Search

Agricultural robotics is a complex, challenging, and exciting research topic nowadays. However, orchard environments present harsh conditions for robotics operability, such as terrain irregularities, illumination, and inaccuracies in GPS signals. To overcome these challenges, reliable landmarks must be extracted from the environment. This study addresses the challenge of accurate, low-cost, and efficient landmark identification in orchards to enable robot row-following. First, deep learning, integrated with depth information, is used for real-time trunk detection and location. The in-house dataset used to train the models includes a total of 2453 manually annotated trunks. The results show that the trunk detection achieves an overall mAP of 81.6%, an inference time of 60 ms, and a location accuracy error of 9 mm at 2.8 m. Secondly, the environmental features obtained in the first step are fed into the DWA. The DWA performs reactive obstacle avoidance while attempting to reach the row-end destination. The final solution considers the limitations of the robot’s kinematics and dynamics, enabling it to maintain the row path and avoid obstacles. Simulations and field tests demonstrated that even with a certain initial deviation, the robot could automatically adjust its position and drive through the rows in the real orchard.

A series-parallel hybrid banana-harvesting robot was previously developed to pick bananas, with inverse kinematics intractable to an address. This paper investigates a deep reinforcement learning-based inverse kinematics solution to guide the banana-harvesting robot toward a specified target. Because deep reinforcement learning algorithms always struggle to explore huge robot workspaces, a practical technique called automatic goal generation is first developed. This draws random targets from a dynamic uniform distribution with increasing randomness to facilitate deep reinforcement learning algorithms to explore the entire robot workspace. Then, automatic goal generation is applied to a state-of-the-art deep reinforcement learning algorithm, the twin-delayed deep deterministic policy gradient, to learn an effective inverse kinematics solution. Simulation experiments show that with automatic goal generation, the twin-delayed deep deterministic policy gradient solved the inverse kinematics problem with a success rate of 96.1% and an average running time of 23.8 milliseconds; without automatic goal generation, the success rate was just 81.2%. Field experiments show that the proposed method successfully guided the robot to approach all targets. These demonstrate that automatic goal generation enables deep reinforcement learning to effectively explore the robot workspace and to learn a robust and efficient inverse kinematics policy, which can, therefore, be applied to the developed series-parallel hybrid banana-harvesting robot.

Potato machinery has become more intelligent thanks to advancements in autonomous navigation technology. The effect of crop row segmentation directly affects the subsequent extraction work, which is an important part of navigation line detection. However, the shape differences of crops in different growth periods often lead to poor image segmentation. In addition, noise such as field weeds and light also affect it, and these problems are difficult to address using traditional threshold segmentation methods. To this end, this paper proposes an end-to-end potato crop row detection method. The first step is to replace the original U-Net’s backbone feature extraction structure with VGG16 to segment the potato crop rows. Secondly, a fitting method of feature midpoint adaptation is proposed, which can realize the adaptive adjustment of the vision navigation line position according to the growth shape of a potato. The results show that the method used in this paper has strong robustness and can accurately detect navigation lines in different potato growth periods. Furthermore, compared with the original U-Net model, the crop row segmentation accuracy is improved by 3%, and the average deviation of the fitted navigation lines is 2.16°, which is superior to the traditional visual guidance method.

A row-following system based on end-to-end learning for an agricultural robot in an apple orchard was developed in this study. Instead of dividing the navigation into multiple traditional subtasks, the designed end-to-end learning method maps images from the camera directly to driving commands, which reduces the complexity of the navigation system. A sample collection method for network training was also proposed, by which the robot could automatically drive and collect data without an operator or remote control. No hand labeling of training samples is required. To improve the network generalization, methods such as batch normalization, dropout, data augmentation, and 10-fold cross-validation were adopted. In addition, internal representations of the network were analyzed, and row-following tests were carried out. Test results showed that the visual navigation system based on end-to-end learning could guide the robot by adjusting its posture according to different scenarios and successfully passing through the tree rows.

Traveling along straight lines and headland turning are two common motions during the automatic guidance of agricultural machines. However, most studies focus on accurately following parallel tracks in the field and seldom consider the maneuvers at the end of each row. Moreover, numerous studies have mainly focused on planning the global trajectories in the entire field to increase field efficiency, and very few works are related to maneuver generation and vehicle control in headlands. In this study, a feedforward-plus-proportional–integral–derivative controller for a differential drive robot in headland turning was developed. The feedforward-plus-proportional–integral–derivative controller consisted of a feedforward and a feedback loop. The feedforward loop was designed based on the heading errors of a lookahead point on the planning path. The feedback loop was designed based on radial errors to improve the tracking accuracy. Field comparison tests of feedforward, feedback, and feedforward-plus-proportional–integral–derivative controllers were conducted. Experimental results showed that the feedforward-plus-proportional–integral–derivative had better tracking results and took less turning time.

This paper proposed a quaternion and complementary filter-based attitude estimation algorithm for agricultural implements in dynamic conditions. Built a minimal attitude estimation system using micro-electro-mechanical-system (MEMS) based sensors and STM32F446RC core-processor. The inertial measurement unit (IMU) ADIS 16445 is composed of a triaxial gyroscope and a triaxial accelerometer. Utilizing quaternion for attitude representation while combining the complementary filtering algorithm, the estimation error from the gyroscope was corrected during the data processing by the accelerometer. The experiments were conducted both on the SGT320E triaxial turntable platform and the ZP9500 self-propelled boom sprayer with the assistance of antenna positioning and attitude module BD982 in the field. The triaxial turntable platform dynamic experiment result showed that the average error of the attitude estimation was 0.15° and the maximal measurement error was 0.45°. Meanwhile the results in the field showed that the average error was 0.42° and the maximal measurement error was 0.62°, which satisfied the precise operation requirement of the agricultural equipment. The method proposed in this paper helps to reduce the costs of resources and improve the quality of precision farming especially the performance of agricultural implements.

In this paper, a PI path tracking controller based on look-ahead point information for a differential-drive crawler-type robot was developed. The robot is equipped with Global Navigation Satellite System (GNSS) antennas and a receiver module which are used as navigation sensors to obtain real-time estimates of the position and heading errors of the robot. To control the position of the robot with respect to the target path, a look-ahead point was generated by using Euler Transformation. The look-ahead point lateral errors were then controlled by the designed controller. Forward and backward path tracking tests with different look-ahead distances were conducted to examine the controller performance of the robot and the effect of changing look-ahead distance. Through experiment results, it was noted that the lateral errors mean value was within 0.05 m for forward path tracking and 0.06 m for backward path tracking while the heading errors mean value was within 0.5 degree for forward path tracking and 1 degree for backward path tracking. The proposed algorithms were suitable for the differential-drive robot platform for forward and backward path tracking with a reasonable look-ahead distance.

A row-following system based on machine vision for a picking robot was designed in our previous study. However, the visual perception could not provide reliable information during headland turning according to the test results. A complete navigation system for a picking robot working in an orchard needs to support accurate row following and headland turning. To fill this gap, a headland turning method for an autonomous picking robot was developed in this paper. Three steps were executed during headland turning. First, row end was detected based on machine vision. Second, the deviation was further reduced before turning using the designed fast posture adjustment algorithm based on satellite information. Third, a curve path tracking controller was developed for turning control. During the MATLAB simulation and experimental test, different controllers were developed and compared with the designed method. The results show that the designed turning method enabled the robot to converge to the path more quickly and remain on the path with lower radial errors, which eventually led to reductions in time, space, and deviation during headland turning.

Background IL36RN mutation has been identified as one pathogenesis of generalized pustular psoriasis, but the existence of GPP patients without mutation makes this controversial. Objective Our study aimed at assessing the differences in clinical profiles of children with GPP, with and without IL36RN mutation. Methods An ambispective case series study involved review of the records of 66 childhood patients with pediatric-onset GPP and without previous psoriasis vulgaris. Results c.115 + 6T > C was the most common mutation in this Chinese population with GPP alone. The age at onset was nearly halved in the homozygotes/compound heterozygotes than in IL36RN-negative patients. Besides a more severe inflammatory progression, three minor signs could prioritize patients with GPP for IL36RN screening (confluent lakes of pus (P = 0.002), perianal erosion (P = 0.014), and flexural erosion (P = 0.007)). More patients with the pathogenic mutation converted to ACH than those without mutation (χ2 = 4.773, P = 0.029). Children with GPP with or without IL36RN mutation responded well to oral low-dose acitretin, but IL36RN-positive cases suffered a much higher half-year recurrence rate after withdrawl of acitretin treatment(χ2 = 10.370, P = 0.001). Conclusions Specific clinical features can remind dermatologists of the necessity of sequencing diagnosis. The mild pustular phenotype of those without mutation may imply the possible role of the epigenetic changes of IL36RN, or other IL36-blockers in the pathogenesis. Pediatric patients with GPP alone, both with and without IL36RN mutation responded well to low-dose acitretin.

Geographic tongue (GT) is a benign inflammatory disorder of unknown etiology. Epidemiology and histopathology in previous studies found that generalized pustular psoriasis (GPP) is a factor associated with GT, but the molecular mechanism remains obscure. To investigate the mechanism of GT, with and without GPP, three cohorts were recruited to conduct genotyping of IL36RN, which is the causative gene of GPP. In a family spanning three generations and diagnosed with only GT (“GT alone”), GT was caused by the c.115+6T>C/p.Arg10ArgfsX1 mutation in the IL36RN gene. An autosomal dominant inheritance pattern with incomplete penetrance was observed. In the cohort consisting of sporadic cases of “GT alone” (n = 48), significant associations between GT and three IL36RN variants (c.115+6T>C/p.Arg10ArgfsX1, c.169G>A/p.Val57Ile and c.29G>A/p.Arg10Gln) were shown. In the GPP patient cohort (n = 56) and GPP family member cohort (n = 67), a significant association between the c.115+6T>C mutation and the simultaneous presence of GPP and GT was observed when compared to the presence of GPP without GT (P

According to the features of headland turning, new path planning and headland turning control algorithms for autonomous agricultural machine were presented in this paper. The turning path planning considered both the minimum turning radius and headland space was created by applying three straight lines. A path tracking algorithm based on the improved pure pursuit model was also proposed. This study used the BP neutral network to implement the dynamical look-ahead distance control for the improved pure pursuit model. Based on simplified bicycle kinematics model parameters, MATLAB/Simulink simulation results showed that the path planning algorithm were simple, occupied small headland space while still had a high tracking accuracy. The control method is feasible and practical.