A multi-modal adaptive deep loosening mechanism based on soil resistance perception

Abstract

The present invention discloses a multi-modal adaptive deep loosening mechanism based on soil resistance perception, including an integral frame, a deep loosening unit, a control system and an energy device. The deep loosening unit detects the soil resistance in real time through a sensor assembly and transmits it to the control system. The control system automatically switches between three modes: traditional rigid connection deep loosening, self-excited vibration deep loosening and forced vibration deep loosening according to the size and change characteristics of the soil resistance. When the soil resistance is less than the first preset threshold, the rigid connection mode is adopted; when the resistance is between the first and second preset thresholds, the self-excited vibration mode is enabled; when the resistance is greater than the second preset threshold, the vibration motor is started to achieve forced vibration. The control system not only monitors the absolute value of the resistance, but also analyzes its time characteristics to achieve intelligent threshold adjustment. The present invention solves the problems of high energy consumption and poor adaptability of traditional deep loosening mechanisms, and has the advantages of intelligent operation mode switching, optimized energy utilization, and adjustable vibration parameters.

Description

Technical Field

[0001] The present invention belongs to the technical field of agricultural machinery, and in particular relates to a multi-modal adaptive deep loosening mechanism based on soil resistance perception. Background Art

[0002] Subsoiling is an important soil tillage technique that effectively breaks up the plow base, loosens the soil, and improves soil structure, making it crucial for increasing crop yields and improving long-term soil health in planted areas. Traditional rigid-connection subsoiling methods are the most common (e.g., Chinese patents CN217336327U and CN118661488B). However, due to the high tillage depth, this inevitably leads to high resistance and energy consumption. To reduce traction resistance during subsoiling operations and improve soil fragmentation, vibration subsoiling technology has emerged, primarily encompassing self-excited and forced vibration methods.

[0003] Self-excited vibration deep tillage mainly relies on the periodic changes in resistance generated during the interaction between the deep tillage shovel and the soil to induce vibration of the entire system (such as Chinese patents CN209806378U, CN205864970U, CM206332984U). This method has a relatively simple structure and may produce a good drag reduction effect under specific working conditions. However, the frequency and amplitude of self-excited vibration depend largely on the soil properties and operating speed, making it difficult to accurately control and adjust. When the working conditions change, the vibration may become unstable or even disappear, resulting in large fluctuations in the operating results and poor adaptability to different soil types.

[0004] Forced vibration deep loosening uses an independent excitation device, such as an eccentric wheel, a crank-connecting rod mechanism, or a vibration motor, to generate an excitation force, forcing the deep loosening components to vibrate at a set frequency and amplitude (such as Chinese patents CN118805458A, CN105940810A, CN101480120B, and CN114402710B). Compared with the self-excitation type, forced vibration has a better soil-breaking effect and better controllability of vibration parameters. However, due to the introduction of an additional energy device, it will inevitably lead to increased power consumption, or a very limited adjustment range and a complicated adjustment process. This leads to the following problems:

[0005] 1. Insufficient adaptability to operating conditions: Fixed vibration parameters are difficult to adapt to the complex and changing field conditions, such as soil type, compaction, and moisture content. Under certain conditions, excessive parameters may lead to increased energy consumption and excessive soil disturbance, while under other conditions, insufficient parameters may result in poor drag reduction and soil crushing.

[0006] 2. Difficulty Balancing Energy Consumption and Efficiency: Due to the inability to adjust based on real-time soil resistance, equipment often struggles to maintain high efficiency and low energy consumption under varying operating conditions. This often results in either high power designs to accommodate the worst-case scenario, resulting in widespread waste, or low efficiency when resistance changes.

[0007] 3. Lack of intelligent and refined operation capabilities: Traditional forced vibration devices lack the ability to perceive the operation process and cannot realize intelligent and refined parameter adjustment and operation mode switching based on actual load.

[0008] Therefore, there is an urgent need to develop a new deep plowing technology and equipment that can overcome the above limitations, sense the workload in real time, and intelligently and adaptively adjust the vibration mode and parameters accordingly, so as to meet the needs of modern precision agriculture for efficient, energy-saving, and high-quality soil cultivation. Summary of the Invention

[0009] The purpose of the present invention is to provide a multimodal adaptive deep loosening mechanism based on soil resistance perception, aiming to solve the technical problems of poor adaptability, high energy consumption and low efficiency of existing deep loosening technology under complex and changeable soil conditions. By sensing the soil resistance characteristics in real time, it automatically switches between the three modes of traditional rigid connection deep loosening, self-excited vibration deep loosening and forced vibration deep loosening, optimizes the operation parameters according to different soil conditions, and realizes the optimization of energy utilization. The present invention can operate efficiently in low-resistance areas and ensure the soil-breaking effect in high-hardness soil areas. At the same time, it adopts precise vibration parameter control to make the deep loosening operation precise, intelligent and low-energy, significantly improving the quality and efficiency of farmland soil cultivation.

[0010] The present invention is achieved through the following solutions:

[0011] A multi-modal adaptive deep loosening mechanism based on soil resistance sensing, characterized by comprising an integral frame, a deep loosening unit, a control system, and an energy device; the deep loosening unit is mounted on the integral frame, the control system is connected to the deep loosening unit for signal transmission, and the energy device provides power or energy for the control system and energy-consuming components in the deep loosening unit;

[0012] The deep loosening unit includes a rear bracket, an electric push rod, a middle bracket, an upper bracket, a tension pin, a sensor assembly, a center bracket, a deep loosening shovel, a spring system and a vibration motor assembly; the deep loosening shovel is fixed to the bottom of the rear bracket; the middle bracket is hinged to the rear bracket through a rotating shaft, and the upper bracket is hinged to the rear bracket, and the rear bracket, the upper bracket and the middle bracket together constitute a parallel four-bar linkage mechanism; the electric push rod is fixed to the front end of the rear bracket, and its extended end can abut against the middle bracket after being extended; the spring system abuts between the rear bracket and the middle bracket; the sensor assembly is installed on the tension pin; the center bracket is connected to the front end of the middle bracket through a tension pin, and the top end of the center bracket is rotatably connected to the bottom of the upper bracket; the vibration motor assembly is installed at the front end of the upper bracket;

[0013] The subsoiling unit senses soil resistance in real time through a sensor assembly and transmits the signal to the control system. The control system decides whether the subsoiling mechanism switches between the following three subsoiling modes based on the size and variation characteristics of the soil resistance and generates corresponding control instructions to adjust the working state of the subsoiling unit:

[0014] When the soil resistance is less than the first preset threshold value T1, the deep loosening mechanism enters the rigid connection deep loosening mode;

[0015] When the soil resistance is between the first preset threshold value T1 and the second preset threshold value T2, the subsoiling mechanism enters the self-excited vibration subsoiling mode;

[0016] When the soil resistance is greater than or equal to the second preset threshold value T2, the deep plowing mechanism enters the forced vibration deep plowing mode.

[0017] Furthermore, in the rigid connection deep plowing mode, the control system controls the electric push rod to fully extend and press against the middle bracket, forming a stroke lock, and the spring system does not participate in the work. At this time, the rear bracket cannot rotate around the rotating shaft; in the self-excited vibration deep plowing mode, the control system controls the electric push rod to retract, releasing the restriction on the rotation of the rear bracket, and the spring system starts to work. The rear bracket can rotate around the rotating shaft. During the movement, the soil resistance and the spring system form a self-excited vibration system, and the deep plowing shovel produces periodic vibrations; in the forced vibration deep plowing mode, the control system starts the vibration motor assembly, and transmits the vibration to the deep plowing shovel through the vibration motor assembly, the upper bracket and the rear bracket.

[0018] Furthermore, in the forced vibration deep loosening mode, the control system adjusts the vibration parameters according to the real-time monitoring of the soil resistance characteristics:

[0019] At the initial start-up of the mode, a frequency sweep test is performed in the range of 1-50Hz. The response signal of the sensor assembly is monitored and the resistance attenuation rate at each frequency is calculated. The frequency with the highest attenuation rate is determined as the optimal resonant frequency. The operating frequency of the vibration motor is adjusted to within ±0.1Hz of the optimal resonant frequency. Fine-tuning is performed in real time based on the resistance change trend.

[0020] At the same time, the control system controls the vibration amplitude and waveform type by adjusting the input voltage waveform and phase parameters of the vibration motor.

[0021] Furthermore, the control system monitors the magnitude of soil resistance and analyzes the time characteristics of the soil resistance changes. It uses a sliding time window to smooth the continuously collected resistance data and judges its persistence characteristics based on the resistance change rate. Only when the soil resistance continues to exceed the corresponding preset threshold within the sliding time window, the switching of the deep plowing mode is triggered. By setting the sliding time window, when the deep plowing shovel encounters an instantaneous resistance mutation caused by an obstacle, resulting in instantaneous resistance fluctuations, the current deep plowing mode is still maintained to avoid frequent switching of the deep plowing mode.

[0022] Preferably, the sliding time window is set to 1 to 3 seconds.

[0023] Furthermore, the control system adopts an adaptive threshold adjustment method based on machine learning to dynamically adjust the preset threshold for switching deep loosening mode, including the following process:

[0024] Resistance data were collected in experimental fields with known soil types, moisture content, and compaction, and a database of the correspondence between soil resistance characteristics and soil conditions was established;

[0025] Extracting time domain features and frequency domain from the resistance signal collected by the sensor assembly;

[0026] An improved support vector machine algorithm combined with a decision tree model was used to construct a classification model. The soil resistance characteristic vector was used as input, and the preset threshold parameters corresponding to the optimal deep loosening pattern were used as output for model training.

[0027] Extract soil resistance feature vectors in real time during the actual operation process, input the soil resistance feature vectors into the trained model, and obtain the prediction results of the best first preset threshold value T1 and the second preset threshold value T2;

[0028] Based on the prediction results, the preset threshold used to control the switching of deep plowing modes is dynamically adjusted.

[0029] Furthermore, the time domain characteristics of the resistance signal include the average resistance value, the maximum resistance value, the minimum resistance value, the standard deviation of the resistance and the resistance change rate; the frequency domain characteristics include the main frequency, the power spectrum density and the fluctuation frequency.

[0030] Furthermore, the sensor assembly includes a sensor base plate, a No. 1 shaft sleeve, a No. 2 shaft sleeve, a No. 1 sensor fixed shaft, a sensor support, a No. 2 sensor fixed shaft, a tension and compression sensor, a sensor top plate, a sensor fixing bolt and a sensor support fixing bolt; two through holes are provided on the sensor base plate for installing the No. 1 shaft sleeve and the No. 2 shaft sleeve; two tension pins respectively pass through the two No. 1 shaft sleeves and the No. 2 shaft sleeves and extend to the sensor top plate, and the sensor assembly, the center bracket and the middle bracket are fixed on the same axis by the top bolts; an open groove is provided on the sensor support for installing the No. 1 sensor fixed shaft and the No. 2 sensor fixed shaft, which pass through the tension and compression sensors respectively and are positioned on the sensor support. When working, the tension and compression sensors are deformed by pressure, causing the No. 1 sensor fixed shaft and the No. 2 sensor fixed shaft to slide in the open groove along the deformation direction of the sensor; the sensor fixing bolt contacts the top of the tension and compression sensor, and the sensor support fixing bolt fixes the sensor support to the sensor base plate.

[0031] Furthermore, the middle bracket includes a connecting plate, a No. 1 fixed side plate, a No. 2 fixed side plate, a hollow limiting beam, a push rod base, a spring base and a bracket bottom plate; the connecting plate is arranged between the No. 1 fixed side plate and the No. 2 fixed side plate; the push rod base and the spring base are respectively fixed on the bracket bottom plate, and when the electric push rod is in an extended state, the push rod base is used to fix the extended push rod and at the same time limit the radial displacement range of the spring system; the hollow limiting beam is clamped on the fixed side plate, and the tension pin shaft passes through the through holes on the bracket bottom plate and the hollow limiting beam and is fixed on the hollow limiting beam.

[0032] Furthermore, the rear bracket includes a bottom plate, a No. 1 side plate, and a No. 2 side plate; the bottom plate is fixed behind the No. 1 side plate and the No. 2 side plate, and the deep plowing shovel is fixed between the No. 1 side plate and the No. 2 side plate through a vibration-damping gasket; one end of the electric push rod is fixedly connected to the bottom plate, and the other end thereof will approach and eventually abut the middle bracket when it is controlled to extend, thereby limiting the range of motion of the rear bracket.

[0033] Compared with the prior art, the multi-modal adaptive deep loosening mechanism based on soil resistance perception of the present invention has the following beneficial effects:

[0034] 1. Intelligent Multi-Mode Switching Technology: This invention uses a sensor assembly to sense the changing characteristics of soil resistance in real time. The control system, based on a dual-threshold judgment mechanism, automatically switches between three modes: traditional rigid-connection subsoiling, self-excited vibration subsoiling, and forced vibration subsoiling. This overcomes the technical bottleneck of existing single-mode subsoiling mechanisms, which cannot adapt to complex soil conditions, and achieves precise matching of operating modes with soil characteristics. This mode switching mechanism achieves a breakthrough in dynamically matching subsoiling operation modes with actual soil resistance characteristics, representing a significant technological innovation in subsoiling machinery, moving from "single-mode fixed operation" to "multi-mode intelligent adaptive operation."

[0035] 2. Energy Optimization Allocation Mechanism: To address the prominent problem of high energy consumption in traditional deep tillage mechanisms, this invention adopts an energy optimization allocation strategy based on soil resistance classification. In low-resistance areas, a traditional rigid connection mode with zero additional energy consumption is used. In medium-resistance areas, a self-excited vibration mode using the self-excited energy of the soil-spring system is utilized. Only in high-resistance areas is a motor-driven forced vibration mode activated. This invention cleverly achieves optimal energy allocation while ensuring operational quality, significantly reducing agricultural production costs.

[0036] 3. High-Precision Vibration Parameter Control System: This system achieves precise control of vibration parameters in the forced vibration subsoiling mode. The vibration frequency can be adjusted within a range of 1-50Hz with a precision of 0.1Hz, and the vibration amplitude can be adjusted within a range of 1-50mm with a precision of 0.5mm. Phase parameter adjustment is also supported to achieve a variety of vibration waveform outputs. By scanning vibration parameters to identify the soil's natural frequency and utilizing the principle of resonance, soil-breaking efficiency is further improved, far exceeding the effectiveness of existing fixed-frequency or roughly adjustable subsoil mechanisms.

[0037] 4. Machine Learning-Driven Intelligent Decision-Making System: This invention innovatively applies machine learning technology to deep tillage mechanism control, establishing a soil property classification model that combines a support vector machine and a decision tree. This model uses soil resistance time-domain characteristics (average, maximum, minimum, standard deviation, rate of change) and frequency-domain characteristics (dominant frequency, power spectral density, and fluctuation frequency) as input to automatically optimize the operating mode switching threshold. The system continuously optimizes model parameters through incremental learning, enabling the mechanism to continuously improve its adaptability to different soil environments over long-term use, thus realizing the intelligent evolution of agricultural machinery.

[0038] 5. Broad Spectrum Adaptability: Thanks to the synergy between multimodal adaptive switching technology and a machine learning system, this invention overcomes the adaptability limitations of existing single-mode deep loosening mechanisms, now covering the full spectrum of soil types, from loose sandy soils to firm clay soils. Furthermore, it enables real-time dynamic adaptation to soil heterogeneity within the same plot (such as differences in soil hardness, moisture content, and obstacle distribution), significantly improving the stability and reliability of agricultural machinery in complex and changing soil environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings used in the description of the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the contents of the embodiments of the present invention and these drawings without any creative work.

[0040] Figure 1 Schematic diagram of the overall structure of a multi-modal adaptive deep loosening mechanism based on soil resistance perception according to Example 1 of the present invention;

[0041] Figure 2 is a schematic structural diagram of the deep loosening monomer described in Example 1;

[0042] Figure 3 is a schematic structural diagram of the rear support described in Example 1;

[0043] Figure 4 is a schematic structural diagram of the middle bracket described in Example 1;

[0044] Figure 5 is a schematic structural diagram of the upper bracket described in Example 1;

[0045] Figure 6 is a schematic structural diagram of the sensor assembly described in Example 1;

[0046] Figure 7 is a schematic structural diagram of the central support described in Example 1;

[0047] Figure 8 This is a schematic diagram of the vibration posture of the deep loosening unit described in Example 1;

[0048] Figure 9 Schematic diagram of the working process of the subsoiling mechanism in Example 1;

[0049] Figure 10 is a hardware schematic diagram of the control system described in Example 1;

[0050] Figure 11 Schematic diagram of the structure of a multi-modal adaptive deep loosening mechanism based on soil resistance perception according to embodiment 2 of the present invention;

[0051] In the picture:

[0052] 1. Deep loosening unit; 2. Control system; 3. Overall frame; 4. Energy device;

[0053] 101. Rear bracket; 102. Electric push rod; 103. Compression spring; 104. Rotating shaft; 105. Middle bracket; 106. Upper bracket; 107. Upper crossbeam; 108. Vibration motor mount; 109. Vibration motor; 110. Tension pin; 111. Sensor assembly; 112. Center bracket; 113 Soiling shovel; 114. Vibration damping washer; 115. Tension spring; 116. Spring limit pin

[0054] 101-1. Bottom plate; 101-2. Side plate No. 1; 101-3. Side plate No. 2; 101-4. Limit plate No. 1; 101-5. Limit plate No. 2;

[0055] 105-1. Connecting plate; 105-2. Fixed side plate No. 1; 105-3. Fixed side plate No. 2; 105-4. Hollow limit beam; 105-5. Push rod base; 105-6. Spring base; 105-7. Bracket base;

[0056] 106-1. Upper support connecting rod; 106-2. Connecting rod base; 106-3. Side support arm No. 1; 106-4. Side support arm No. 2;

[0057] 111-1. Sensor base plate; 111-2. Shaft sleeve No. 1; 111-3. Shaft sleeve No. 2; 111-4. Sensor mounting shaft No. 1; 111-5. Sensor support; 111-6. Sensor mounting shaft No. 2; 111-7. Tension and compression sensor; 111-8. Sensor top plate; 111-9. Sensor mounting bolts; 111-10. Sensor support mounting bolts;

[0058] 112-1. Pin shaft sleeve No. 1; 112-2. Pin shaft sleeve No. 2; 112-3. Clamp No. 1; 112-4. Support arm base No. 1; 112-5. Support arm base No. 2; 112-6. Center support beam; 112-7. Clamp No. 2; 112-8. Support arm No. 1; 112-9. Support arm No. 2; 112-10. Square plate No. 1; 112-11. Square plate No. 2. DETAILED DESCRIPTION

[0059] The present invention will be further described in detail below with reference to the accompanying drawings and examples. It will be understood that the specific embodiments described herein are intended only to illustrate the present invention and are not intended to limit the present invention. It should also be noted that, for ease of description, the accompanying drawings only illustrate portions relevant to the present invention, not all structures.

[0060] In the description of this embodiment, terms such as "upper," "lower," "front," and "rear" are used to refer to positions or locations based on those shown in the accompanying drawings. These terms are intended solely to facilitate description and simplify operation, and are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limitations on the present invention. Furthermore, the terms "first" and "second" are used solely for descriptive purposes and have no special meaning.

[0061] Example 1

[0062] like Figure 1As shown, this embodiment is a multi-modal adaptive deep loosening mechanism based on soil resistance perception, which is mainly composed of a deep loosening unit 1, a control system 2, an overall frame 3, and an energy device 4. The integral frame 3 constitutes the basic support structure of the deep plowing mechanism of the present invention, which is suitable for installation or connection to the base of traction equipment such as tractors, and is used to carry and fix other components of the deep plowing mechanism; the deep plowing unit 1 is the core functional unit for performing actual deep plowing operations and realizing adaptive vibration adjustment. The deep plowing unit 1 can be detachably or fixedly installed on the integral frame 3 through connectors such as U-bolts; the control system 2 realizes signal and energy transmission connection with the deep plowing unit 1 and the energy device 4. The control system 2 is configured to receive sensor signals from the deep plowing unit 1 and process the signals based on preset logic or algorithms, and then generate control instructions to adjust the working state of the deep plowing unit 1. The control system 2 can be integrated on the integral frame 3 or installed on traction equipment; the energy device 4 is configured to provide the required power or energy for the control system 2 and the energy-consuming components (such as vibration motors, electric push rods, etc.) in the deep plowing unit 1; the energy device 4 can provide, for example but not limited to, a voltage in the range of 110V to 380V and a current in the range of 5A to 16A to meet the system operation requirements. To sum up, through the coordinated work of the overall frame 3, deep plowing unit 1, control system 2 and energy device 4, the present invention realizes a deep plowing mechanism that can sense soil resistance and adaptively adjust the working mode and vibration parameters accordingly, aiming to improve working efficiency and optimize energy consumption.

[0063] Figure 2 This is the internal structure of the deep plowing unit 1, which is mainly composed of the following components: a rear bracket 101, an electric push rod 102, a rotating shaft 104, a middle bracket 105, an upper bracket 106, an upper crossbeam 107, a tension pin 110, a sensor assembly 111, a center bracket 112, a deep plowing shovel 113, a vibration-damping gasket 114, a spring system and a vibration motor assembly; the spring system includes a compression spring 103, a tension spring 115, and a spring limit pin 116; the vibration motor assembly includes a vibration motor base 108 and a vibration motor 109.

[0064] Specifically, the deep plowing shovel 113 is fixed to the rear bracket 101 through a vibration-damping gasket 114 and moves synchronously with it. The rear bracket 101 and the middle bracket 105 are hinged by the rotating shaft 104, which enables the rear bracket 101 to rotate around the rotating shaft 104 under the thrust of the soil. At the same time, the rear bracket 101 is also connected to the upper bracket 106 through a pin shaft. When the rear bracket 101 rotates, it will push the upper bracket 106 accordingly. Therefore, the rear bracket 101, the upper bracket 106 and the middle bracket 105 together constitute a parallel four-bar linkage mechanism, which ensures the synchronous and coordinated movement of related components. The center bracket 112 is connected to the front end of the middle bracket 105 through the tension pin shaft 110. The center bracket 12 is rotatably connected to the upper bracket 106, and the sensor assembly 111 is installed on the tension pin shaft 110. After the soil resistance acts on the deep plowing shovel 113, it is transmitted to the middle bracket 105 through the rotation of the rear bracket 101, and then acts on the sensor assembly 111 through the series tension pin 110, thereby realizing real-time sensing of the operation resistance.

[0065] Specifically, the rear portion of the rear bracket 101 provides mounting and support for the electric push rod 102 (preferably with a thrust greater than 2000N and a travel of at least 50cm) and the spring stop pin 116. The compression spring 103 is mounted on the extension rod of the electric push rod 102, with its axial ends respectively abutting the rear bracket 101 and the middle bracket 105. Similarly, the tension spring 115 is mounted on the spring stop pin 116, with its axial ends also respectively abutting the rear bracket 101 and the middle bracket 105. The electric push rod 102 extends and retracts according to the instructions of the control system 2, and its state directly determines the working mode of the deep plowing unit: when the electric push rod 102 is fully extended to support the middle bracket 105, a stroke lock is formed. At this time, the compression spring 103 and the tension spring 115 are both ineffective, and the deep plowing shovel 113 cannot rotate around the rotating shaft 104. The mechanism exhibits a traditional rigid connection deep plowing mode; when the electric push rod 102 retracts, the restriction on the rotation of the rear bracket 101 is released, the compression spring 103 and the tension spring 115 start working, and the deep plowing shovel 113 can rotate around the rotating shaft 104. The spring system interacts with the soil resistance. At this time, the mechanism is in or can enter the self-excited vibration deep plowing mode.

[0066] In addition, the upper bracket 106 is connected to the upper crossbeam 107 and the vibration motor base 108 by bolts. The vibration motor 109 (preferably providing adjustable vibration with a frequency of 1-50 Hz and an amplitude of 1-50 mm) is installed on the vibration motor base 108. When the soil resistance is too large in the self-excited mode, causing the spring system to reach its limit stroke, that is, when the rear bracket 101 rotates to the maximum angle and cannot maintain effective self-excited vibration, the control system 2 can start the vibration motor 109, and the forced vibration generated by it will be transmitted to the deep plowing shovel 113 through the vibration motor base 108, the upper crossbeam 107, the upper bracket 106, and the rear bracket 101 in sequence, so that the mechanism switches to the forced vibration deep plowing mode to cope with high resistance conditions.

[0067] like Figure 3 The structure of the rear bracket 101, specifically, the rear bracket 101 includes a bottom plate 101-1, a No. 1 side plate 101-2, a No. 2 side plate 101-3, a No. 1 limit plate 101-4, and a No. 2 limit plate 101-5. The bottom plate 101-1 is welded behind the No. 1 side plate 101-2 and the No. 2 side plate 101-3, and the deep plowing shovel 113 is fixed between the No. 1 side plate 101-2 and the No. 2 side plate 101-3 through a vibration-damping gasket 114. One end of the electric push rod 102 is fixedly connected to the bottom plate 101-1, and the other end thereof is not continuously connected to the middle bracket 105. Instead, when the electric push rod 102 is controlled to extend, its end will approach and eventually contact the corresponding structure on the middle bracket 105 (i.e. Figure 4 The push rod base 105-5 shown in the figure is used to limit the range of motion of the rear bracket 101. The push rod portion of the electric push rod 102 can extend from the through hole of the first limit plate 101-4. When the push rod is extended, it plays a role in limiting the radial displacement of the compression spring 103. Similarly, the tail of the limit pin 116 is fixed to the second limit plate 101-5 to prevent the tension spring 115 from radial displacement. In addition, combined with Figure 2 As shown in the connection relationship, the rear bracket 101 is hinged to the middle bracket 105 via the rotating shaft 104, allowing relative rotation between the two.

[0068] Figure 4This is the structure of the middle bracket 105, which consists of precision components such as the connecting plate 105-1, the first fixed side plate 105-2, the second fixed side plate 105-3, the hollow limit beam 105-4, the push rod base 105-5, the spring base 105-6, and the bracket base 105-7. As one of the core support and force transmission structures of the deep loosening unit 1, the middle bracket 105 works in conjunction with the linkage mechanism including the rear bracket 101 and the upper bracket 106. It plays a key role in the multimodal adaptive deep loosening process, ensuring the stability, functional implementation, and movement coordination of the entire mechanism under complex soil conditions. The connecting plate 105-1 is positioned between the first and second fixed side plates 105-2, 105-3. By controlling the structural spacing between the rigid plates, it improves the mechanical strength and torsional rigidity of the overall structure. The bracket base plate 105-7 features four functional through-holes: the upper and lower holes are used to secure the push rod base 105-5 and the spring base 105-6, respectively. The left and right holes are coaxially arranged with the two through-holes in the hollow limit beam 105-4, providing precise positioning for the installation of the tension pin 110. When the electric push rod 102 is in the extended state, the push rod base 105-5 not only performs the critical function of securing the extended push rod, but also limits the radial displacement range of the compression spring 103, preventing lateral instability during operation. The spring base 105-6 specifically limits the radial range of motion of the tension spring 115, ensuring a stable and controllable motion trajectory of the elastic element during the operating cycle. The hollow limiting beam 105-4 is clamped in the installation groove at the front end of the two fixed side plates. The two through holes of the hollow limiting beam 105-4 are concentrically arranged with the left and right through holes of the bracket base plate 105-7. The tension pin 110 passes through these two groups of through holes at the same time and is fixed on the hollow limiting beam 105-4, forming a reliable mechanical connection, thereby realizing effective force transmission and overall stability of the structure during deep plowing operations.

[0069] Figure 5 The structure of the upper bracket 106 is composed of an upper bracket connecting rod 106-1, a connecting rod base 106-2, a No. 1 side arm 106-3, and a No. 2 side arm 106-4, forming a motion transmission and mechanical support unit in the deep loosening unit 1. The upper bracket connecting rod 106-1 is designed with a through-hole structure on both sides, and a reliable hinged connection is formed with the upper through hole of the rear bracket 101 and the connecting rod base 106-2 through a high-strength pin shaft. This hinged structure ensures that when the rear bracket 101 produces angular displacement, the upper bracket connecting rod 106-1 can efficiently and accurately transmit the motion state to the connecting rod base 106-2, thereby realizing synchronous motion and force balance among the various components of the system. Combined with Figure 2Further analysis reveals that upper crossbeam 107 is fixed to side arm 106-3 and side arm 106-4. Connecting rod base 106-2, side arm 106-3, and side arm 106-4 are all located on the same side of upper crossbeam 107, forming a rigidly connected whole. Vibration motor mount 108, located on the opposite side of upper crossbeam 107, is bolted together to form a rigid structural unit. Therefore, upper bracket 106, upper crossbeam 107, vibration motor mount 108, and vibration motor 109 form a highly coordinated motion system, enabling precise displacement control and vibration regulation during multimodal adaptive deep loosening.

[0070] Figure 6 The figure is a schematic diagram of the structure of the sensor assembly 111. The assembly consists of a sensor base plate 111-1, a No. 1 sleeve 111-2, a No. 2 sleeve 111-3, a No. 1 sensor fixed shaft 111-4, a sensor support 111-5, a No. 2 sensor fixed shaft 111-6, a tension and compression sensor 111-7, a sensor top plate 111-8, a sensor fixing bolt 111-9, and a sensor support fixing bolt 111-10, forming a complete mechanical transmission path. In terms of structural layout, the sensor base plate 111-1 serves as the basic support platform and is provided with two through-holes for mounting the No. 1 sleeve 111-2 and the No. 2 sleeve 111-3. The tension pin 110 passes through these two sleeves and extends to the sensor top plate 111-8. The top bolts fix the three independent structural components of the sensor assembly 111, the center bracket 112, and the middle bracket 105 on the same axis. The sensor support 111-5 is located in the middle position and is designed with a special open groove for installing the No. 1 sensor fixing shaft 111-4 and the No. 2 sensor fixing shaft 111-6. The two fixing shafts first pass through the fixing holes on both sides of the tension and compression sensor 111-7, and then are positioned on the sensor support 111-5. During operation, the tension and compression sensor 111-7 is deformed by pressure, causing the No. 1 sensor fixing shaft 111-4 and the No. 2 sensor fixing shaft 111-6 to slide in the open groove along the sensor deformation direction. The sensor fixing bolt 111-9 contacts the top of the tension and compression sensor 111-7, and the sensor support fixing bolt 111-10 ensures that the sensor support 111-5 and the sensor base plate 111-1 remain fixed and stable.

[0071] The core function of the sensor assembly 111 is to achieve accurate conversion and measurement of mechanical signals. When the deep plowing shovel 113 is working in the soil, the soil resistance causes the rear bracket 101 to produce rotational motion. This motion converts the horizontal force into a direct force on the tension sensor 111-7 through a mechanical connection, thereby realizing real-time monitoring of the soil resistance. In the signal transmission path, the tension pin 110 serves as the main force transmission element, transmitting the tension from the rear bracket 101 through the sensor top plate 111-8 to the sensor fixing bolt 111-9, and then acting on the tension sensor 111-7. To ensure measurement accuracy, special attention should be paid during installation: during the tightening process of the tension pin 110, the sensor top plate 111-8 should not be directly fixed to the sensor support 111-5, but a measurement gap of 5 to 20 mm must be accurately reserved. This gap ensures effective transmission and accurate measurement of force. At the same time, in order to prevent the sensor support 111-5 from sinking and moving due to its own weight, which would affect the measurement accuracy, the design uses sensor support fixing bolts 111-10 to fix it on the sensor base plate 111-1 to ensure the structural stability of the entire sensing system in a dynamic working environment.

[0072] Figure 7The structure of the central support 112, a key force-transmitting component in the deep loosening mechanism, is composed of a No. 1 pin shaft sleeve 112-1, a No. 2 pin shaft sleeve 112-2, a No. 1 clamping plate 112-3, a No. 1 support arm base 112-4, a No. 2 support arm base 112-5, a central support beam 112-6, a No. 2 clamping plate 112-7, a No. 1 support connecting arm 112-8, a No. 2 support connecting arm 112-9, a No. 1 square plate 112-10, and a No. 2 square plate 112-11. The central support beam 112-6 serves as the core supporting structure, with a No. 1 clamping plate 112-3 and a No. 2 clamping plate 112-7 on either side, forming a "sandwich structure" of clamping plate, crossbeam, and clamping plate. During installation, the No. 1 pin shaft sleeve 112-1 and the No. 2 pin shaft sleeve 112-2 pass through the through hole of the No. 1 splint 112-3, the through hole of the center support beam 112-6 and the through hole of the No. 2 splint 112-7 to form a stable coaxial connection. This two-side splint design provides effective lateral protection for the beam. The No. 1 pin shaft sleeve 112-1 and the No. 2 pin shaft sleeve 112-2 are respectively connected to a tension pin shaft 110. The upper surface of the center support crossbeam 112-6 is mounted with a No. 1 support arm base 112-4 and a No. 2 support arm base 112-5, and the lower surface is mounted with a No. 1 square plate 112-10 and a No. 2 square plate 112-11. The support arm bases and the square plates are connected by bolts. The No. 1 support arm base 112-4 and the No. 2 support arm base 112-5 are respectively connected to the No. 1 side support arm 106-3 and the No. 2 side support arm 106-4 of the upper support 106 via pins, thereby achieving a rotational connection between the center support 12 and the upper support 106. At the same time, the No. 1 support connecting arm 112-8 and the No. 2 support connecting arm 112-9 are fixed to the front side of the center support crossbeam 112-6 by precision welding or high-strength bolts, providing the overall structure with a connection interface with the overall frame 3. That is, the front end of the center support 112 is connected to the overall frame 3 via the No. 1 support connecting arm 112-8 and the No. 2 support connecting arm 112-9.

[0073] Figure 8 For deep loosening single vibration posture, combined with Figure 2 and Figure 5 Further analysis reveals that the entire mechanism forms a parallelogram, with the rear support 101, upper support 106, and middle support 105 serving as the three key links, connected by a pivot axis to form a closed-loop kinematic chain. Within this structure, the rear support 101 maintains a constant parallel relationship with the first and second side arms 106-3, 106-4, while the upper support link 106-1 and middle support 105 also maintain a strictly parallel relationship. This geometric constraint ensures the coordinated and stable movement of the entire mechanism. The pivot axis 104, serving as the center of rotation for the rear support 101, provides the reference rotation axis for the entire mechanism. Figure 8 The left picture shows the initial state of deep loosening. Figure 8The right image shows the motion during vibration. Under soil resistance, the rear support 101 can rotate clockwise relative to its initial position, with a maximum rotation angle of no more than 60°. The middle support 105 is rigidly connected to the center support 112 via a tension pin 110, maintaining a relatively fixed position during vibration. Furthermore, the rear support 101 and the subsoiler 113 form an integrated structure, allowing soil resistance to be directly transmitted to the motion system. The operating principle is as follows: When the subsoiler 113 encounters soil resistance during operation, this external load drives the rear support 101, to which it is fixed, to angular displacement about the rotation axis 104. Based on the kinematic principles of a parallelogram linkage, the rotation of the rear support 101 is mechanically transmitted, causing the upper support 106 to move synchronously. Thanks to the geometric constraints in the design that ensure the rear support 101 remains parallel to the first and second side arms 106-3, and the upper support link 106-1 remains parallel to the middle support 105, the entire system achieves highly coordinated motion.

[0074] Figure 9 The multi-modal adaptive deep plowing mechanism of this embodiment continuously senses changes in soil resistance through the tension and compression sensors 111-7 based on the control system 2, automatically switches working modes and optimizes working parameters according to the working environment, realizing intelligent switching between three working modes: traditional rigid connection deep plowing, self-excited vibration deep plowing, and forced vibration deep plowing. The specific working process is as follows:

[0075] (1) When the soil resistance is small, below the resistance threshold T1, the system operates in the traditional rigid connection deep loosening mode. The electric push rod 102 is in the maximum stroke state, and the spring system does not participate in the work. At this time, the deep loosening shovel forms a rigid connection with the frame, completing the loosening operation in the simplest and most efficient way. This mode has the lowest energy consumption and is suitable for loose soil areas.

[0076] (2) When the soil resistance reaches the threshold value T1 but is lower than the threshold value T2, the system switches to the self-excited vibration deep loosening mode. The control system 2 instructs the electric push rod 102 to retract appropriately, so that the tension spring and the compression spring 103 start to work. During the movement, the soil resistance and the spring system form a self-excited vibration system, and the deep loosening shovel generates periodic vibrations, effectively breaking up medium-hard soil blocks while reducing traction resistance.

[0077] (3) When soil resistance continues to rise and exceeds threshold value T2, the system switches to forced vibration deep loosening mode. Control system 2 activates vibration motor 109, generating controlled forced vibration to maximize the breaking up of the hard soil. At this point, the spring system can no longer provide sufficient vibration energy, and external energy input is required to ensure work quality.

[0078] The working process of this embodiment is as follows:

[0079] When the system starts, it defaults to the traditional rigid connection deep loosening mode, and the electric push rod 102 is in the maximum stroke state; the tension and compression sensor 111-7 continuously monitors the soil resistance and transmits it to the control system 2; the control system 2 analyzes the resistance characteristics based on the intelligent resistance judgment strategy and switches to the appropriate working mode according to the resistance size; in the forced vibration deep loosening mode, the system dynamically adjusts the working parameters of the vibration motor according to the real-time resistance changes; the system continuously monitors the soil resistance changes, and when the resistance decreases and stabilizes below a specific threshold, it falls back to the lower energy consumption working mode as needed. Through the above-mentioned intelligent adaptive mechanism, the present invention can intelligently select the most suitable deep loosening mode and parameter combination for different soil conditions, realize efficient and energy-saving deep loosening operations, and significantly improve the work quality and equipment service life.

[0080] Furthermore, in this embodiment, the setting of the first preset threshold value T1 and the second preset threshold value T2 is closely related to the soil type. After a large number of experimental verifications, for general agricultural soils, the first preset threshold value T1 is preferably set in the range of 400N to 900N, and the second preset threshold value T2 is preferably set in the range of 1000N to 2000N. Specifically, in sandy soil, T1 is preferably set to 450N to 550N, and T2 is preferably set to 1100N to 1300N; in loam, T1 is preferably set to 550N to 650N, and T2 is preferably set to 1300N to 1500N; under clay conditions, T1 is preferably set to 750N to 850N, and T2 is preferably set to 1700N to 1900N. These precisely set threshold ranges ensure that the system can accurately identify different soil conditions and switch the operating mode at the most appropriate time to achieve optimal energy utilization.

[0081] Furthermore, to improve the system's adaptability to complex and changeable soil environments, this embodiment adopts an intelligent resistance judgment mechanism rather than a single threshold trigger. The system not only monitors the absolute value of the resistance, but also analyzes the time characteristics of the resistance change. It uses a sliding time window technology to smooth the continuously collected resistance data and judge its persistence characteristics based on the resistance change rate. For example, when a deep plowing shovel encounters obstacles such as gravel and tree roots, which cause instantaneous resistance fluctuations, the system maintains the current working mode to avoid mechanical shock and energy waste caused by frequent switching. Mode switching is triggered only when the resistance continues to exceed the threshold within a specific time window. The sliding time window is set to 1 to 3 seconds.

[0082] Furthermore, in the forced vibration deep loosening mode, control system 2 precisely adjusts vibration parameters based on real-time monitoring of soil resistance characteristics. The system can adjust the vibration frequency within a range of 1-50 Hz with an accuracy of 0.1 Hz, and the vibration amplitude within a range of 1-50 mm with an accuracy of 0.5 mm. By varying the phase parameters, different vibration waveform outputs are achieved to adapt to different soil conditions. Control system 2 identifies the soil's natural frequency through a vibration parameter sweep test and adjusts the operating frequency of the vibration motor 109 to a value close to the soil's natural frequency, utilizing the principle of resonance to maximize the vibration effect. The specific steps include: performing a frequency sweep test within the 1-50 Hz range at the initial start-up of the mode; monitoring the resistance sensor response signal and calculating the resistance attenuation rate at each frequency; determining the frequency point with the maximum attenuation rate as the optimal resonant frequency; precisely adjusting the vibration motor's operating frequency to within ±0.1 Hz of this frequency point; and fine-tuning in real time based on resistance fluctuations to maintain the optimal vibration effect. Simultaneously, the system precisely controls the vibration amplitude and waveform type by adjusting the vibration motor's input voltage waveform and phase parameters. This multi-parameter collaborative optimization enables the system to make fine-grained parameter adjustments based on the specific characteristics of the soil, significantly improving operating efficiency and reducing energy consumption.

[0083] Furthermore, the control system 2 of the present invention employs an adaptive threshold adjustment mechanism based on machine learning. This mechanism establishes a machine learning model to analyze the soil resistance variation pattern collected during operation in real time and continuously optimize the threshold parameters. The implementation process of the adaptive threshold adjustment mechanism based on machine learning includes the following steps:

[0084] First, the control system 2 establishes an initial threshold database containing standard T1 and T2 value ranges corresponding to different soil resistance patterns. During the system development phase, researchers collected resistance data in a test field with known soil type, moisture content, and compaction to establish a database of the corresponding relationship between resistance characteristics and soil conditions.

[0085] Next, the system extracts time-domain features (average, maximum, minimum, standard deviation, and rate of change) and frequency-domain features (dominant frequency, power spectrum density, and fluctuation frequency) from the resistance signal collected by tension and compression sensor 111-7 to form a feature vector. Research has shown that different soil types and states produce distinct resistance characteristic patterns. For example, sandy soils typically exhibit high-frequency, low-amplitude resistance fluctuations, while clay soils exhibit low-frequency, high-amplitude resistance variations.

[0086] Then, control system 2 uses an improved support vector machine algorithm combined with a decision tree model, takes the above soil resistance characteristic vector as input, and uses the optimal deep loosening pattern as output to train and obtain the mapping relationship between resistance characteristics and optimal thresholds, without the need to directly measure soil physical parameters in real time;

[0087] During the actual operation, the system uses a sliding time window (length of 1 to 3 seconds) to extract the resistance feature vector in real time, inputs the feature vector into the trained model, and indirectly infers the current soil condition through the resistance feature and predicts the optimal T1 and T2 thresholds;

[0088] Finally, based on the prediction results, the system threshold parameters are dynamically adjusted to achieve precise mode switching control. Simultaneously, the system continuously collects new resistance characteristic data and corresponding operational performance data (such as observed soil crushing rate, actual energy consumption, and optimal operating speed), storing this newly collected data in a database for online incremental learning of the model.

[0089] The preset threshold values ​​adjusted by machine learning significantly improved accuracy, narrowing the originally wide threshold range to precise values ​​suitable for specific soil conditions. For example, for a resistance mode characterized by sandy soil, the initial T1 threshold range was 450N-550N, which was refined to 480N-500N after machine learning optimization; the initial T2 threshold range was 1100N-1300N, which was refined to 1160N-1200N. Through this indirect inference method based on resistance characteristics, Control System 2 gradually formed and continuously optimized a mapping database between resistance characteristics and optimal threshold parameters, enabling the mechanism to accurately select the optimal mode switching time based solely on data collected by tension and compression sensors.

[0090] Figure 10 This is a block diagram of the hardware composition principle of control system 2. Control system 2 adopts a layered modular design and is mainly composed of five functional modules: central processing unit 201, signal acquisition module 202, execution drive module 203, human-computer interaction module 204 and data storage module 205, forming a complete information perception-analysis processing-execution control closed loop.

[0091] The central processing unit 201 serves as the calculation and decision-making core of the system, and is responsible for real-time processing of soil resistance data and executing the modal switching algorithm. It is preferred to use a high-performance processor based on the ARM Cortex-A series architecture, especially a SoC chip with an integrated neural network acceleration unit, which can realize rapid intelligent analysis and judgment of complex soil conditions, support real-time reasoning of deep learning models, and provide accurate decision-making basis for multi-modal switching. The unit also integrates a dedicated machine learning operation unit, including a support vector machine algorithm accelerator and a decision tree model parsing engine, which can realize the feature extraction, model training and prediction functions described in the claims, and support real-time threshold parameter optimization and adjustment. In addition, the unit is equipped with a dedicated soil characteristic recognition algorithm library, which can accurately identify and classify different soil types by analyzing soil resistance patterns.

[0092] Signal acquisition module 202 focuses on high-precision acquisition of soil resistance and mechanism displacement signals. It is preferably constructed using a multi-channel digital signal processing chip, equipped with a 16 / 24-bit high-precision ADC conversion unit, and integrated self-calibration and interference suppression circuitry to ensure reliable signal acquisition in harsh agricultural environments. This module utilizes multi-stage filtering and signal conditioning technology to effectively eliminate high-frequency noise interference caused by mechanical vibration. Specifically, this module incorporates a spectrum analysis processing unit, employing FFT hardware acceleration technology to calculate the frequency domain characteristics of the soil resistance signal in real time, including the dominant frequency, power spectral density, and fluctuation frequency. This provides critical data support for identifying soil natural frequencies and optimizing vibration parameters.

[0093] The execution drive module 203 is responsible for controlling the vibration mode and operating depth adjustment, preferably using a closed-loop drive system based on vector control to achieve precise control of the vibration motor and electric push rod. More preferably, the module integrates a vibration parameter self-optimization algorithm, which can automatically adjust the vibration frequency, amplitude and phase parameters according to the soil resistance characteristics, so that the vibration energy is concentrated in the most effective frequency band, and adjust the resonance parameters in real time according to the operating speed to achieve optimal vibration performance and energy utilization efficiency. The module adopts a programmable logic controller architecture, supports multiple vibration waveform generation and precise phase control, and can achieve precise adjustment of the vibration motor 1-50Hz vibration frequency and 1-50mm amplitude.

[0094] The human-computer interaction module 204 allows operators to set parameters and monitor system status. It preferably utilizes an industrial-grade touchscreen interface, supporting multi-level permission management and remote control. This module integrates a data visualization engine that displays real-time soil resistance distribution, vibration performance evaluation, and energy consumption analysis data, helping operators intuitively understand equipment operating status and soil conditions.

[0095] The data storage module 205 is specifically used to record operational data and system operating status, preferably using a redundant storage architecture to ensure data security. This module supports a dual-layer structure of high-speed cache and low-power storage, capable of storing multi-dimensional information such as soil resistance maps, operational trajectories, and equipment status, providing data support for subsequent precision agriculture decision-making. In particular, this module is equipped with an incremental learning data management system that supports the dynamic integration of newly collected data and real-time updating of model parameters. It can continuously optimize the mapping database between soil texture and optimal threshold parameters, enabling online adaptive optimization of machine learning models and ensuring that the control system can continuously improve its adaptability to varying soil conditions.

[0096] Control System 2 offers a variety of flexible power configuration options that can be adapted to the specific application scenario. The system can be powered by a standard battery and supports conventional voltage levels such as 12V and 24V. It integrates multiple DC-DC conversion circuits to provide independent operating voltages such as 5V and 3.3V for different functional modules. Another configuration method is to draw power directly from Energy Device 4. Using a high-efficiency voltage-stabilizing step-down circuit, the high-voltage power supply is converted to a voltage level suitable for the control system, ensuring stable operation of the system under various power supply conditions. The power management unit employs an intelligent power consumption scheduling strategy to maximize battery life while maintaining system performance.

[0097] Example 2

[0098] Figure 11 This is a schematic diagram of the structure of a multi-modal adaptive deep loosening mechanism based on soil resistance perception as described in Example 2. This embodiment includes three of the deep loosening units 1, which are integrated into a configuration scheme for a multi-row deep loosening mechanism. Preferably, the number of the deep loosening units 1 can be reasonably selected according to the size design of the overall frame 3, and the whole machine can be flexibly configured into two rows, three rows or more rows. Multiple rows of whole machines share a set of the control system 2 and energy device 4, and there is no need to configure an additional control system and energy device for each row, which simplifies the structure of the whole machine and reduces costs.

[0099] In this embodiment, the control system 2 can process the operating status of multiple deep loosening units 1 at the same time, and receive the soil resistance feedback signal of each deep loosening unit sensor assembly 111 respectively. According to the soil resistance conditions at different positions, the control system 2 can independently control the working status of the electric push rod 102 and the vibration motor 109 of each deep loosening unit, so that each deep loosening unit can automatically switch between the three modes of traditional rigid connection deep loosening, self-excited vibration deep loosening, and forced vibration deep loosening. This design enables the whole machine to achieve precise adaptation according to actual soil conditions at different positions in the horizontal direction (perpendicular to the direction of travel), greatly improving operating efficiency and adaptability. The multi-row whole machine embodiment not only maintains all the technical advantages of the single-row deep loosening mechanism, but also significantly improves the operating efficiency and operating width. It is an important extension of the present invention in actual agricultural machinery applications. It is particularly suitable for deep loosening operations in large areas of farmland and can meet the needs of modern agricultural mechanization, intelligence, efficiency and precision.

[0100] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A multi-modal adaptive deep loosening mechanism based on soil resistance perception, characterized in that: It includes an integral frame, a subsoil unit, a control system and an energy device; the subsoil unit is installed on the integral frame, the control system is connected with the subsoil unit for signal transmission, and the energy device provides power or energy for the control system and the energy-consuming components in the subsoil unit; The deep loosening unit includes a rear bracket, an electric push rod, a middle bracket, an upper bracket, a tension pin, a sensor assembly, a center bracket, a deep loosening shovel, a spring system and a vibration motor assembly; the deep loosening shovel is fixed to the bottom of the rear bracket; the middle bracket is hinged to the rear bracket through a rotating shaft, and the upper bracket is hinged to the rear bracket, and the rear bracket, the upper bracket and the middle bracket together constitute a parallel four-bar linkage mechanism; the electric push rod is fixed to the front end of the rear bracket, and its extended end can abut against the middle bracket after being extended; the spring system abuts between the rear bracket and the middle bracket; the sensor assembly is installed on the tension pin; the center bracket is connected to the front end of the middle bracket through a tension pin, and the top end of the center bracket is rotatably connected to the bottom of the upper bracket; the vibration motor assembly is installed at the front end of the upper bracket; The subsoiling unit senses soil resistance in real time through a sensor assembly and transmits the signal to the control system. The control system decides whether the subsoiling mechanism switches between the following three subsoiling modes based on the size and variation characteristics of the soil resistance and generates corresponding control instructions to adjust the working state of the subsoiling unit: When the soil resistance is less than the first preset threshold value T1, the deep loosening mechanism enters the rigid connection deep loosening mode; When the soil resistance is between the first preset threshold value T1 and the second preset threshold value T2, the subsoiling mechanism enters the self-excited vibration subsoiling mode; When the soil resistance is greater than or equal to the second preset threshold value T2, the deep plowing mechanism enters the forced vibration deep plowing mode.

2. The multi-modal adaptive deep loosening mechanism based on soil resistance perception according to claim 1, characterized in that: In the rigid connection deep loosening mode, the control system controls the electric push rod to fully extend and press against the middle bracket, forming a travel lock, and the spring system does not participate in the work. At this time, the rear bracket cannot rotate around the rotating shaft; In the self-excited vibration deep loosening mode, the control system controls the electric push rod to retract, releasing the restriction on the rotation of the rear bracket, and the spring system starts to work, so that the rear bracket can rotate around the rotating shaft. During the movement, the soil resistance and the spring system form a self-excited vibration system, and the deep loosening shovel generates periodic vibration; In the forced vibration deep tillage mode, the control system starts the vibration motor assembly and transmits the vibration to the deep tillage shovel through the vibration motor assembly, the upper bracket and the rear bracket.

3. The multi-modal adaptive deep loosening mechanism based on soil resistance perception according to claim 2, characterized in that: In the forced vibration deep loosening mode, the control system adjusts the vibration parameters according to the real-time monitoring of the soil resistance characteristics: At the initial start-up of the mode, a frequency sweep test is performed in the range of 1-50Hz. The response signal of the sensor assembly is monitored and the resistance attenuation rate at each frequency is calculated. The frequency with the highest attenuation rate is determined as the optimal resonant frequency. The operating frequency of the vibration motor is adjusted to within ±0.1Hz of the optimal resonant frequency. Fine-tuning is performed in real time based on the resistance change trend. At the same time, the control system controls the vibration amplitude and waveform type by adjusting the input voltage waveform and phase parameters of the vibration motor.

4. The multi-modal adaptive deep loosening mechanism based on soil resistance perception according to claim 1, characterized in that: The control system monitors the magnitude of soil resistance and analyzes the time characteristics of soil resistance changes. It uses a sliding time window to smooth the continuously collected resistance data and determines its persistence characteristics based on the resistance change rate. The switching of deep plowing mode is triggered only when the soil resistance continuously exceeds the corresponding preset threshold within the sliding time window; by setting the sliding time window, when the deep plowing shovel encounters an obstacle, causing a sudden change in instantaneous resistance and resulting in instantaneous resistance fluctuations, the current deep plowing mode is still maintained, avoiding frequent switching of deep plowing modes.

5. The multi-modal adaptive deep loosening mechanism based on soil resistance perception according to claim 4, characterized in that: The sliding time window is set to 1 to 3 seconds.

6. The multi-modal adaptive deep loosening mechanism based on soil resistance perception according to claim 1, characterized in that: The control system uses an adaptive threshold adjustment method based on machine learning to dynamically adjust the preset threshold for deep loosening mode switching, including the following process: Resistance data were collected in experimental fields with known soil types, moisture content, and compaction, and a database of the correspondence between soil resistance characteristics and soil conditions was established; Extracting time domain features and frequency domain from the resistance signal collected by the sensor assembly; An improved support vector machine algorithm combined with a decision tree model was used to construct a classification model. The soil resistance characteristic vector was used as input, and the preset threshold parameters corresponding to the optimal deep loosening pattern were used as output for model training. Extract soil resistance feature vectors in real time during the actual operation process, input the soil resistance feature vectors into the trained model, and obtain the prediction results of the best first preset threshold value T1 and the second preset threshold value T2; Based on the prediction results, the preset threshold used to control the switching of deep plowing modes is dynamically adjusted.

7. The multi-modal adaptive deep loosening mechanism based on soil resistance perception according to claim 6, characterized in that: The time domain characteristics of the resistance signal include the average resistance value, the maximum resistance value, the minimum resistance value, the standard deviation of the resistance and the resistance change rate; Frequency domain characteristics include main frequency, power spectrum density and fluctuation frequency.

8. The multi-modal adaptive deep loosening mechanism based on soil resistance perception according to claim 1, characterized in that: The sensor assembly includes a sensor base plate, a No. 1 shaft sleeve, a No. 2 shaft sleeve, a No. 1 sensor fixed shaft, a sensor support, a No. 2 sensor fixed shaft, a tension and compression sensor, a sensor top plate, a sensor fixing bolt and a sensor support fixing bolt; two through holes are provided on the sensor base plate for installing the No. 1 shaft sleeve and the No. 2 shaft sleeve; two tension pins respectively pass through the two No. 1 shaft sleeves and the No. 2 shaft sleeves and extend to the sensor top plate, and the sensor assembly, the center bracket and the middle bracket are fixed on the same axis by the top bolts; an open groove is provided on the sensor support for installing the No. 1 sensor fixed shaft and the No. 2 sensor fixed shaft, which pass through the tension and compression sensors respectively and are positioned on the sensor support. When working, the tension and compression sensors are deformed by pressure, causing the No. 1 sensor fixed shaft and the No. 2 sensor fixed shaft to slide in the open groove along the deformation direction of the sensor; the sensor fixing bolt contacts the top of the tension and compression sensor, and the sensor support fixing bolt fixes the sensor support to the sensor base plate.

9. The multi-modal adaptive deep loosening mechanism based on soil resistance perception according to claim 1, characterized in that: The middle bracket includes a connecting plate, a No. 1 fixed side plate, a No. 2 fixed side plate, a hollow limiting beam, a push rod base, a spring base and a bracket bottom plate; the connecting plate is arranged between the No. 1 fixed side plate and the No. 2 fixed side plate; the push rod base and the spring base are respectively fixed on the bracket bottom plate, and when the electric push rod is in an extended state, the push rod base is used to fix the extended push rod and at the same time limit the radial displacement range of the spring system; the hollow limiting beam is clamped on the fixed side plate, and the tension pin shaft passes through the through holes on the bracket bottom plate and the hollow limiting beam and is fixed on the hollow limiting beam.

10. The multi-modal adaptive deep loosening mechanism based on soil resistance perception according to claim 1, characterized in that: The rear bracket includes a bottom plate, a No. 1 side plate, and a No. 2 side plate; the bottom plate is fixed behind the No. 1 side plate and the No. 2 side plate, and the deep plowing shovel is fixed between the No. 1 side plate and the No. 2 side plate through a vibration-damping gasket; one end of the electric push rod is fixedly connected to the bottom plate, and the other end thereof will approach and eventually abut the middle bracket when it is controlled to extend, thereby limiting the range of motion of the rear bracket.

Images