Umudike Journal of Engineering and Technology

In recent years, tricycles have gained popularity as a cost-effective and flexible mode of transportation in developing economies, such as Nigeria. However, there has been a lack of attention paid to reducing tricycle accident rates. This paper introduces a Tricycle Accident Prevention and Control System (TAPCS), which employs a Feed Forward Neural Network (FFNN) and a Rule-Based Accident Control (RBAC) algorithm to mitigate this issue. The system utilizes data collection, augmentation, and feature extraction to train the FFNN for Inter-vehicle Collision Detection (IVC). The RBAC algorithm ensures safe inter-vehicle distances, aligning with Nigerian Federal Road Safety Corps standards (less than 152 meters). Evaluation metrics, including cross-entropy, specificity (94.3%), sensitivity (96.2%), and accuracy (96.9%), validate the effectiveness of the IVC algorithms. Practical testing confirms the system's ability to detect and alert drivers to potential accidents accurately. Comparative analysis demonstrates its superiority over existing algorithms, highlighting reliability as a key strength of the proposed system.