deep groove ball bearings

[0144] Substituting into the equivalent radial load:

[0145] Pc=1×(2300+2*11222 / 87)+0.45×87.77=2597.47N≈2.597kN

[0146] If the roller pitch diameter of a customized cross roller bearing is 85, Pc = 0.39 kN.

[0147] We selected the Xinkai RA8008 bearing and customized it based on it. The roller diameter is 4mm. With a pitch circle of 87mm, the safety factor reaches 6.35 for static load and 3.6 for dynamic load.

[0148] To facilitate the installation of the absolute encoder, the stator hollow shaft is extended using an adapter. Since the motor itself does not have a connection interface for the extended shaft, interference fit and bonding are required for connection. Figure 5The absolute encoder's magnetic ring is connected to the harmonic reducer output shaft via a stainless steel adapter sleeve using an H7 / P6 interference fit. The readhead is fixed to the inner side of the joint housing, 0.5 mm from the magnetic ring.

[0149] The connecting flange is immersed in glue to ensure axial positioning. The flange is provided with notches to ensure circumferential positioning.

[0150] Considering the sealing of the harmonic reducer, a labyrinth seal is used at the high-speed end and an O-ring seal is used at the low-speed end. Chamfers and process holes are designed to facilitate assembly. The gaps between all parts with relative motion are appropriate, neither too small to cause deformation and disappearance, nor too large to cause waste of space. Figure 6 shown.

[0151] (1) Power system design

[0152] In addition to driving the brushless DC motor, the lightweight arm hardware system also needs to power the FPGA controller, motor drive, and sensor processing circuits. This requires the lightweight arm's power supply system to provide multiple voltage levels. Using multiple DC power inputs to supply different modules would require a large number of external devices. Therefore, this design uses a single-level input and then converts the single level to meet the different voltage requirements of each module. This improves system integration and ease of use, while reducing external wiring.

[0153] According to the working mode of the power tube in the stable power supply circuit, the stable power supply can be divided into two categories: switching power supply and linear power supply.

[0154] Linear power supplies offer high voltage and load stability, low output ripple voltage, and fast circuit transient response. However, they also have high power consumption and low efficiency. They are bulky and heavy, making them difficult to miniaturize, and require large-capacity filter capacitors. They are particularly suitable for applications requiring high power stability.

[0155] Switching power supplies have low power consumption, high efficiency, small size, light weight, wide voltage regulation range and good voltage regulation effect. However, they can generate spike interference and harmonic interference, which can easily affect the normal operation of the entire power supply system. They are suitable for applications with high requirements on power supply space.

[0156] Considering the advantages and disadvantages of switching power supply and linear power supply, this paper adopts a new power supply system that combines switching power supply and linear power supply. Figure 7 As shown, the power supply system principle is as follows Figure 8 shown.

[0157] After the external power supply is introduced, a portion is used to directly drive the motor, while the remaining portion first passes through the DC / DC converter circuit to obtain a series of rough voltage values, then undergoes linear voltage regulation before being output to power the components. The main DC / DC converter module uses TI's LM5176 controller to convert the 48V input into 12V / 20A and 5V / 10A outputs. The linear voltage regulator module uses the LT3045 ultra-low-noise LDO, and the multi-stage filtering circuit includes π-type LC filtering and tantalum capacitor decoupling. The common-mode inductor is TDK's ACM2012-102-2P-T00 to suppress the switching power supply's noise above 10MHz.

[0158] (2) Xilinx FPGA-based underlying control system

[0159] The underlying controller uses an FPGA as the main control chip. By implementing hardware units within the FPGA, the number of electrical components within the joints is significantly reduced, thereby reducing the weight of the robotic arm and paving the way for modular joints. Furthermore, the FPGA's flexible parallel processing programming and soft-core development tools not only allow sensor data to be processed locally at the joint level, reducing transmission noise, but also enable the simultaneous and rapid implementation of boundary layer equations, motor drive, and communication functions. These advantages of the FPGA make it far more efficient than traditional controllers such as microcontrollers and DSPs.

[0160] At the same time, because FPGAs allow users to customize their internal logic and functions and are capable of unlimited reconfiguration, coupled with PC-based CAD-assisted design software and powerful simulation tools, users can design their own ASIC devices to meet specific user-defined applications, greatly increasing the flexibility of electronic system design. Unlike ASICs, FPGAs are simply standard cell arrays without the functions of conventional ICs. However, users can reprogram their internal components as needed using specialized layout and routing tools, designing their own specialized integrated circuits in a very short time, greatly enhancing design flexibility. Furthermore, because they perform parallel processing purely in hardware and do not consume CPU resources, they can achieve very high system performance.

[0161] The FPGA uses the XC2V3000BQ728 chip from Xilinx's Virtex2 series. Xilinx's Virtex2 series FPGAs are highly reliable and stable, and can be embedded with MicroBlaze. MicroBlaze's peripheral interfaces include the serial peripheral interface SPI, the universal serial port UART, parallel I / O, SDRAM controller, memory interface, Ethernet, etc. The selected XC2V3000BQ728 chip has rich on-chip resources, including 2 phase-locked loops, 20,060 logic units, and a maximum of 728 I / O pins. The designed underlying control system with FPGA as the core is as follows: Figure 8 The hardware part includes sensor data acquisition, robotic arm joint motor control, PPSECO communication, etc. The software part completes the coordination and data processing of each module, such as joint angle calculation, motor speed calculation, sensor signal filtering, etc. The specific structure of the system is as follows Figure 9 shown.

[0162] 3) Upper control system based on PCI bus

[0163] The main purpose of a robotic arm is to manipulate objects in the external environment. Therefore, in addition to precise position control, it is also expected to have the ability to sense forces when interacting with the external environment, achieve compliant control, and protect itself and the manipulated objects from damage. To achieve these functions, a high-performance hardware platform is required. This hardware platform should have the following features:

[0164] 1. It has high-speed communication function, which can upload the entire robot sensor data collected in real time to the upper controller in the shortest possible control cycle, and at the same time feed back the control signal of the upper controller to the bottom controller.

[0165] 2. It has high-speed processing capabilities and can quickly implement complex control algorithms.

[0166] 3. It can communicate with PC at high speed, facilitating human-computer interaction.

[0167] Therefore, the upper hardware control structure of DSP / FPGA based on PCI bus is adopted. The control system structure is as follows Figure 10 The inverse kinematic solution uses the DH parameter method, and the feedforward compensation is calculated based on the Newton-Euler dynamics model. The dual-port RAM uses an IDT70V09L chip with a memory depth of 18K × 16 bits. The second-order Butterworth filter cutoff frequency is set to 50Hz, and group delay is eliminated using a phase compensation algorithm.

[0168] At the bottom layer, we use a Xilinx FPGA as the main control chip. Through the FPGA's flexible parallel processing programming and soft-core development tools, the underlying layer not only processes sensor data but also rapidly implements motor drive, communication, and data fusion functions. These advantages of the FPGA make its processing efficiency far superior to traditional controllers such as microcontrollers and DSPs.

[0169] At the top level, we employ a DSP / FPGA control structure based on the PCI bus. The PCI bus primarily connects the PC and DSP units via a PCI bridge controller, enabling convenient human-machine interface operations at the top level through teleoperation, force feedback, and other methods. The DSP is the primary control chip, utilizing a high-speed floating-point DSP unit to efficiently calculate desired trajectories, gravity compensation, and kinematics. The FPGA serves only as a precise bus communication interface, obtaining joint state information and transmitting joint commands from the underlying FPGA via a serial bus, and then communicating with the DSP via a parallel bus. The purpose of the FPGA design is to separate bus communication from the traditional DSP unit, reducing the amount of computation required by the DSP.

[0170] Based on the aforementioned DSP / FPGA-FPGA hardware control framework, the lightweight robotic arm can be efficiently controlled. The upper layer of the PCI-based DSP / FPGA board performs extensive kinematic and matrix operations, while the lower layer of the FPGA board achieves a high control frequency and can instantly update the desired torque. Both communicate via a 25Mbps PPSECO serial bus.

[0171] (4) High-speed serial bus based on low-voltage differential signaling (LVDS)

[0172] To achieve real-time control of the robotic arm, the top layer requires rapid feedback on joint position and velocity from the bottom layer (the joint layer), which also requires immediate updates of desired torque. This is particularly critical during rapid robotic arm motion. Furthermore, to minimize the number of wiring in the robotic arm, the control architecture requires high-speed serial bus communication to meet these control requirements.

[0173] Point-to-point low voltage differential signaling (PP-LVDS) complies with the ANSI / TIA / EIA-899M-LVDS bus standard. It allows point-to-point LVDS serial connections and can be expanded to point-to-multipoint (i.e., M-LVDS). The maximum communication rate is 500Mbps. Figure 10As shown, the data is connected via two low-voltage differential data lines and a 100-120 Ω termination resistor RT. The data is driven by an LVDS driver chip and processed in the top and bottom FPGAs. The checksum field is generated using the CRC-16 algorithm, and the node identification field is an 8-bit address code. The impedance matching circuit uses a 120Ω termination resistor connected in parallel to the ends of the LVDS differential lines. The communication cycle is configured to be 1ms, meeting the 100Hz control frequency requirement of the robotic arm. The FPGA is configured with an edge-triggered interrupt controller. When the encoder data changes by more than ±1LSB or a new command packet is received, an LVDS transmission interrupt is immediately triggered, with a response latency of <10μs.

[0174] The receiving and transmitting modes of the LVDS serial bus are both written in the hardware description language VHDL. Making this independent module into an IP core is conducive to the implementation of the module on different FPGA chips and software upgrades. At the same time, the bus's transceiver module is an intelligent IO module implemented by FPGA. It can complete data encoding and decoding, data packet address identification, interrupt generation, CRC data verification and other functions without the intervention of the microprocessor, greatly reducing the load on the microprocessor. The serial bus system diagram is as follows Figure 11 shown.

[0175] Before data transmission, the nodes are assigned addresses. To prevent the joint from receiving the data frames sent by itself, the joint address includes the receiving address and the sending address. The process of system data transmission is divided into the following steps:

[0176] (1) During actual data transmission, the DSP receives the PC's operation control instructions from the dual-port RAM through the PCI bus. After internal calculations, it sends the desired position, speed, and nonlinear compensation values ​​to the FPGA in the PCI board.

[0177] (2) The FPGA in the PCI board is responsible for packaging the data of each joint and then sending the data packet to the robotic arm via the bus according to a fixed cycle.

[0178] (3) After the joint receives the data packet through the interrupt, it unpacks and judges the data and puts it into the buffer area, and then packages its own state variables and sends them to the bus, and then waits for the joint to read, calculate and control;

[0179] (4) The FPGA on the PCI card is responsible for interrupt reception, unpacking, judgment, and then sending the data address inherent to the joint, waiting for the DSP to read the joint data;

[0180] Example 3 describes in detail the specific principles and steps of the robot arm control algorithm of the present invention. The robot arm control algorithm includes: the design of the speed loop and position loop regulator based on the joint motor servo system; the speed loop design; the position loop design and the current loop controller design.

[0181] (1) Design of speed loop and position loop regulator based on joint motor servo system

[0182] The current loop uses i d =0 control strategy, the role of the current loop is to transform the transfer function of the inner loop control object and improve the speed of the system; timely suppress the interference inside the current loop, especially the back electromotive force interference; and limit the maximum current so that the system has a sufficiently large acceleration torque and ensures the safe operation of the system. Since the electrical time constant of the motor is much smaller than the mechanical time constant, it can be considered that the motor torque reaches the given value instantaneously. Therefore, a well-designed current loop can be approximated as a small inertia link, and the mathematical model of the PMSM system with current loop control can be approximated as a second-order system. The joint motor servo system adopts a three-loop series structure, with the current loop as the inner loop and the speed loop and position loop as the outer loop. In order to suppress the influence of load disturbances, inertia changes and torque fluctuations, the speed loop is controlled by a 2-degree-of-freedom PID regulator; in order to eliminate the static error of position tracking, the position loop adopts a feedforward control scheme. The control block diagram of the joint motor servo system is as follows Figure 11 shown.

[0183] (2) Speed ​​loop design

[0184] The speed of the motor in the joint is calculated using the M / T method, which has good accuracy over the entire speed range of the motor.

[0185] Servo systems place high demands on the speed loop, as it suppresses load disturbances, overcomes inertia changes, and dampens torque fluctuations. The speed loop essentially determines the servo system's frequency response and accuracy. The speed loop is typically controlled using a PI regulator, which offers a simple structure and excellent steady-state accuracy. However, as a single-degree-of-freedom controller, it is generally difficult to achieve both good tracking performance and interference rejection in the system's dynamic response. A two-degree-of-freedom controller, on the other hand, can meet both tracking and interference rejection requirements.

[0186] Assuming that the model of the permanent magnet synchronous motor controlled by the current loop is P(s), the speed loop control diagram is obtained, as shown in Figure 12 shown.

[0187] The actual motor speed is w, the reference input speed is w*, and the regulators C1(s) and C2(s) use PID control:

[0188] C1(s)=(1-α)K p +K I ·1 / s+(1-b)K D ·s (1)

[0189] C2(s)=aK p +b·KD ·s (2)

[0190] The transfer function and sensitivity function of the closed-loop system are:

[0191]

[0192] Since the sensitivity function of formula (4) does not contain parameters a and b, it is only related to K p , K I , K D The parameters are related, so by adjusting this set of parameters, the anti-interference ability of the system can be changed, and by adjusting parameters a and b, the dynamic response characteristics of the system can be changed. Therefore, by adjusting these two sets of parameters, two-degree-of-freedom control of the speed can be achieved.

[0193] (3) Position loop design

[0194] The positioning accuracy of the motor servo system depends on the position accuracy of the magnetic code disk, the motor position feedback element. The line count of the magnetic code disk we use is 1200PPR. The code disk information is processed by 4 times the frequency, and its resolution can reach 0.075°. The control period of the position loop is 1ms.

[0195] The design of the position loop is relatively simple, consisting of a proportional regulator and a feedforward compensation link, such as Figure 11 The position control link is shown in the dotted line. Adding a low-order filter to the feedforward link can effectively suppress position fluctuations and reduce overshoot of the position control system. Proportional regulator gain K p =1.0, speed feedforward coefficient K v =0.95; the trajectory smoothing filter uses a second-order Butterworth low-pass filter with a cutoff frequency of 50 Hz, and the group delay is compensated by the inverse filter phase; the position correction synthesis module is implemented in the FPGA with a fixed-point adder, and the output resolution is 16 bits.

[0196] (4) Design of current loop controller

[0197] The current correction stage requires high bandwidth, fast response, and excellent tracking performance. Therefore, a high proportional gain and a high adjustment frequency are required, which means short system latency. This necessitates a simple and practical current regulator algorithm. The most commonly used current correction method is the PI correction method, which typically calibrates the current loop to either a typical Type I or Type II system. The typical Type I system is characterized by low overshoot and good tracking performance, while the typical Type II system is characterized by good interference immunity.

[0198] When designing a current loop using the PI control principle, cross-feedback caused by back EMF occurs. The electromagnetic time constant Tl of a real system is much smaller than the electromechanical time constant Tm, and the current regulation process is often much faster than the speed change process. To the current loop, back EMF is merely a slowly changing disturbance. During the rapid regulation of the current regulator, back EMF E can be considered essentially unchanged, or ΔE ≈ 0. Because back EMF is proportional to speed, its influence on the current loop increases when the motor runs at high speeds. This can lead to deviations in amplitude and phase between the output current of the current control loop and its set current signal, making it difficult for the motor current to maintain its sinusoidal properties and deteriorating the current loop's control performance.

[0199] According to the PMSM mathematical model in the dq rotor synchronous rotating coordinate system, the stator voltage equation can be rewritten as:

[0200]

[0201] in:

[0202]

[0203] From the formula, we can see that the back EMF can be regarded as a detectable interference quantity and compensated by feedforward control. The corresponding compensation amount is:

[0204]

[0205] The compensation amount is superimposed on the output of the current regulator to form a voltage feedforward control, which can approximately offset the actual coupling amount related to the speed, thereby eliminating the rotating potential p n ωψ f That is, the influence of back electromotive force on the current control loop.

[0206] Figure 13 The control block diagram of the current control loop with back-electromotive force compensation adopts a modular design concept and designs the motor current sampling, SPWM generation module, vector coordinate transformation module, code disk information quadruple frequency and M / T method speed measurement module and PI regulation module to realize the motor magnetic field oriented control.

[0207] Example 1: Robotic arm grasping operation process

[0208] The AGV mobile platform activates the lidar to scan the working area, builds an environmental map and plans an obstacle avoidance path, and transports the robotic arm to the vicinity of the target scrap steel.

[0209] The robotic arm collects scrap steel shape data through the end-point vision sensor, and the upper-level DSP calculates the grasping posture based on the point cloud information to generate a seven-degree-of-freedom joint motion trajectory.

[0210] The underlying FPGA reads the joint encoder data in real time, calculates the speed of each motor using the M / T method, and drives the EC60flat and EC45flat motors to run according to the trajectory;

[0211] The disc motor of the dexterous gripper drives the synchronous belt through the harmonic reducer, which drives the wire mechanism to realize the linkage closure of the four fingers, increasing the gripping friction of the plastic packaging.

[0212] The cross roller bearing and four-point contact bearing work together to offset the joint overturning moment, ensuring the end repeat positioning accuracy of ±1mm.

[0213] The upper system receives joint torque sensor data in real time through the LVDS bus and dynamically corrects trajectory parameters to prevent scrap steel from slipping.

[0214] Example 2: Power and control collaborative workflow

[0215] After the external 48V power supply is input, the main DC / DC module outputs 12V / 5V voltage, the linear voltage regulator module performs secondary voltage regulation to power the FPGA / DSP, and the flyback isolated power supply drives the brushless motor.

[0216] The underlying FPGA synchronously processes the encoder's quadruple frequency signal and SPWM waveform generation to drive the joint motor at a frequency of 20kHz;

[0217] The upper DSP receives the PC-side grabbing instructions through the PCI bus, solves the inverse kinematics and generates feedforward compensation parameters, which are then buffered in the dual-port RAM and sent to the joints through the LVDS bus.

[0218] After parsing the data packet, the FPGA updates the current loop setting value and implements magnetic field oriented control in combination with the vector coordinate transformation module.

[0219] The power ripple monitoring module provides real-time feedback on voltage fluctuations, triggering the FPGA to switch to the backup power path. At the same time, the DSP operates at a reduced frequency to ensure system safety.

[0220] Example 3: Multi-loop control algorithm implementation process

[0221] The motor current sampling module obtains i d 、i q Component, based on i d =0 strategy generates q-axis current reference value;

[0222] The feedforward compensation module calculates u dc =p n ωi q with u qc =-p n ωi d -p n ωψ f, added to the PI regulator output to offset the back electromotive force interference.

[0223] The forward channel regulator is (1-α)K p +K I / s+(1-b)K D s improves trajectory tracking speed;

[0224] The feedback channel regulator is αK p +bK D s suppresses speed fluctuations caused by sudden load changes;

[0225] The parameter tuning module automatically optimizes the α and b values ​​according to the step response, and uses the Ziegler-Nichols critical proportionality method to adjust K first. p Make the system oscillate with equal amplitude and record the critical gain K cr With period T cr , according to a=0.6K cr ,b=0.5T cr Initialize the parameters to control the overshoot within 5%. The step response self-tuning adopts the critical proportionality method, with initial parameters α=0.6, b=0.4; the integral time constant T of the forward channel regulator is I =K p / K I Set to 10ms, differential time constant T D =K D / K p The FPGA is equipped with a dedicated multiplier and accumulator MAC unit to accelerate PID calculation.

[0226] Proportional regulator with K p =1.0 generates the basic position error correction;

[0227] Speed ​​feedforward module introduces K v =0.95 pre-compensation, output after smoothing by second-order Butterworth filter to eliminate trajectory lag;

[0228] The position filter filters out high-frequency vibration noise with a cutoff frequency of 50Hz to ensure smooth end motion.

[0229] Example 4: High-speed communication and maintenance process

[0230] The joint controller packages the encoder data into a data frame containing a 16-bit CRC check code and uploads it via the LVDS differential driver chip at a rate of 500Mbps.

[0231] After the upper-layer FPGA parses the data frame, it writes the joint status into the dual-port RAM through the interrupt mechanism for real-time call by the DSP.

[0232] After removing the shoulder joint housing, directly replace the EC60flat motor and cross roller bearing assembly;

[0233] The plastic packaging of the dexterous gripper can be quickly disassembled through a snap-fit ​​structure, and the wire mechanism remains in place to avoid recalibration;

[0234] The harmonic reducer's hollow shaft design allows for tool-free assembly and disassembly, shortening the maintenance cycle to less than 30 minutes.

[0235] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solutions, which should all be included in the scope of the claims of the present invention.

Claims

1. A control system for a mechanical device for scrap steel grabbing, characterized by: include: The mobile platform uses an omnidirectional AGV mobile robot with a laser radar and path planning module installed on its top; A seven-degree-of-freedom collaborative robot arm, whose base is fixed on the mobile platform, includes a shoulder joint, a wrist joint and a hollow shaft structure, has a working radius of ≥900mm, and a repeatability accuracy of ±1mm; The end effector is installed at the end of the robotic arm and adopts a coupled transmission dexterous gripper to drive multi-finger linkage through a steel wire mechanism; A power supply system, electrically connected to the mobile platform, the robotic arm, and the end effector, adopting a hybrid power supply architecture combining a switching power supply and a linear power supply; The underlying control system is integrated into the joints of the robotic arm and implements motor drive and real-time processing of sensor signals based on FPGA; The upper control system communicates with the lower control system via the PCI bus and uses a DSP / FPGA architecture to perform kinematics calculations and trajectory planning; A high-speed serial bus connects the bottom layer and the upper layer control system based on low voltage differential signal (LVDS) to transmit joint status data and control instructions; The control algorithm module is deployed in the upper and lower control systems, including the current loop controller, the speed loop controller and the position loop controller, wherein the current loop adopts i d =0 control strategy.

2. The control system for a mechanical device for scrap steel grabbing according to claim 1, characterized in that: The joint structure of the seven-degree-of-freedom collaborative robot arm includes: The shoulder joint is fixed to the upper surface of the mobile platform through a flange connection, and an outer rotor brushless motor and a harmonic reducer are integrated therein, and the output shaft of the motor is rigidly connected coaxially with the wave generator of the harmonic reducer; The wrist joint is connected in series with the output end of the harmonic reducer of the shoulder joint through a hollow shaft sleeve, and has a built-in outer rotor brushless motor and a four-point contact bearing. The inner ring of the bearing is fixed to the motor rotor, and the outer ring is connected to the joint housing; The dual encoder system includes a relative encoder on the motor side and an absolute encoder on the joint side, where: The magnetic ring of the relative encoder is fixed to the rotor end face of the outer rotor brushless motor and rotates coaxially with the motor; The reading head of the absolute encoder is installed on the inner side of the joint housing, and its magnetic ring is interference-fitted with the output shaft of the harmonic reducer through an adapter; Wires run through the hollow shafts of the shoulder joint and the wrist joint, and the wires are electrically connected to the FPGA signal interface of the underlying control system.

3. The control system for a mechanical device for scrap steel grabbing according to claim 1, characterized in that: The dexterous gripper of the end effector comprises: The drive unit is composed of a disc motor and a harmonic reducer, wherein the disc motor is fixed inside the base joint of the finger, and its output shaft is connected to the input shaft of the harmonic reducer through a synchronous belt mechanism; The multi-finger linkage mechanism includes four sets of parallel steel wire transmission assemblies. One end of each steel wire assembly is wound around the output wheel of the harmonic reducer, and the other end passes through the guide pulley of the finger joint and is fixed to the end knuckle. A plastic packaging piece, which is detachably covered on the outside of the finger joint by a snap-fit ​​structure, and has a friction pattern on its inner surface; The transmission ratio of the wire transmission assembly and the synchronous belt mechanism is configured to be 1:2 to 1:4, and the closing stroke of the terminal knuckle is controlled by the output angle of the harmonic reducer.

4. The control system for a mechanical device for scrap steel grabbing according to claim 1, characterized in that: The power supply system comprises: A main DC / DC conversion module, whose input end is electrically connected to an external 48V DC power supply, and whose output end includes a first output end and a second output end, wherein the first output end provides a 12V voltage and the second output end provides a 48V voltage; The linear voltage regulator module is connected to the 12V output terminal of the main DC / DC conversion module through a copper busbar, and provides a 5V regulated power supply to the FPGA, DSP and sensor interface after passing through a multi-stage filtering circuit; The motor drive dedicated isolated power supply adopts a flyback topology. Its primary side winding is connected to the 48V output terminal of the main DC / DC conversion module, and the secondary side winding supplies power to the brushless motor drivers of each joint through a magnetic isolation transformer; A common-mode inductor is provided between the linear voltage stabilization module and the motor drive dedicated isolation power supply to suppress high-frequency conducted interference.

5. The control system for a mechanical device for scrap steel grabbing according to claim 1, characterized in that: The underlying control system includes: An FPGA chip, whose signal input terminal is electrically connected to the magnetic encoder of each joint, is configured to convert the original encoder signal into a quadruple frequency pulse signal through a code disk quadruple frequency module; An M / T speed measurement module is integrated into the FPGA chip, receives the quadruple frequency pulse signal and calculates the real-time speed of the joint motor; An SPWM generation module generates a three-phase SPWM waveform for driving a brushless motor according to the output of the M / T method speed measurement module and the current loop control instruction; The communication interface of the FPGA chip is connected to the DSP processor of the upper control system through a high-speed serial bus to transmit the real-time speed and joint position data; The resolution of the code disk quadruple frequency module is adapted to the line number of the magnetic encoder, ensuring that the joint positioning accuracy meets ±1mm.

6. The control system for a mechanical device for scrap steel grabbing according to claim 1, characterized in that: The upper control system includes: a DSP processor, electrically connected to an external control terminal via a PCI bus, configured to receive a grasping instruction and perform an inverse kinematic solution of the manipulator to generate a desired trajectory of joint angles; A dual-port RAM memory, wherein the first port thereof is electrically connected to the PCI bus and the second port thereof communicates with the FPGA of the underlying control system via an LVDS bus interface, thereby realizing bidirectional buffering of control instructions and joint status data; A feedforward compensation module, integrated into the DSP processor, receives the desired trajectory of the joint angle and calculates gravity compensation and nonlinear friction force feedforward to generate compensation parameters; A trajectory smoothing filter is connected between the DSP processor and the feedforward compensation module, and uses a second-order Butterworth filter to perform phase compensation and high-frequency noise suppression on the desired trajectory; The compensation parameters and the expected trajectory of the joint angle are sent to the underlying control system via the LVDS bus to drive the motors of each joint to move according to the planned trajectory.

7. The control system for a mechanical device for scrap steel grabbing according to claim 1, characterized in that: The communication protocol of the high-speed serial bus includes: The data packet structure includes a checksum field and a node identification field; the checksum field is used to verify data integrity, and the node identification field uniquely identifies each joint controller of the robotic arm; the data packet is generated by the joint controller, encoded by the LVDS differential driver chip, and then sent to the FPGA of the upper control system; The interrupt trigger mechanism is configured to trigger data transmission immediately via hardware interrupt when joint status data is updated or control instructions arrive; Impedance matching circuit, set between the transmitting end and the receiving end of the LVDS bus, suppresses signal reflection to ensure transmission stability; The data packets are transmitted via the LVDS bus in a time-division multiplexing manner, and the communication cycle between the upper and lower control systems is adapted to the real-time control requirements of the robotic arm.

8. The control system for a mechanical device for scrap steel grabbing according to claim 1, characterized in that: The current loop controller implements: The motor current sampling module has an input end electrically connected to the three-phase current sensor of each joint brushless motor and an output end connected to the Clarke transformation module; A vector coordinate transformation module is configured to perform Clarke transformation and Park transformation to convert the three-phase current into i d 、i q Component, where i d The reference value is set to 0; Clarke transformation converts the three-phase current i a ,i b ,i c Convert to i α ,i β , Park transformation is further mapped to the rotor coordinate system i d ,i q The feedforward compensation module calculates the compensation value in real time through the FPGA's built-in DSP slice. The SPWM generation module adopts a symmetrical regular sampling method with a carrier frequency of 20kHz. The feedforward compensation module receives the motor rotor position signal and speed signal and calculates the back electromotive force compensation voltage u dc =p n ωi q and u qc =-p n ωi d -p n ωψ f , superimposed on the output end of the current loop; PI adjustment module, for i q The component errors are subjected to proportional-integral operation to generate the q-axis voltage command; An SPWM generation module generates a three-phase drive waveform according to the q-axis voltage command and the compensation voltage, and outputs the waveform to the brushless motor via an isolation drive circuit; The vector coordinate transformation module, feedforward compensation module and PI adjustment module are all integrated into the FPGA chip of the underlying control system in the form of hardware logic.

9. The control system for a mechanical device for scrap steel grabbing according to claim 1, characterized in that: The speed loop controller adopts a two-degree-of-freedom PID structure, including: The forward channel regulator receives the speed reference signal from the upper control system at its input, and its transfer function is C1(s)=(1-α)K p +K I ·1 / s+(1-b)K D s, the output end is connected to the command input end of the current loop controller; Feedback channel regulator, whose input end is electrically connected to the real-time speed signal of the joint encoder, and the transfer function is C2(s)=aK p +b·K D s, the output end of which is superimposed on the output end of the forward channel regulator; The parameter tuning module is integrated into the DSP processor of the upper control system, dynamically adjusts the α and b parameters according to the step response curve, and sends the tuned parameters to the FPGA of the lower control system via the LVDS bus; The operation logic of the forward channel regulator and the feedback channel regulator is solidified in the FPGA chip with the hardware description language, and the regulator output is connected to the current loop controller i q The command end is electrically connected.

10. The control system for a mechanical device for scrap steel grabbing according to claim 1, characterized in that: The position loop controller comprises: The proportional regulator receives the position command from the upper control system and the position feedback signal from the joint absolute encoder at its input, and generates the basic position correction at its output; A speed feedforward module is electrically connected to the DSP processor of the upper control system, receives the trajectory planning speed instruction signal generated by the DSP processor, and generates a pre-compensation amount through a feedforward coefficient K; A trajectory smoothing filter, the input of which is connected to the output of the velocity feedforward module, and uses a second-order low-pass filter to perform phase lag compensation on the pre-compensation amount; A position correction synthesis module, which superimposes the basic position correction amount and the filtered pre-compensation amount to generate a final position control instruction; Position feedback interface, which collects the position data of each joint's absolute encoder in real time through the LVDS bus and uploads it to the DSP processor of the upper control system; The final position control instruction is converted into a motor drive signal by the FPGA of the underlying control system to drive the joint motor to move according to the planned trajectory.

Images