The insufficient loading of vehicles, which leads to a low logistics load factor is a common problem in the logistics industry. This study addresses this issue by utilizing actual shipment data from an automotive company. An effective method has been proposed to improve the company’s logistics efficiency through a scenario-based approach. Two real- world scenarios were developed to enhance vehicle loading performance. Machine learning algorithms were employed to evaluate the shipment performance of these scenarios. For the study, a dataset was generated from the company’s ERP system and real-world shipment data. Since this is a real-world problem, the dataset consisted of unlabeled data. To solve this problem, both supervised and unsupervised learning algorithms were applied. First, unsupervised clustering algorithms were used to group the shipment performance based on similarities. Then, supervised learning algorithms were utilized to classify the data within each group. The average cost was used to evaluate the clusters obtained through the unsupervised methods, while classification performance was measured using supervised machine learning techniques. The scenario-based approach has significantly improved the performance of the shipments as it shows the changes in load factor more clearly. In the actual case, only %25.7 of shipments were high performing, while this percentage gradually increased to %98.4 in the scenarios. The results show that optimizing the load factor makes the transports more efficient and balanced.

Edge computing, a key component of Internet of Things (IoT) systems, enables data processing close to the data source. In low-power, resource-constrained IoT environments, it can reduce dependency on centralized cloud systems, lower communication load, and minimize overall energy consumption. However, transmitting all sensor data from edge devices still incurs significant communication and energy costs. The key research gap is that existing approaches rarely address this inefficiency through selective transmission; specifically, few studies have explored how filtering and sending only anomalous data can improve energy efficiency. To bridge this gap, we propose an anomaly detection-based data reduction method operating at the edge using LoRa technology. We apply two unsupervised learning algorithms, DBSCAN and Isolation Forest, to identify and transmit only anomalous sensor instances. The proposed methods are evaluated against a full-data transmission (FDT) baseline. The key contributions are demonstrated through experimental results: DBSCAN reduces data volume by 98.19% and energy consumption by 98.10%, while Isolation Forest achieves reductions of 97.32% and 97.32%, respectively. These findings confirm that anomaly-driven selective transmission significantly reduces the communication load while ensuring high energy efficiency in edge computing systems.

Bu çalışmada günlük hayatta en sık kullanılan vücut bölümlerinin başında gelen üst uzuvlara ait omuz ve dirsek eklemlerindeki kuvvet ve tork değerlerinin belirlenmesi amaçlanmıştır. Denekten dik konumda iken masada duran belirli bir yükü alması kendisine yaklaştırması ve geri bırakması istenmiştir. Deneğin bu görevi yerine getirirken gerçekleştirdiği hareketler esnasında omuz ve dirseğinde meydana gelen eklem kuvvet ve torklarının belirlenebilmesi için Newton-Euler metodu kullanılarak dinamik bir model oluşturulmuştur. Eklemlere ait konum verilerinin ölçülmesinde giyilebilir hareket sensörleri kullanılmıştır. Bu sensörlerden alınan hareket verileri ile oluşturulan dinamik model kullanılarak eklemlerdeki tork ve bağ kuvveti değerleri hesaplanmıştır. Elde edilen hesaplama sonuçları karşılaştırmalı olarak değerlendirilerek eklem kuvvet ve torklarının hangi durumlarda arttığı ve ne gibi tedbirlerle azaltılabileceği ortaya konulmuştur.

Manual material lifting is associated with risk factors that can lead to musculoskeletal disorders. Body mass index (BMI) plays a critical role in determining the health status of individuals and the burden on the musculoskeletal system. This study aims to compare the lifting behaviors of underweight and healthy weight female participants through the net reaction moment and g-force effects on the L5/S1 joint. For this purpose, participants are asked to perform two different types of manual material lifting actions: stooping and squatting. Joint movements are analyzed using IMU sensors and a 3D kinematic model and dynamic calculations are performed using the Newton-Euler method. Squatting lifting caused higher peak extension moments at the L5/S1 joint compared to stooping lifting. Underweight participants experienced lower moments during squat lifting than healthy weight participants. While there is no significant difference between subjects with healthy weight and subjects with underweight in the evaluations made with peak net reaction moments, a significant difference is revealed in the comparisons made on g-force, which is an evaluation criterion free of mass. This study revealed the critical role of BMI on manual material lifting biomechanics. Although underweight females have a mechanical advantage when manual material lifting, they may be less maneuverable due to imbalance, increasing their risk of injury, so ergonomic training and individualized lifting strategies should be developed.