Agricultural Robot-Centered Recognition of Early-Developmental Pest Stage Based on Deep Learning: A Case Study on Fall Armyworm (Spodoptera frugiperda).

Accurately detecting early developmental stages of insect pests (larvae) from off-the-shelf stereo camera sensor data using deep learning holds several benefits for farmers, from simple robot configuration to early neutralization of this less agile but more disastrous stage. Machine vision technology has advanced from bulk spraying to precise dosage to directly rubbing on the infected crops. However, these solutions primarily focus on adult pests and post-infestation stages. This study suggested using a front-pointing red-green-blue (RGB) stereo camera mounted on a robot to identify pest larvae using deep learning. The camera feeds data into our deep-learning algorithms experimented on eight ImageNet pre-trained models. The combination of the insect classifier and the detector replicates the peripheral and foveal line-of-sight vision on our custom pest larvae dataset, respectively. This enables a trade-off between the robot's smooth operation and localization precision in the pest captured, as it first appeared in the farsighted section. Consequently, the nearsighted part utilizes our faster region-based convolutional neural network-based pest detector to localize precisely. Simulating the employed robot dynamics using CoppeliaSim and MATLAB/SIMULINK with the deep-learning toolbox demonstrated the excellent feasibility of the proposed system. Our deep-learning classifier and detector exhibited 99% and 0.84 accuracy and a mean average precision, respectively.

Design factors for determining the radula shape of Euhadra Peliomphala.

Biomimetics present useful ideas for various product designs. However, most biomimetics only mimic the features of living organisms. It has not been clarified how a given shape is attained through natural selection. This paper presents the design factors that optimize the radula shape of Euhadra peliomphala. Clarifying the important design factors would help designers in solving several problems simultaneously in order to adapt to complicated and multi-functionalized design mechanisms. We measured the radula of Euhadra peliomphala by using a microscope and modeled the grinding/cutting force using the finite element analysis (FEA). We reproduced the natural selection using multi-objective genetic algorithm (MOGA). We compared the solutions when optimizing the radula shape using objective functions of each combination of stress, cutting force, abrasion, or volume. The results show that the solution obtained through two-objective optimization with stress and cutting force was the closest to the actual radula shape.

Spiral Folded Adhesive Plaster Optimization for Laparoscopic Surgery.

Laparoscopic surgery has the advantage of the minimally invasive for patients. However, the surgery is technically difficult for surgeon because high dexterity is required for suturing in the narrow patient's body. This paper presents a sealing method to locate the adhesive plaster at the incision instead of suturing. The objective is to optimize the plaster material and structure. We made the plaster with the thermally cross-linked gelatin film in a spiral fold because thermally cross-linked gelatin film has the high biocompatibility and tackiness, and a spiral fold has great storage efficiency. In 3 experiments, we measured expansion rate, expansion tension, peeling force, and sealing pressure in a variety of gelatin volume and concentration, and the films diameter. From these experimental results, we optimized the films using response surface method. As a result, the plaster is optimal at gelatin volume 10 mL, gelatin concentration 4 wt %, and films diameters 75 mm. We concluded that the optimized spiral folded adhesive plaster is sufficient in terms of the expansion, tackiness, and sealing properties.

Enhanced Frequency Difference of Tumor inside Vibrated Tissue by a Compression Cylinder.

Breast cancer diagnosis has been mostly accomplished by imaging technologies. These methods have the great advantages of detecting the presence and location of breast cancer. However, it's difficult to distinguish between a benign and malignant tumor in a deep position because both tumor types look similar. In this paper, we vibrated the tissue including tumor from skin with a compression cylinder to analyze the frequency difference for distinguishing the tissue type. Before distinguishing a benign and malignant tumor, it's necessary to validate to distinguish between normal tissue and tumor. The objective is to validate the feasibility of using a compression cylinder that emphasizes the differences in frequency between normal tissue and tumor. In two experiments, we measured the displacement on the surface of a breast phantom vibrated by an impulse hammer. We compared the frequency difference with and without a cylinder. We also studied the frequency changes in the relationship between tumor and cylinder position. We found a 5.0 Hz difference in compliance between normal tissue and the simulated tumor using a compression cylinder. The difference in frequency correlated negatively with distance from the simulated tumor to a compression cylinder. We concluded that a compression cylinder would enhance the frequency difference between normal tissue and a simulated tumor with appropriate configuration.

Route bundling in polygonal domains using Differential Evolution.

Route bundling implies compounding multiple routes in a way that anchoring points at intermediate locations minimize a global distance metric to obtain a tree-like structure where the roots of the tree (anchoring points) serve as coordinating locus for the joint transport of information, goods and people. Route bundling is a relevant conceptual construct in a number of path-planning scenarios where the resources and means of transport are scarce/expensive, or where the environments are inherently hard to navigate due to limited space. In this paper we propose a method for searching optimal route bundles based on a self-adaptive class of Differential Evolution using a convex representation. Rigorous computational experiments in scenarios with and without convex obstacles show the feasibility and efficiency of our approach.