A constante influência humana em atividades de manejo animal, além de aumentar os custos de produção, torna-se uma adicional fonte geradora de estresse sobre os animais. Nesse sentido, é necessária a busca pelo desenvolvimento de métodos alternativos de acompanhamento, à distância e em tempo real, das condições físicas dos animais, em conjunto com o controle das instalações. Para a identificação da condição de conforto ou estresse calórico dos animais, alguns indicativos podem vir a auxiliar, tais como a temperatura retal (TR), sendo esse um bom indicador da temperatura do núcleo corporal, bem como, a frequência respiratória (FR). Porém, com a crescente preocupação em relação ao bem-estar animal, vários questionamentos são feitos acerca de métodos invasivos, motivando a busca por alternativas à mensuração da TR. Surge então, como variável alternativa, a temperatura da superfície corporal, buscando-se correlacioná-la com a TR e FR. Sendo assim, com essa pesquisa objetivou-se identificar a região mais adequada da superfície corporal dos suínos, em fase de creche, que apresente a melhor correlação com a TR e FR. Para tal, foi conduzido um experimento, dividido em duas etapas: etapa 1) pré-experimento, sendo conduzido com dois animais em câmara climática, variando-se as condições de temperatura e testando-se métodos de fixação de sensores e coleta de dados inicialmente propostos; e etapa 2) experimento principal. Este último foi conduzido em uma câmara climática, com cinco leitões da raça Landrace x Large White, com 30 dias de idade, provenientes de uma mesma ninhada e do mesmo sexo (fêmea). Variaram-se as condições de temperatura no interior da câmara climática de 14°C a 35,5°C, de modo a atingir situações de estresse tanto por frio quanto por calor, sendo calculada a entalpia para os propósitos do presente estudo. O delineamento estatístico utilizado foi o inteiramente casualizado, com um único fator, a entalpia ambiente, com sete níveis (31,26; 39,56; 51,12; 59,24; 74,82; 82,96; 94,26 kJ.kg de ar seco-1). Foram realizadas medidas repetidas em intervalos de 30 minutos, em seis diferentes regiões corporais: cabeça (A), paleta (B), lombo (C), pernil (D), orelha (E) e timpânica (F). Para as regiões de A a E foram utilizados dois métodos de medida diferentes: datalogger de temperatura Thermochron iButton® - DS1921G e outro via termômetro de infravermelho Fluke® 566. Para a região F, utilizou-se um termômetro de infravermelho de testa e ouvido G-Tech - T1000. Todos com cinco repetições das medidas para cada variável, em cada situação ambiente. Com os resultados obtidos foi possível propor equações de regressão múltipla para a TR e FR, sendo esta última apontada pela análise de componentes principais como a melhor candidata a correlações com as temperaturas da superfície corporal e por ser um bom indicador da situação de estresse térmico. Por meio desses resultados foi possível observar que a região timpânica mostrou-se como a melhor opção para acompanhamento tanto da TR quanto da FR via termômetro de infravermelho (TiF), enquanto que ao utilizar sensores de temperatura da superfície corporal, a melhor opção foi a orelha (TbE) para predição de TR, e a região do lombo (TbC) para predição de FR. / Human constant influence in handling activities, besides raising production costs, becomes another stress source for the animals. In this sense, it becomes necessary the development of alternative methods, that can remotely monitor, in real time, animal\'s physical conditions, together with remote facilities control. In terms of identifying comfort or stressful thermal situations for animals, some indicators can be handy, such as rectal temperature (RT), which is a good indicator of the core body temperature, as well as, the respiratory rate (RR). Although, with the raising concerns about animal welfare, several questions are raised against invasive methods, encouraging the search for alternatives to RT measuring. The determination of body surface temperature values, trying to correlate them to RT and RR, emerges as an alternative. Thus, it\'s aimed, with this research, to identify the most adequate swine body surface region, in nursery phase, which presents better correlation with RT and RR. For that, an experiment was conducted, divided in two stages: stage 1) pre-experiment, being conducted with two animals in a climate chamber, varying temperature conditions and testing sensor fixation and data collection methods previously proposed; and stage 2) main experiment. The last one was conducted in a climate chamber, with five Landrace x Large White piglets, 30 days aged, from the same litter and of the same sex (female). Temperature conditions inside the chamber were varied from 14°C to 35.5°C, attaining stressful situations both for cold and heat, being calculated the enthalpy for this study purposes. The statistical design used was the completely randomized, with one factor only, the ambient enthalpy, in seven levels (31.26; 39.56; 51.12; 59.24; 74.82; 82.96; 94.26 kJ.kg of dry air-1). Repeated measures were taken in 30 minutes intervals, in six different body regions: head (A), shoulder (B), loin (C), ham (D), ear (E) and tympanic (F). For regions from A to E, two different methods were used: temperature datalogger Thermochron iButton® - DS1921G and infrared thermometer Fluke® 566. For region F, a forehead and ear infrared thermometer G-Tech - T1000 was used. All of them had five replicates of measures for each variable, in each environment situation. With the obtained data, it was possible to propound multiple regression equations for RT and RR, the last one being shown by principal components analysis as a better candidate to correlate to body surface temperatures and because it\'s a good indicator of the animal\'s thermal stress situation. By means of these results it was possible to observe that the tympanic region arises as the better option for monitoring RT and RR through infrared thermometer (TiF), while when using body surface temperature sensors, the best option was the ear (TbE) for predicting RT, and the loin region (TbC) for predicting RR.

Artificial Intelligence (AI) is increasingly used in different parts of society for providing decision support in various activities. The agricultural sector is anticipated to benefit from an increased usage of AI and smart devices, a concept called smart farming technologies. Since the agricultural sector faces several simultaneous challenges, such as shrinking marginals, complicated pan-European regulations, and demands to mitigate the environmental footprint, there are great expectations that smart farming will benefit both individual farmers and industry stakeholders. However, most previous research focuses only on a small set of characteristics for implementing and optimising specific smart farming technologies, without considering all possible aspects and effects. This thesis investigates both technical and non-technical opportunities and hurdles when implementing AI in Swedish agricultural businesses. Three sectors in agriculture are scrutinized: arable farming, milk production and beef production. As a foundation for the thesis, a literature review revises former research on smart farming. Thereafter, an interview study with 27 respondents both explores the susceptibility and maturity of smart farming technologies and provides examples of technical requirements of three chosen applications of AI in agriculture. Findings of the study include a diverse set of aspects that both enable and obstruct the transition. Main identified opportunities are the importance smart farming has on the strategic agendas of several industry stakeholders, the general trend towards software technology as a service through shared machinery, the vast amount of existing data, and the large interest from farmers towards new technology. Contrasting, the thesis identifies main hurdles as technical and legislative challenges to data ownership, potential cybersecurity threats, the need for a well-articulated business case, and the sometimes lacking technical knowledge within the sector. The thesis concludes that the macro trend points towards a smart farming transition but that the speed of the transformation will depend on the resolutions for the identified obstacles.

The live production sector of the poultry industry has a growing interest in robotics. Robotics have the possibility to monitor environmental conditions, assess bird welfare, and reduce labor for farm workers and owners. However, interactions of poultry with robotic systems in commercial poultry house environments is largely unknown. Therefore, the goal of this research was to assess the effect of ground and aerial robots on bird stress using avoidance distance (AD) and fleeing speed (FS) as indirect indicators. A low-cost, autonomous robot was also developed to aid in collecting data on environmental conditions in commercial broiler houses. AD and FS were measured for multiple breeds (broilers, brown hens, and white hens) at different bird ages. Poultry-robot AD was greater than poultry-human AD for both broilers and laying hens, indicating that birds tended to avoid the ground robot more than humans. However, birds did become accustomed to the ground robot as reflected by decreasing AD and FS over the trial periods. Aerial drones operated in a commercial broiler house were found to induce a larger AD and higher FS than a moveable sensor package attached to a fixed, overhead rail system. No significant difference was found in the performance of the low-cost, autonomous robot when tested on different substrates (hard tile and litter). However, some differences were found when the robot was operated at different speeds. Results from these studies have provided useful insight into the operation of ground and aerial robots in commercial poultry settings.