A method for tracking control of unmanned harvesting machine based on touch signal correction and adaptive preview
By using tactile signal correction and adaptive pre-aiming control methods, the problem of high-precision trajectory tracking of unmanned harvesters in farmland environments was solved, achieving efficient and stable harvester control and avoiding crop crushing and steering wheel vibration.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- TIANJIN UNIV
- Filing Date
- 2023-11-24
- Publication Date
- 2026-06-26
AI Technical Summary
Existing unmanned harvesting machines struggle to achieve high-precision trajectory tracking in farmland environments. Traditional control algorithms perform poorly in complex road conditions and lack sufficient computing resources, failing to meet the high-precision row tracking requirements of the harvesting machines.
A control method based on tactile signal correction and adaptive preview is adopted. The matching point is found by the bisection method, the preview point is adaptively selected, and the heading error and steering wheel angle are calculated by combining the touch sensor signal and proportional-integral control to achieve precise trajectory control.
It improves the trajectory tracking accuracy of the harvester, reduces the risk of crushing crops, enhances the stability and computing efficiency of the controller, and ensures the stable operation of the harvester under complex road conditions.
Smart Images
Figure CN117519179B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of unmanned harvester control technology, and in particular to a tracking control method for unmanned harvesters based on tactile signal correction and adaptive pre-aiming. Background Technology
[0002] Cotton, corn, soybeans, and other agricultural products are among the most important crops globally, possessing immense economic and food security value and thus receiving widespread attention and cultivation worldwide. However, large-scale crop harvesting is often hampered by high labor costs and several problems. For instance, inexperienced drivers may cause problems such as blocked harvester heads or crop damage due to improper speed and directional control. Therefore, unmanned harvesters capable of accurately tracking the harvesting trajectory hold immense potential to significantly improve the quality and efficiency of crop harvesting.
[0003] To address this, Miao Zhonghua and colleagues at Shanghai University employed a speed-parameter adaptive proportional-integral-derivative (PID) control algorithm to study the automatic alignment assisted driving control technology for cotton harvesters. However, due to the changing state of the vehicle over time and the uncertainty of road conditions, the PID controller struggled to maintain optimal control quality under different conditions and could not be extended to the tracking control of cotton harvester steering. Furthermore, for similarly structured corn harvesters, soybean harvesters, and others, most can only achieve automatic alignment, but cannot automatically turn or change direction, thus failing to achieve fully automated driving throughout the entire process.
[0004] In the traditional field of autonomous driving for passenger vehicles, algorithms such as pre-aiming, model predictive control, and adaptive control have been widely studied. Some scholars have designed trajectory tracking controllers using pure tracking methods. Their results show that this algorithm performs well on paths with low curvature, but on roads with high curvature, the tracking accuracy decreases because the pre-aiming point crosses some points. Model prediction-based trajectory tracking controllers can be designed using vehicle dynamics models, and experimental results demonstrate that this method has better control performance than traditional control algorithms. However, due to the complex road conditions and structure of agricultural machinery, using overly complex prediction models can lead to insufficient computing power in the computing units. On the other hand, simply using control based on a simple model can easily degrade control quality due to deviations between the model and the actual object. Directly using traditional trajectory tracking control algorithms on agricultural machinery cannot meet the high-precision row alignment requirements of harvesters.
[0005] In summary, a tracking control algorithm that adapts to the working environment and has high control accuracy is needed to address the trajectory tracking control problem of unmanned harvesters. Summary of the Invention
[0006] The purpose of this invention is to address the technical deficiencies in the existing technology by providing a tracking control method for unmanned harvesters based on tactile signal correction and adaptive pre-aiming.
[0007] The technical solution adopted to achieve the purpose of this invention is:
[0008] A tracking control method for an unmanned harvester based on tactile signal correction and adaptive pre-aiming includes the following steps:
[0009] Step 1: Based on the target trajectory of the unmanned harvester and its current actual position (X,Y), use a binary search method to find the nearest matching point (X,Y) on the target trajectory that matches the current actual position. r ,Y r ,θ r ,κ r ), where (X) r ,Y r ) represents the coordinates of the matching point, θ r To match the heading angle, κ r To determine the curvature of the matching points, the distance error e between the matching points is then calculated. rd ;
[0010] Step 2: The pre-aiming point adaptive selection module uses a binary search method to find the pre-aiming point ahead of the matching point on the target's driving trajectory, with a pre-aiming distance of L from the matching point. p Calculate the heading error e of the aiming point. pθ ;
[0011] Step 3: The heading correction module calculates the heading correction amount Δe based on the median offset voltage ΔU. θ ;
[0012] Step 4, based on the heading error e of the pre-aiming point obtained in Step 2 pθ and the heading correction amount Δe obtained in step 3 θ Calculate the heading error e θ ;
[0013] Step 5, the model-based tracking controller calculates the matching point distance error e obtained in Step 1. rd and the heading error e obtained in step 4 θ Calculate the target steering wheel angle δ wr .
[0014] In the above technical solution, in step 1, the target driving trajectory is given by the upper-level planning of the unmanned harvester, and the current actual position of the vehicle is obtained by the combined inertial navigation on the unmanned harvester.
[0015] In the above technical solution, in step 1, the matching point distance error
[0016] In the above technical solution, in step 2, the pre-aiming distance L p =k l v+L0, where v is the actual vehicle speed obtained from the sensor, and k l L0 represents the aiming distance speed coefficient and the preset aiming distance.
[0017] In the above technical solution, in step 2, e pθ =θ pr -θ, θ pr θ represents the heading at the target point, and θ represents the actual heading of the current vehicle.
[0018] In the above technical solution, in step 3, ΔU=UU mid U mid This indicates the zero-position voltage output by the touch sensor when the vehicle's trajectory does not deviate from the crop ridge line, where U is the voltage output by the touch sensor in real time.
[0019] In the above technical solution, in step 3, Δe θ =k p ΔU+k i ∫ΔUdt, where k p For the proportional gain coefficient and k i This is the integral gain coefficient.
[0020] In the above technical solution, k p and k i The results were obtained through parameter adjustments in simulation and real experimental tests, respectively.
[0021] In the above technical solution, in step 4, e θ =e pθ +[1-sgn(|κ r |)]Δe θ .
[0022] In the above technical solution, in step 5, i sw k is the transmission ratio between the steering wheel angle and the wheel angle. v These are the controller parameters.
[0023] Compared with the prior art, the beneficial effects of the present invention are:
[0024] 1. This invention enhances the control effect of the kinematic model-based tracking controller by selecting a preview point based on vehicle speed adaptation, avoiding the problems of poor cornering tracking accuracy caused by excessive preview distance and poor tracking accuracy caused by excessive preview distance at different vehicle speeds.
[0025] 2. This invention significantly improves the controller's accuracy by introducing touch sensor signals into the controller, enabling the harvester's trajectory to align with the crop ridge line. This not only reduces crop crushing caused by poor trajectory tracking control accuracy but also avoids crop crushing due to erroneous tracking during unmanned harvester operation caused by inaccurate target trajectories.
[0026] 3. By introducing pre-aiming point information, this invention reduces fluctuations in the search for matching points caused by sensor and computer errors, and increases the stability margin of the entire controller. This not only avoids the situation where the vehicle deviates due to the introduction of heading correction, but also reduces the situation where the harvester's steering wheel vibrates rapidly and the vehicle's heading is constantly adjusted. Attached Figure Description
[0027] Figure 1 Overall architecture diagram of the control method of the present invention. Detailed Implementation
[0028] The present invention will be further described in detail below with reference to specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
[0029] A tracking control method for an unmanned harvester based on tactile signal correction and adaptive pre-aiming includes the following steps:
[0030] Step 1: Based on the target trajectory planned by the upper layer of the unmanned harvester and the current actual vehicle position (X,Y) obtained by the combined inertial navigation system, use the binary search method to find the nearest matching point (X,Y) on the target trajectory that matches the current actual vehicle position. r ,Y r ,θ r ,κ r ), where (X) r ,Y r ) represents the coordinates of the matching point, θ r To match the heading angle, κ r To match the curvature of the points, then apply the formula for the distance between two points. Calculate the distance error e between the matching points rd .
[0031] Step 2: The pre-aiming point adaptive selection module quickly finds a pre-aiming point on the target's driving trajectory that is ahead of the matching point and whose pre-aiming distance from the matching point is L, using a binary search method. p The aiming point, and the heading θ based on the aiming point. pr The heading error e of the aiming point is calculated by subtracting the actual heading θ of the vehicle obtained from the sensor. pθ That is, e pθ =θ pr -θ. Where, the aiming distance L p =k lv+L0, where v is the actual vehicle speed obtained from the sensor, and k l L0 and L0 are the aiming distance velocity coefficient and preset aiming distance, respectively, which are obtained by simulation test parameter tuning, and can generally be set to 0.2 and 0.1.
[0032] Step 3: The touch sensor is installed inside the harvester's head. When the vehicle's trajectory does not coincide with the crop ridge line, the crop stem will cause the contour gripper of the touch sensor to shift, thus causing the shaft of the mechanically connected angle sensor to rotate (for the structure of the touch sensor, see Miao Zhonghua, Tian Daqing, He Chuangxin, et al. Research on Automatic Row Alignment Assist Driving Control Technology of Intelligent Cotton Harvester [J]. Agricultural Mechanization Research, 2022(044-010).). This inputs a voltage signal U to the controller, which serves as the output voltage of the touch sensor. At this time, if the curvature κ of the matching point is... r When the value is 0, meaning the harvester enters the straight-line alignment state, the median offset voltage ΔU is calculated. This value represents the degree of deviation between the vehicle's trajectory and the crop ridge line, i.e., the row offset. The calculation formula is ΔU = UU mid U mid This indicates the zero-position voltage output by the touch sensor when the vehicle's trajectory is aligned with the crop ridge line. To accurately adjust the vehicle's trajectory in real time, the alignment correction module calculates the heading correction amount Δe based on the median offset voltage ΔU. θ That is, Δe θ =k p ΔU+k i ∫ΔUdt, where k p For the proportional gain coefficient and k i k is the integral gain coefficient. p and k i These values are obtained through simulation and real experimental testing and parameter tuning, and can generally be set to 0.002 and 0.00001.
[0033] Step 4, calculate the heading error e θ The heading error is caused by the heading error e at the aiming point. pθ And the heading correction amount Δe after the curvature judgment at the matching point θ Composition, the formula can be expressed as e θ =e pθ +[1-sgn(|κ r |)]Δe θ , where [1-sgn(|κ r |)] indicates that Δe is only applied when the curvature of the matching point is zero. θ Calculate the control rate.
[0034] Step 5: In the model-based tracking controller, the harvester is simplified to a two-wheeled bicycle kinematic model. The control law, i.e., the target steering wheel angle, can be designed based on this model. Where i sw The transmission ratio between the steering wheel angle and the wheel angle is an inherent parameter of the harvester, k. v These are controller parameters, which can be obtained through simulation and real experimental testing and adjustment. They can generally be set to 1.
[0035] The above description is only a preferred embodiment of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.
Claims
1. A tracking control method for an unmanned harvester based on tactile signal correction and adaptive pre-aiming, characterized in that, Includes the following steps: Step 1, based on the target travel trajectory of the unmanned harvester and the current actual position of the vehicle ( The binary search method is used to find the closest matching point on the target driving trajectory that is the current actual position of the vehicle. ),in( () represents the coordinates of the matching point. To match the heading angle, To determine the curvature of the matching points, the distance error between the matching points is then calculated. ; Step 2: The pre-aiming point adaptive selection module uses a binary search method to find the pre-aiming point ahead of the matching point on the target's driving trajectory, with a pre-aiming distance of [missing value]. Calculate the heading error of the aiming point. ; Step 3, the line correction processing module adjusts the line offset based on the median offset voltage. Calculate heading correction amount ; Step 4, based on the heading error of the pre-aiming point obtained in Step 2 and the heading correction amount obtained in step 3 Calculate heading error ; Step 5: The model-based tracking controller calculates the matching point distance error obtained in Step 1. and the heading error obtained in step 4 Calculate the target steering wheel angle ; , This refers to the gear ratio between the steering wheel angle and the wheel angle. These are the controller parameters.
2. The tracking control method for an unmanned harvester based on tactile signal correction and adaptive pre-aiming as described in claim 1, characterized in that, In step 1, the target driving trajectory is given by the upper-level planning of the unmanned harvester, and the current actual position of the vehicle is obtained by the combined inertial navigation system on the unmanned harvester.
3. The tracking control method for an unmanned harvester based on tactile signal correction and adaptive pre-aiming as described in claim 1, characterized in that, In step 1, the distance error of the matching point .
4. The tracking control method for an unmanned harvester based on tactile signal correction and adaptive pre-aiming as described in claim 1, characterized in that, In step 2, the pre-aiming distance , The actual vehicle speed obtained by the sensor. and The aiming distance speed coefficient and preset aiming distance.
5. The tracking control method for an unmanned harvester based on tactile signal correction and adaptive pre-aiming as described in claim 1, characterized in that, In step 2, , , This represents the current actual heading of the vehicle.
6. The tracking control method for an unmanned harvester based on tactile signal correction and adaptive pre-aiming as described in claim 1, characterized in that, In step 3 ,in This indicates the zero-position voltage output by the touch sensor when the vehicle's trajectory is not deviated from the crop ridge line. The voltage output in real time for the touch sensor.
7. The tracking control method for an unmanned harvester based on tactile signal correction and adaptive pre-aiming as described in claim 1, characterized in that, In step 3 ,in For proportional gain coefficient and This is the integral gain coefficient.
8. The tracking control method for an unmanned harvester based on tactile signal correction and adaptive pre-aiming as described in claim 7, characterized in that, and The results were obtained through parameter tuning via simulation and real experimental testing, respectively.
9. The tracking control method for an unmanned harvester based on tactile signal correction and adaptive pre-aiming as described in claim 1, characterized in that, In step 4 .