Artificial Intelligence-Based Decision Support System for Sustainable Animal Environment Management and Welfare Assessment

Abstract

The increasing demand for sustainable livestock production has intensified the need for intelligent systems capable of continuously monitoring animal welfare, environmental conditions, and farm management practices. Traditional livestock monitoring methods rely heavily on manual observation, resulting in delayed identification of welfare issues, environmental stress, disease outbreaks, and inefficient resource utilization. Recent advances in Artificial Intelligence (AI), Internet of Things (IoT), computer vision, and edge computing provide unprecedented opportunities to automate animal environment management while supporting evidence-based decision-making. However, most existing systems focus on isolated monitoring tasks such as disease detection, behaviour recognition, or environmental sensing, with limited integration of welfare assessment and sustainability indicators into a unified decision support framework.  This study proposes an Artificial Intelligence-Based Decision Support System (AI-DSS) for sustainable animal environment management and welfare assessment. The proposed framework integrates multimodal data acquired from environmental sensors, wearable devices, surveillance cameras, and farm management records. Environmental variables including temperature, humidity, ammonia concentration, carbon dioxide, particulate matter, illumination, and noise levels are continuously monitored alongside animal behavioural indicators such as feeding frequency, locomotion, resting behaviour, drinking activity, posture, social interaction, vocalisation, and body condition. Computer vision techniques combined with deep learning models automatically identify abnormal behavioural patterns, while explainable machine learning algorithms classify welfare status and generate transparent recommendations for farm managers. The proposed AI-DSS employs a hybrid architecture consisting of IoT-enabled data acquisition, intelligent feature extraction, multimodal data fusion, explainable deep learning, and a multi-criteria decision support engine. Sustainability indicators including environmental quality, resource utilization efficiency, animal health, welfare compliance, and productivity are integrated into a comprehensive Animal Sustainability Index (ASI) that assists stakeholders in prioritizing management interventions. The decision engine further provides predictive alerts for environmental stress, disease risk, welfare deterioration, and management optimization through real-time inference.