How Well Do Pseudo-Rigid Body Models Capture Real Plants? In-Field Validation of Simulated Blueberry Canes
Authors: Hannah Kolano, Chelse VanAtter, Wei Yang, Cindy Grimm, Joseph R. Davidson
Organizations: Collaborative Robotics and Intelligent Systems (CoRIS) Institute, Oregon State University, Corvallis OR 97331, USA · Department of Horticulture, Oregon State University, Corvallis, OR 97331, USA
Many labor-intensive tasks in fruit production such as pruning require physically interacting with the plant (e.g., pushing, pulling, bending limbs, etc.). Due to increasing labor shortages, there is widespread interest in the adoption of robotics in this area. When the robot must physically interact with the system, a deformable model of the plant is beneficial. Coupled, spring-loaded beams have been used in prior work to build deformable plant models in simulation for learning, planning, and control, but this approach has rarely been validated against the mechanical properties of real-world plants. In this paper, we use this rigid-body model, grounded in real material properties and beam bending theory, to simulate the deformation of blueberry canes under loading. To validate the model, we simulate the canes in MuJoCo and compare their behavior with real-world data collected from probing canes at a commercial farm with a custom testbed. We perform sensitivity analysis on several key modeling variables and show that this approach is highly sensitive to the plant diameter calculations and a priori flexural modulus estimation, which is dependent on season and blueberry variety. We also share our dataset of live blueberry cane deformation, including RGB-D images and measured forces and displacements.
Figures & tables
Fig. 1 : Workflow adopted in this study. Blueberry plants ( left ) are imaged during the dormant season. Individual canes are then extracted from the point cloud and discretized into serial chains of links with compliant joints ( right ). These chains are probed in a physics simulator, with the predicted behavior compared with in-field measurements from a commercial farm.
Fig. 2 : Left: Mobile robot with custom 2 degree-of-freedom cartesian gantry system. Right: Up-close view of components during linear probing of a blueberry cane. Two IMUs installed on the plant are shown at the bottom right of the image.
Fig. 3 : The full testing pipeline. a. Three test canes were manually segmented from the blueberry bush point cloud. The push points are visible as orange marks. b. CloudCompare’s polyline tool was used on each cane. c. The polyline was discretized into segments until the maximum distance between the segmented chain and the polyline was 5 mm. Black dots represent joints. d. MuJoCo was used to simulate the force applied to the cane as a lumped-parameter, rigid body model. The green dot represents the push point. e. The force-displacement slopes from the field data and simulation were compared.
Fig. 4 : Fixture for 3-point flexural testing of blueberry cane samples in a Mark-10 tensile tester. Left: Force was applied with a quarter-inch cylinder at the loading point. Right: The support width could be modified to accommodate various sample geometries.
Fig. 5 : At the 20 second mark on this trial, the pushed cane hits an obstruction, likely the other cane. Left: One accelerometer channel on the pushed cane and three accelerometer channels on an adjacent cane show a noise distribution change at the 20s mark, indicating contact. The change points are detected by the ruptures library. Right: The force-displacement graph shows an elbow at 20s, indicating that more force is suddenly required to push the cane, corroborating that contact occurred. This trial meets the criteria for contact, so the data is truncated at 20s during analysis.
Fig. 6 : The data shows a drop in the flexural modulus of northern highbush blueberry plants between the dormant season ( n=19 ) and the fruiting season ( n=40 )
Fig. 7 : Fruiting season flexural modulus measurements. Averages are indicated with black dashed lines. Left: Fruiting data categorized by the age of the sample. Right: Fruiting data categorized by the variety of the blueberry. Significant differences are marked via Welch’s t-test.
Fig. 8 : A comparison of effective stiffness values on a per-trial basis between the real-world data and the simulation. Left: The full plot of stiffness values plotted against a 1:1 ideal line. Trials from the same bush have the same marker color; the shape of the marker indicates the branch number, and the size indicates the height of the probing point (Low, Middle, High). Trials above the line indicate that the simulation has a stiffness value higher than the real data; trials under the line indicate the simulation was too compliant. Right: An enlarged plot of the gray boxed area from left .
Fig. 9 : Output visualization of the Method of Morris study. Each dot is one study performed on one trial; the cross represents the mean over all simulation trials.
Fig. 10 : The log-ratio error of each trial on the Naïve case and the lowest log-ratio error found when sweeping flexural modulus and diameter factor.
This study presents a method for modeling diverse plant branches by iteratively estimating material parameters to support delicate branch manipulation. Branch manipulation is necessary in agricultural robotics for plant repositioning, stabilizing, and clearing visual obstructions in dense foliage. The proposed method builds a tetrahedral branch model from point-cloud data and simulates its behavior using the finite element method. Using real observed deformation data, it iteratively estimates branch parameters and then computes an optimal path with a deformation-aware motion planner to move and stabilize branches within another robot's field of view. Across 30 trials on branches with varying geometries and material properties, the proposed method reduced the deformation energy by 35.69% while increasing the path length by 8.10% on average.
Madhav Rijal, Rashik Shrestha, Trevor Smith +1
Department of Mechanical and Aerospace Engineering, West Virginia University, Morgantown, WV 26505, USA
The global agri-food sector faces increasing challenges from labour shortages, high consumer demand, and supply-chain disruptions, resulting in substantial losses of unharvested produce. Robotic harvesting has emerged as a promising alternative; however, evaluating and training soft grippers for delicate fruits remains difficult due to the highly variable mechanical properties of natural produce. This makes it difficult to establish reliable benchmarks or data-driven control strategies. Existing testing practices rely on large quantities of real fruit to capture this variability, leading to inefficiency, higher costs, and waste. The methodology presented in this work aims to address these limitations by developing tunable soft physical twins that emulate the stiffness characteristics of real fruits at different ripeness levels. A fiber-reinforced pneumatic physical twin of a kiwi fruit was designed and fabricated to replicate the stiffness at different ripeness levels. Experimental results show that the stiffness of the physical twin can be tuned accurately over multiple trials (97.35 - 99.43% accuracy). Gripping tasks with a commercial robotic gripper showed that sensor feedback from the physical twin can reflect the applied gripping forces. Finally, a stress test was performed over 50 cycles showed reliable maintenance of desired stiffness (0.56 - 1.10% error). This work shows promise that robotic physical twins could adjust their stiffness to resemble that of real fruits. This can provide a sustainable, controllable platform for benchmarking and training robotic grippers.
School of Engineering and Materials Science, Queen Mary University of London, Mile End Rd, Bethnal Green, London E1 4NS, United Kingdom. · Department of Robotics, Ritsumeikan University, Japan
Robotic tree-fruit harvesting is a flagship problem for agricultural automation, but progress is bottlenecked by the cost and irreproducibility of field experiments: an orchard is available only weeks a year, every tree is different, and a control error can permanently damage the crop or the plant. The tree models used in graphics and agronomy are geometrically detailed but physically inert, while the GPU-parallel simulators used in robot learning contain no plausible trees. We present OrchardBench, a physically-grounded, GPU-parallel simulation of apple-orchard trees on the Newton engine. Each tree is grown by a stochastic L-system and instantiated as a fully articulated body: branches are compliant torsional spring-dampers whose stiffness follows Euler-Bernoulli beam theory, they break at a wood modulus of rupture and fall as free hinges, and apples are independent bodies on stem tethers that detach at literature-grounded pull forces and load the branch when pulled. A moving, density-controllable foliage layer occludes the canopy as real leaves do. Every physical parameter is tied to a published source. Per-environment domain randomization makes each batched world a distinct tree, and a mobile manipulator with a wrist depth camera closes the loop with geometric fruit perception and an autonomous harvesting baseline. Careful engineering of the solver and the model lets OrchardBench run many parallel environments at interactive rates on a laptop GPU. We define the tasks and a metric suite spanning harvest completeness, throughput, and plant damage (with a per-canopy-zone breakdown), and report baseline results across foliage, fruit load, terrain, canopy zone, and parallelism. The analytic baseline succeeds on about 40% of the fruit it detects and harvests only about an eighth of the reachable fruit on a tree, leaving clear headroom for novel autonomy approaches.