Ethiopian Journal of Technical and Vocational Education and Training

Improving Wireless Image Transmission Using Machine Learning-Based Joint Source–Channel Coding

2026-10-10T14:58:45-07:00

Transmitting digital images reliably over bandwidth-limited wireless links remains difficult, particularly when signal-to-noise ratios (SNRs) are low. Under unfavorable channel conditions, conventional source–channel separation architectures may exhibit substantial deterioration and abrupt reconstruction failure. Although deep joint source–channel coding (DeepJSCC) provides an alternative, improving reconstruction fidelity and resilience remains an important objective. This study proposes Enhanced Deep Joint Source–Channel Coding (EDeepJSCC), an end-to-end convolutional neural network (CNN) framework for image delivery through additive white Gaussian noise (AWGN) and Rayleigh fading channels. Joint optimization of the coding and reconstruction stages enables the approach to operate without a separate explicit channel-estimation procedure. Experiments use CIFAR-10, comprising 50,000 training images and 10,000 test images, and are implemented in TensorFlow with Adam optimization over multiple SNR settings and bandwidth ratios. Performance is characterized using peak signal-to-noise ratio (PSNR), mean squared error (MSE), and channel SNR. The reported comparisons suggest improved image recovery and more progressive performance degradation than conventional JPEG/BPG compression paired with LDPC coding, particularly when channel quality and available bandwidth are limited. The findings support further investigation of EDeepJSCC for resilient visual communication in future wireless systems.

Bridging the Skills Gap: English Language Communication as a Tool for Improving TVET Graduate Employability

2026-10-10T14:43:57-07:00

The growing complexity of international labor markets has increased the importance of employability competencies in addition to technical expertise within Technical and Vocational Education and Training (TVET). Despite sustained investment in Ethiopia’s TVET expansion, graduate unemployment and underemployment remain significant concerns, and English communication is frequently identified as an area of insufficient competence. This study investigated the relationship between English-language communication and employability among Ethiopian TVET graduates. A convergent mixed-methods approach combined quantitative responses from 210 graduates with qualitative evidence obtained from 35 TVET instructors and 25 employers. Descriptive analysis indicated moderate competence across the four language skills, with writing (M = 2.95) and speaking (M = 3.12) performing less strongly than reading and listening. English communication was strongly and positively correlated with employability (r = .68, p < .01). Multiple regression further identified a statistically significant predictive relationship (R² = .54, β = .62, p < .001). Qualitative accounts supported these findings and identified limitations in curriculum provision, instructional practice, and institutional collaboration with industry. The findings indicate the importance of strengthening workplace-relevant English communication through curriculum improvement, English for Specific Purposes (ESP) integration, and increased industry participation to enhance graduates’ preparedness for employment.

Identification of Soil Ero1sion Hotspots and Estimation of Sediment Yield Using Integrated GIS and Hydrological Modeling in the Dirma River Watershed, Lake Tana Basin, Ethiopia

2026-10-10T15:16:48-07:00

Soil erosion and sediment transport pose substantial threats to agricultural production, reservoir sustainability, and water quality in the Dirma River watershed of the Lake Tana Basin. This investigation sought to delineate erosion-prone areas and quantify sediment yield by integrating GIS-based multicriteria assessment with hydrological modeling. Spatial datasets from field surveys and government agencies were processed, and missing observations were estimated using the arithmetic mean method. Erosion susceptibility was mapped in ArcGIS by combining nine thematic layers through Weighted Overlay Analysis (WOA) and the Analytical Hierarchy Process (AHP). ArcSWAT, incorporating the Modified Universal Soil Loss Equation (MUSLE), was used to simulate sediment yield from long-term meteorological, hydrological, and suspended sediment concentration records. The model was calibrated for 2000–2013 and validated for 2014–2020 using SWAT-CUP. The spatial assessment categorized 2.1% of the watershed as low risk, 35.8% as moderate risk, 61.7% as high risk, and 0.4% as very high risk. Soil erodibility (K), support practice (P), and minimum land cover (C) emerged as the most influential model parameters. Model performance was satisfactory, with R² and NSE values of 0.82 and 0.73 during calibration and 0.78 and 0.87 during validation, respectively. Estimated average annual sediment yields ranged from 38 to 345 t ha⁻¹ yr⁻¹. These results offer evidence for targeting conservation interventions, reducing soil loss, safeguarding reservoirs, and supporting agricultural productivity and sustainable watershed management in areas with limited monitoring data.

Identifying Key Predictors of Success and Failure on the Exit Exam at the FDRE Technical and 1Vocational Training Institute

2026-10-10T15:04:06-07:00

The national exit examination serves as a critical quality assurance mechanism in Ethiopian higher education, yet persistently low and fluctuating pass rates at the FDRE Technical and Vocational Training (FTVT) Institute raise concerns about graduate competency and institutional effectiveness, necessitating the identification of specific predictors of examination performance to develop targeted interventions. This mixed-method study identified key predictors of success and failure on the national exit examination at the FTVT Institute, examining academic, psychological, institutional, and preparation-related factors through quantitative analysis of data from 1,836 students who took the exit examination between 2023-2025 using Mann-Whitney U tests, Kolmogorov-Smirnov tests, and Spearman's rank correlation, complemented by qualitative data collected through semi-structured interviews and focus group discussions with students, trainers, and faculty deans analyzed using thematic analysis. The findings revealed that academic preparedness (p = .004) and methods of exit examination preparation (p = .003) significantly discriminated between pass and fail groups, with Spearman's correlation demonstrating a moderate positive relationship between CGPA and exit examination results (rₛ = .456, p < .001), explaining 20.8% of the variance, while psychological and personal factors (p = .419) and institutional and training support (p = .870) showed no statistical difference between groups; however, qualitative findings revealed widespread student anxiety, inadequate resources, disorganized tutorials, technical failures, and systemic institutional deficiencies affecting all students. The study concludes that a dual-track intervention strategy is required, combining targeted academic remediation and preparation skill-building with psychological support and systemic institutional improvements, and recommends implementation of early preparation, structured tutorial sessions, mock examinations, counselling services, reliable technology, curriculum alignment with exam blueprints, and competency-based assessment to improve exit examination outcomes at the FTVT Institute.

Labor Market Demand Analys1is of Technical Vocational Training (TVT) Graduates: Manufacturing Sector in Focus

2026-10-10T14:53:38-07:00

Labor demand analysis is a fundamental foundation for workforce planning, economic policy formation, and sustainable development. The primary goal of this study is to analyze the labor market demand for Technical and Vocational Education and Training (TVET) graduates in Ethiopia's manufacturing sector by identifying important factors that affect labor demand, evaluating their effects, and estimating how labor demand changes over time. This study employed an explanatory research design with a mixed research approach. It relied primarily on secondary data, including twenty-five years (1988 to 2012 EC) of manufacturing survey data from the Ethiopian Statistics Service (ESS), as well as interviews with 31 respondents. A purposive sampling approach was used to select respondents for the interviews. Descriptive, inferential, and thematic analyses were employed to analyze the data. Standard deviation, mean, and percentage were used in descriptive analyses. In addition, Quantile regression, ordinary least squares (OLS), and ARIMA were utilized to explore the factors influencing labor demand and to estimate labor demand over the next five years. The study's key findings were that labor demand was positively correlated with working capital, wages, sales, bank payments (loan payments), and raw material costs. The ranking of labor demand across industries from textile and leather, non-metallic minerals, food and beverage processing, chemicals, pharmaceuticals and personal care, wood and printing, basic metals, to automotive and transport equipment was also identified. Furthermore, labor demand in the manufacturing sector is predicted to decrease over the next five years.

Optimized Hybrid 1CNN with Late Feature Fusion and Batch Normalized Fully Connected Layer for Traffic Sign Classification

2026-10-10T15:22:53-07:00

The objective of this study was to develop a digital traffic sign recognition system using artificial intelligence, particularly deep learning, to improve recognition performance. Novel hybrid CNN architecture was proposed by combining two fine-tuned models through feature fusion before the fully connected layer. To enhance training stability and convergence speed, batch normalization was applied to the fully connected layer. The proposed approach improved the model‟s ability to extract discriminative features and achieved higher accuracy compared with single-model methods. The model was trained on an augmented dataset of 11,000 Ethiopian traffic sign images representing 156 unique classes. Using transfer learning, the convolutional bases of MobileNet and VGG16 were fine-tuned, and their extracted features were fused before the classifier layer. The batch normalization layers reduced over-fitting by simplifying the computational load of the fully connected layer. Experimental results demonstrated that the hybrid model outperformed the individual models, achieving a testing accuracy of 99.18% compared with the baseline accuracies of VGG16 and MobileNet, which were 96.82% and 89.95%, respectively. The hybrid model successfully integrated the strengths of both architectures, resulting in faster training performance and improved classification accuracy.

Fermentation-Driven Changes in Microbial Dynamics, Physicochemical Characteristics, and Proximate Composition of Yam (Dioscorea spp.) in Southwestern Ethiopia

2026-10-10T15:09:40-07:00

Yam (Dioscorea spp.) is an important root and tuber crop in several African communities and has particular cultural value as a prestigious food in southwestern Ethiopia. Although it is widely consumed, traditional processing, mainly boiling, provides limited opportunities for shelf-life extension and contributes to postharvest losses. The present investigation assessed the influence of natural fermentation on the physicochemical characteristics, microbial dynamics, and proximate composition of yam. Mature white (T1) besides brown (T2) yam tubers were processed and fermented for seven days in laboratory micro-silo jars. Changes in physicochemical properties, microbial populations, and nutrient composition were examined during fermentation. Titratable acidity increased progressively from 0.45 to 1.74, whereas pH and moisture declined significantly (p < 0.05), from 6.73 to 3.89 and from 61.3% to 54%, respectively. Yeast and mold counts declined from 5.34 to 4.57 CFU/ml, and Enterobacteriaceae fell below the detection threshold (<1.00 CFU/ml) by fermentation day five. Total viable counts remained comparatively elevated, while lactic acid bacteria increased initially before a subsequent decline. The proximate analysis indicated increases in crude protein and total carbohydrate content, accompanied by reductions in crude fat, crude fiber, and ash. These findings suggest that yam fermentation may enhance nutritional characteristics and restrict spoilage-associated microbial populations, thereby supporting improved preservation and reduced postharvest losses. Further work should assess appropriate starter cultures and the effects of different fermentation periods on antinutritional compounds.