Salah Al-Darraji

Dr. Salah Al-Darraji

University of Basrah

Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
3 November 2021 in Computers
Verified
Journal Article
Open access
PDF file
Salah Al-DarrajiSalah Al-Darraji
Salah Al-Darraji
Dhafer G. HoniDhafer G. Honi
Dhafer G. Honi
Francesca FallucchiFrancesca Fallucchi
Francesca Fallucchi
AAAyad I. Abdulsada
Ayad I. Abdulsada
Romeo GiulianoRomeo Giuliano
Romeo Giuliano
Hussam Al-AmeenHussam Al-Ameen
Hussam Al-Ameen
Decision-making plays an essential role in the management and may represent the most important component in the planning process. Employee attrition is considered a well-known problem that needs the right decisions from the administration to preserve high qualified employees. Interestingly, artificial intelligence is utilized extensively as an efficient tool for predicting such a problem. The proposed work utilizes the deep learning technique along with some preprocessing steps to improve the prediction of employee attrition. Several factors lead to employee attrition. Such factors are analyzed to reveal their intercorrelation and to demonstrate the dominant ones. Our work was tested using the imbalanced dataset of IBM analytics, which contains 35 features for 1470 employees. To get realistic results, we derived a balanced version from the original one. Finally, cross-validation is implemented to evaluate our work precisely. Extensive experiments have been conducted to show the practical value of our work. The prediction accuracy using the original dataset is about 91%, whereas it is about 94% using a synthetic dataset.

29 February 2024 in Iraqi Journal of Science
Verified
Journal Article
Open access
PDF file
FKFetoh H. Ketafa
Fetoh H. Ketafa
Salah Al-DarrajiSalah Al-Darraji
Salah Al-Darraji
Path planning is a challenging navigation problem that can be handled using multi-objective methods. This paper, present a thre Path planning is a challenging navigation problem that can be handled using multi-objective methods. This paper presents a three-stage multi-objective path-planning method. The first stage is to locate the best or near-best solution path and avoid detected obstacles using a hybrid of the red fox–gray wolf optimizer (RFO–GWO), which finds a route from the start position to the target position. In the second step, a mutation operation using an evolutionary algorithm is utilized to enhance the length, integrity, and smoothness of the route generated by the RFO–GWO algorithm. The final step of the suggested method is refined further using a multiphase technique. By integrating the real sizes of the mobile robots and the size of the barriers and phrasing the issue as a traveling object in the available area, the suggested path-planning method resembles the actual world. The simulation results indicate that this strategy creates the most viable path even in complicated surroundings, overcoming the disadvantages of traditional approaches. Furthermore, when compared to prior path-planning methods, the simulation’s outcomes indicate that the suggested RFO–GWO method is effective in terms of the route, and the strategy is extremely competitive. The results showed a significant improvement, where the total percentage convergence time (in seconds) for RFO–GWO for the three maps was 15%, 12%, and 10%, respectively, whereas it was 35%, 41%, and 43% seconds in GWO and 34%, 35%, and 37% seconds in RFO. There was also a significant improvement in the number of nodes for RFO-GWO (2%, 3%, and 2%) compared to GWO nodes (64%, 65%, and 62%), and RFO nodes (32%, 30%, and 35%) for the same three maps. Subsequently, the smoothness of the path formed by the recommended approach was enhanced using the evolutionary algorithm (EA), where the total percentage length of the path in the worst scenario for GWO was 28% and for RFO was 26% in units, but after improvement with the RFO-GWO with EA, it became 22% in units. stages multi-objective path planning method: The first stage is to locate the best or near-best solution path and avoid the detected obstacles using a hybrid of the Red Fox-Grey Wolf optimizer (RFO-GWO) method, which finds a rout from the start position to the target position. In the second step, a mutation operation by evolutionary, are utilized to enhance the length, integrity, and smooth of the rout generated by the RFO-GWO method. the final step the suggested method is refined further by using the multiphase technique. By integrating both the real sizes of the mobile robots and the size of the barriers and phrasing the issue as a traveling object in the available area, the suggested path planning method resembles the actual world. The simulation results indicate that this strategy creates the best viable path even in complicated surroundings, overcoming the disadvantages of traditional approaches. Furthermore, when contrasted with prior path-planning methods, simulation outcomes indicate that the suggested RFO-GWO in terms of route effectivity, the strategy is extremely competitive.

28 June 2023 in IEEE Access
Verified
Journal Article
Open access
SASuhaib Al-Ansarry
Suhaib Al-Ansarry
Salah Al-DarrajiSalah Al-Darraji
Salah Al-Darraji
ASAsmaa Shareef Shareef
Asmaa Shareef Shareef
Dhafer G. HoniDhafer G. Honi
Dhafer G. Honi
FFFrancesca Fallucchi
Francesca Fallucchi
Path planning is a fundamental aspect of mobile robots and autonomous systems. Methods of path planning are used in robotics to create a path for a robot or autonomous system to follow from a starting position to a goal one while avoiding obstacles and satisfying any additional conditions. There are many different methods to plan the path, including probabilistic methods, heuristics-based approaches, and optimization-based methods. In this paper, we propose a new path planning method called Dynamic Adaptive RRT-connect with a Triangular Segmented Interpolation. Our method improves the traditional RRT algorithms by using an Adaptive-RRT approach, where a random node is chosen as a new node to increase tree exploration. Then, we use a Bi-directional scheme to further enhance the convergence time and cost. Additionally, our method employs Triangular Segmented Interpolation (TSI) method to improve path length and smoothness. Finally, we operate this method within a dynamic environment depending on the Dynamic Window Approach (DWA). Experiments on a variety of environments have shown that our proposed method achieves better than the RRT and RRT-connect algorithms individually in terms of computation time (reduced by 90-80%), cost (reduced by 82-63%), and path length (shorten by 17-12%) besides the ability to avoid dynamic obstacles efficiently.

1 December 2022 in Bulletin of Electrical Engineering and Informatics
Verified
Journal Article
Open access
PDF file
ASAsmaa Shareef Shareef
Asmaa Shareef Shareef
Salah Al-DarrajiSalah Al-Darraji
Salah Al-Darraji
Autonomous mobile robots have become very popular and essential in our life, especially in industry. One of the crucial activities of the robot is planning the path from a start point to a target point, avoiding obstacles in the environment. Recently, path planning received more attention, and many methodologies have been proposed. Path planning studies have shown the effectiveness of swarm intelligence in complex and known or unknown environments. This paper presents a global path planning method based on grasshopper optimization algorithm (GOA) in a known static environment. This algorithm is improved using the bias factor to increase the efficiency and improve the resulting path. The resulting path from this algorithm is further enhanced using an improved version multinomial logistic regression algorithm (MLR). The algorithms were evaluated using three different large environments of varying complexities. The GOA algorithm has been compared with the ant colony optimization algorithm (ACO) using the same environments. The experiments have shown the superiority of our algorithm in terms of time convergence and cost.

29 November 2016
Verified
Book Chapter
Closed access
PDF file
Salah Al-DarrajiSalah Al-Darraji
Salah Al-Darraji
KBKarsten Berns
Karsten Berns
ARAleksandar Rodić
Aleksandar Rodić

30 June 2024 in Basrah Researches Sciences
Verified
Journal Article
Open access
PDF file
HSHala Shaker
Hala Shaker
Salah Al-DarrajiSalah Al-Darraji
Salah Al-Darraji
The growing prevalence of face recognition technology in various applications, including mobile devices, access control, and financial transactions, highlights its importance. However, the vulnerability of face recognition systems to attacks has been demonstrated, underscoring the necessity of addressing potential weaknesses that attackers may exploit. The paper delves into face presentation attack detection (PAD) within biometric systems, which is crucial for ensuring the reliability and security of face recognition algorithms. To address this issue, the paper proposes a method for face presentation attack detection using ResNet-50 in conjunction with multi-modal data, incorporating RGB, depth, infrared (IR), and thermal channels. The method explores diverse strategies to combine results from each modality, investigating various fusion techniques such as majority voting, weighted voting, average pooling, and a stacking classifier. The system has been tested on the WMCA dataset. It exhibits strong performance compared to existing methods, notably achieving an impressive ACER ratio of 0.087% with the stacking classifier. This approach proves effective by consolidating multiple modalities without requiring individual scenario-specific models, indicating promise for real-world applications