Dataset for herding and predator detection with the use of robots.

Precision livestock farming involves the use of new technologies to improve the performance of farms with low profit margins. Since extensive livestock farming is demanding work requiring continuous supervision, it has not improved as drastically as agriculture. Furthermore, nowadays the world is more aware of the importance of respecting biodiversity and reducing the carbon footprint, for which sustainable animal production is recommended. This is the case of small livestock farms, generally located in unpopulated areas and with difficult generational replacement, due to the tasks involved. The use of robots and other devices equipped with intelligent systems can be useful to the farmer in his daily work. In this way, livestock, specifically flocks of sheep, can be monitored and the presence of potential predators such as the wolf identified. Encountering said predator can be avoided by moving the herd to other, safer pasture areas. This work presents a dataset that contains images and videos that allow detecting, classifying and analyzing flocks of sheep and one of their usual predators, wolves. The dataset includes videos of flocks in different locations, with different lighting conditions and different types of sheep. In addition, it contains images of wolves in natural spaces, which are not usually included in the most common datasets used in computer vision. This dataset can be very useful for the work being carried out in extensive precision livestock farming, to develop intelligent systems, such as a robot, that allow autonomous monitoring and control of a herd. Furthermore, it can be used to analyze animal behavior in the presence of a robot, since some of the images have been acquired with the cameras of a quadruped robot. This dataset has been split into three different Zenodo repositories due to its size. Images of sheep can be downloaded from https://zenodo.org/records/11313800 The images of classes 'Person', 'Wolf' and the depth maps for simulation are publicly available at https://zenodo.org/records/11313966 YOLO annotations are at https://zenodo.org/records/11313165.

Effects of 4 Different Velocity-Based Resistance-Training Programming Models on Physical Performance.

PURPOSE: To examine the effects of 4 programming models (linear [LP], undulating [UP], reverse [RP], and constant [CP]) on physical performance. METHODS: Forty-eight moderately strength-trained men were randomly assigned to LP, UP, RP, and CP groups according to their 1-repetition maximum (1RM) in the full-squat exercise (SQ) and followed an 8-week training intervention using the SQ and monitoring movement velocity for every repetition. All groups trained with similar mean relative intensity (65% 1RM), number of repetitions (240), sets (3), and interset recovery (4 min) throughout the training program. Pretraining and posttraining measurements included, in the SQ, 1RM load, the average velocity attained for all absolute loads common to pretests and posttests (AV), and the average velocity for loads that were moved faster (AV > 1) and slower (AV < 1) than 1 m·s-1 at pretraining tests. Moreover, countermovement jump height and 20-m running sprint time were measured. RESULTS: A significant time effect was found for all variables analyzed (P < .05), except for 20-m running sprint time. Significant group × time interactions were observed for 1RM, AV > 1, and AV (P < .05). After training, all groups attained significant strength gains on 1RM, AV, AV > 1, and AV < 1 (P < .001-.01). LP and RP groups improved their countermovement jump height (P < .01), but no significant changes were observed for UP and CP. No significant improvements were achieved in 20-m running sprint time for any groups. CONCLUSIONS: These different programming models are all suitable for improving physical performance. LP and RP induce similar or greater gains in physical performance than UP and CP.

Vision-Based Module for Herding with a Sheepdog Robot.

Livestock farming is assisted more and more by technological solutions, such as robots. One of the main problems for shepherds is the control and care of livestock in areas difficult to access where grazing animals are attacked by predators such as the Iberian wolf in the northwest of the Iberian Peninsula. In this paper, we propose a system to automatically generate benchmarks of animal images of different species from iNaturalist API, which is coupled with a vision-based module that allows us to automatically detect predators and distinguish them from other animals. We tested multiple existing object detection models to determine the best one in terms of efficiency and speed, as it is conceived for real-time environments. YOLOv5m achieves the best performance as it can process 64 FPS, achieving an mAP (with IoU of 50%) of 99.49% for a dataset where wolves (predator) or dogs (prey) have to be detected and distinguished. This result meets the requirements of pasture-based livestock farms.

Genetic Algorithm Approach to the 3D Node Localization in TDOA Systems.

Positioning asynchronous architectures based on time measurements are reaching growing importance in Local Positioning Systems (LPS). These architectures have special relevance in precision applications and indoor/outdoor navigation of automatic vehicles such as Automatic Ground Vehicles (AGVs) and Unmanned Aerial Vehicles (UAVs). The positioning error of these systems is conditioned by the algorithms used in the position calculation, the quality of the time measurements, and the sensor deployment of the signal receivers. Once the algorithms have been defined and the method to compute the time measurements has been selected, the only design criteria of the LPS is the distribution of the sensors in the three-dimensional space. This problem has proved to be NP-hard, and therefore a heuristic solution to the problem is recommended. In this paper, a genetic algorithm with the flexibility to be adapted to different scenarios and ground modelings is proposed. This algorithm is used to determine the best node localization in order to reduce the Cramér-Rao Lower Bound (CRLB) with a heteroscedastic noise consideration in each sensor of an Asynchronous Time Difference of Arrival (A-TDOA) architecture. The methodology proposed allows for the optimization of the 3D sensor deployment of a passive A-TDOA architecture, including ground modeling flexibility and heteroscedastic noise consideration with sequential iterations, and reducing the spatial discretization to achieve better results. Results show that optimization with 15% of elitism and a Tournament 3 selection strategy offers the best maximization for the algorithm.

3D Tdoa Problem Solution with Four Receiving Nodes.

Time difference of arrival (TDOA) positioning methods have experienced growing importance over the last few years due to their multiple applications in local positioning systems (LPSs). While five sensors are needed to determine an unequivocal three-dimensional position, systems with four nodes present two different solutions that cannot be discarded according to mathematical standards. In this paper, a new methodology to solve the 3D TDOA problems in a sensor network with four beacons is proposed. A confidence interval, which is defined in this paper as a sphere, is defined to use positioning algorithms with four different nodes. It is proven that the separation between solutions in the four-beacon TDOA problem allows the transformation of the problem into an analogous one in which more receivers are implied due to the geometric properties of the intersection of hyperboloids. The achievement of the distance between solutions needs the application of genetic algorithms in order to find an optimized sensor distribution. Results show that positioning algorithms can be used 96.7% of the time with total security in cases where vehicles travel at less than 25 m/s.

Design and Implementation of a Stereo Vision System on an Innovative 6DOF Single-Edge Machining Device for Tool Tip Localization and Path Correction.

In the current meso cutting technology industry, the demand for more advanced, accurate and cheaper devices capable of creating a wide range surfaces and geometries is rising. To fulfill this demand, an alternative single point cutting device with 6 degrees of freedom (6DOF) was developed. Its main advantage compared to milling has been the need for simpler cutting tools that require an easier development. To obtain accurate and precise geometries, the tool tip must be monitored to compensate its position and make the proper corrections on the computer numerical control (CNC). For this, a stereo vision system was carried out as a different approach to the modern available technologies in the industry. In this paper, the artificial intelligence technologies required for implementing such vision system are explored and discussed. The vision system was compared with commercial measurement software Dino Capture, and a dedicated metrological microscope system TESA V-200GL. Experimental analysis were carried out and results were measured in terms of accuracy. The proposed vision system yielded an error equal to ±3 µm in the measurement.