Simulation-Based Learning of Feedback Control Through Planar Robot Motion

Denis Mosconi; João Domingos Pereira, Marcelo Becker — Tue, 05 May 2026 00:00:00 +0000

The understanding of feedback control concepts can be challenging for students when theoretical analysis is not directly connected to observable system behavior. Simulation environments can help bridge this gap by allowing learners to visualize the effects of control actions on system dynamics. This work presents a didactic framework for teaching classical proportional feedback control using the ROS~2 Turtlesim simulator. The proposed approach combines mathematical modeling, controller design, and robotic simulation to illustrate how proportional gain influences system response, convergence speed, and trajectory behavior. Simulation results demonstrate that the framework clearly exposes both theoretical properties and practical implementation effects, including actuator saturation and discretization phenomena. Because the system is simple to implement and relies on widely accessible open-source tools, it provides an effective platform for introducing students to feedback control concepts while simultaneously connecting classical control theory with mobile robotics applications.

Computer Vision for Emotion Identification on Sheep Images

Tue, 05 May 2026 00:00:00 +0000

The rising global demand for meat and dairy products has accelerated the expansion of livestock farming, underscoring the need for advanced technologies to ensure animal welfare and productivity. This research explores the potential of automated monitoring systems, leveraging depth sensors and time-of-flight cameras, to provide valuable insights into environmental conditions, nutrition, health, and productivity. These systems enable the early detection of abnormal behaviors in large-scale farming operations, paving the way for more effective management practices. The study emphasizes the importance of understanding animal needs and introduces advanced models, including EfficientNet, ResNet50, ResNet101, ResNet152, and VGG16, for emotion recognition in sheep. These models achieve impressive accuracy rates ranging from 88% to 93%, significantly enhancing the ability to detect and classify emotional states such as pain. This capability represents a vital component of precision livestock farming, a practice that integrates real-time data with machine learning to support informed decision-making, optimize yields, and mitigate risks. Furthermore, the research highlights methodologies for animal identification, body condition assessment, and pain estimation, showcasing the potential of sophisticated imaging and perception technologies to revolutionize livestock farming. By improving welfare and operational efficiency, these advancements offer a sustainable approach to addressing the growing challenges in modern agriculture.

Journal of Mechatronics Engineering

Simulation-Based Learning of Feedback Control Through Planar Robot Motion

Computer Vision for Emotion Identification on Sheep Images