Aug 2026· INMATEH Agricultural Engineering· 0 citations· 9 references
Abstract
In addressing the high labor costs and low operational efficiency of greenhouse tomato harvesting and separate packaging workflows, this study develops an integrated tomato harvesting robotic system embedded with RGB-D machine vision, 5-degree-of-freedom manipulator and vertical heat-sealing net bag packaging mechanism. The YOLOv8s lightweight detection model trained on self-built multi-light greenhouse tomato dataset (1260 annotated images covering unobstructed, semi-occluded and heavily occluded fruits) is adopted to identify ripe tomatoes with a recognition accuracy of 95.2%, and ImageJ software is introduced to conduct secondary maturity screening via RGB chromatographic analysis. A* global path planning combined with TEB local trajectory optimization realizes autonomous obstacle avoidance navigation of the wheeled mobile platform, while RRT-Connect bidirectional random tree algorithm is applied for obstacle-free grasping trajectory planning inside dense tomato canopies. A total of 120 valid cyclic tests are carried out in simulated greenhouse environment to verify the full-chain automation including fruit detection, in-situ picking and instant bagging. Experimental results show that the average single-fruit processing cycle is 12.1 s, with a picking success rate of 90.8% and bagging success rate of 98.3%. Compared with skilled manual picking and packaging, the overall working efficiency is improved by approximately 30%. This system firstly realizes continuous integrated harvesting and commercial packaging operation for greenhouse tomatoes, providing a feasible technical solution for full-process intelligent protected agriculture.
LLMs are increasingly used for code generation, yet they frequently hallucinate non-existent software packages, creating exploitable entry points into the software supply chain. We make four contributions to this problem. First, we show that prior evaluation methodologies systematically inflate hallucination rates by misclassifying standard-library modules as hallucinations in some languages. For Python, the overestimation reaches 9.4 percentage points. Second, we evaluate seven inference-time defenses for mitigating package hallucinations, including five guided decoding strategies (Greedy, Contrastive, DoLa, Nudging, and Active Layer-Contrastive Decoding), an iterative self-refinement approach (Self-Refine), and a Retrieval-Augmented Generation (RAG)-based defense.. Across eight models spanning five families and four programming languages (Python, JavaScript, Ruby, Rust), RAG reduces the package hallucination rate (PHR) in 18 of 32 model--language configurations. Third, we introduce Package Utility (PU) to assess whether defenses preserve valid and task-relevant recommendations. Among strategies evaluated, Greedy decoding provides the strongest average mitigation--utility trade-off. Fourth, we stress-test all strategies under adversarial prompts seeded with fabricated package names and find that PHR surges by up to 45 percentage points relative to standard prompts, with Ruby consistently the most vulnerable language (80.9--95.2\%). Under adversarial conditions, RAG and Self-Refine outperform all decoding-only strategies, indicating that robust defense requires either external grounding or iterative self-verification when prompts are actively hostile. Our results recast package hallucination as both a measurement problem and a decoding-time control problem, and they demonstrate that the choice of defense must be matched to the threat model and recommendation utility.
Albérick Euraste Djiré, Iyiola E. Olatunji, Melissa Tessa et al.· 1 citation
MAGE explains how externalized knowledge, bounded action, independent evaluation, and retained human authority can compose into a governed engineering environment, and proposes tests of when that environment turns commodity intelligence into durable engineering progress.
James C. Davis, Kelechi G. Kalu, Huiyun Peng et al.· 1 citation
An audit-and-placebo protocol is proposed that separates verifier artifacts, interaction scaffolding, and grounded feedback credit in evaluations of self-evolving test generators in evaluations of self-evolving test generators.
Yunhao Liang, Chengguang Gan, Ruixuan Ying et al.· 0 citations
To address the problems of low bale-forming rate and loose compaction during the mechanized recovery of
residual plastic film in farmland, a three-stage residual film baling device was developed. Based on the discrete
element method (DEM), a flexible thin-shell model of the residual film was established, and dynamic simulation
and parameter optimization of the baling process were carried out using Rocky DEM software. Combined with
single-factor experiments and response surface methodology, the effects of baling chamber inclination angle,
belt type, belt linear speed, and upper belt inclination angle on bale integrity and compaction were analyzed.
The results showed that the bale-forming rate of the three-stage device reached 99.13%, and the bale density
reached 83.75 kg/m³ under the optimal conditions: baling chamber inclination angle of 30°, upper belt
inclination angle of 30°, belt linear speed of 2.3 m/s, and the use of a rough-surface belt. Field validation tests
showed that the average density of the residual film bales was 91.4 kg/m³, with a prediction error of only 2.23
kg/m³ compared with the simulation results. The device operated stably and reliably. This study presents a
high-efficiency residual film baling device with improved bale-forming rate and compaction performance
through structural optimization and parameter tuning, providing technical support for efficient recovery and
pollution control of residual plastic film in farmland.
Wang Guan, Hao-Lin Wang, Zhi-Hong Tian et al.· INMATEH Agricultural Enginee...· 0 citations
This prototype MRG image translocation software was helpful to 69% of patients with binocular diplopia, but limited by large angle strabismus because of the limited instrument field of view.
Edsel B Ing, Kevin Sha, Sarosh Dandoti et al.· Journal of neuro-ophthalmolo...· 0 citations
A diagnostic support system based on a unified web platform that classifies patients according to the risks of developing three diseases based on regularly collected clinical or audio data using classical supervised learning algorithms is presented.
Vedamurthy D R, Dr. Anup Ritti, A. Bibi et al.· International Journal for Re...· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduAug 17, 2026
A USAF cadet and a Lincoln Laboratory researcher found AI chatbots can help nontechnical service members produce viable software applications for their unique problems.
