Design Of A Biometric Based Vehicle Security System For Theft Deterrence And Intruder Identity Communication

Abstract

Vehicle Theft Remains A Significant Issue, With Over A Million Vehicles Stolen In 2023 Alone. This Thesis Adopts A Comprehensive Biometric-Based Vehicle Security System Designed To Address These Concerns By Identifying The Limitations Of Current Security Technologies, Such As Susceptibility To False Alarms, Dependency On Physical Keys, Inability To Prevent Unauthorized Engine Starts, Limited Remote Monitoring, And Lack Of Intruder Evidence. The Proposed System Seeks To Overcome These Limitations By Integrating Advanced Biometric Modalities, Making The Sophisticated Security Measure Accessible Beyond High-End Luxury Vehicles. The Development Process Involved Selecting Suitable Biometric Modalities, Optimizing The System For Better Accuracy, And Enabling Real-Time Communication Through A Dedicated Android Application. A Vital Feature Of The System Is Its Ability To Provide Immediate Notifications Of Unauthorized Access Attempts. When Motion Or Vibration Is Detected, The System Triggers An Alarm, Initiates A Phone Call, And Captures Images And Fingerprints Of The Intruder, Which Are Then Sent To The Vehicle Owner Via The App. The Owner Can Also Lock The Engine And Track The Vehicle's Location Through The App. Additionally; A Panic Button Is Available To Alert Emergency Responders In Case Of Armed Coercion. Fingerprint And Face Recognition Were Chosen For Their Accuracy, Cost-Effectiveness, Availability, And Adaptability. The R307s Optical Fingerprint Module, Tested Across 310 Scans, Showed A Specificity Of 100%, Sensitivity Of 99.68%,And Accuracy Of 99.84%, With An Average Authentication Time Of 0.30 Seconds. After Undergoing Hyper-Parameter Optimization In Three Stages, The Cnn-Based Deep Learning Face Recognition Model Was Tested On 229 Test Images, And Achieved A Precision, Recall, And F1-Score Of 99%, With An Average Detection Time Of 0.13 Seconds. Hardware Implementation Utilized Arduino And Raspberry Pi 4b Platforms, And The System Was Tested On A Nissan Diesel Td27 Engine And A Jos-Hansen Can Dashboard Training Model To Evaluate Functionality. This Biometric Security System Enhances Vehicle Safety And Recoverability By Providing Prompt Alerts And Identification Of Intruders, Serving As A Deterrent To Theft And Aiding Law Enforcement. The Study Highpoints The Necessity Of Advancing Vehicle Security Technologies Beyond Conventional Methods. By Integrating Intelligent Biometric Sensors And Effective Communication Protocols, The Proposed System Significantly Contributes To Vehicle Safety And The Fight Against Vehicle-Related Crimes.