Chetan Patil@chetangoudar2947484
1
Events attended
1
Submissions made
India
1 year of experience
About me
I am an Electronics and Communication Engineering student at KLE Technological University with a strong interest in digital IC design and semiconductor technology. Through my academic projects, I have developed hands-on experience in Verilog RTL design, FPGA development, RTL verification, and the ASIC design flow.
🤓 Latest Submissions
.png&w=828&q=75)
DriveSense AI
DriveSense AI is an AI-powered driving behavior analysis and personalized coaching system designed to help drivers understand and improve their driving habits. The system analyzes vehicle telemetry and motion data such as vehicle speed, engine RPM, throttle position, acceleration, trip information, and inertial sensor measurements to identify driving events and patterns. The proposed system will detect behaviors such as harsh acceleration, harsh braking, aggressive cornering, inefficient gear changes, and abnormal road impacts. High-frequency accelerometer data will be processed alongside lower-frequency vehicle telemetry to create a synchronized representation of each driving trip. Machine learning models will then classify driving behavior and identify potentially unsafe or inefficient patterns. Rather than simply displaying raw vehicle statistics, DriveSense AI will transform detected events into an interpretable driver profile and personalized recommendations. The system will track behavior across multiple trips, generate a driver score based on safety, smoothness, and efficiency, and highlight areas where the driver can improve. The project will initially use open vehicle-telematics and sensor datasets, including the LEVIN Vehicle Telematics dataset, and will investigate additional road-event datasets for pothole and road-surface detection. The architecture will be designed to support future integration of GPS, road information, and additional vehicle sensors. AMD technologies will be incorporated into the machine-learning workflow for accelerated model training and inference, with ROCm/AMD compute used where appropriate. Performance will be evaluated using measurable metrics such as inference latency, throughput, and model performance. The final goal is to create an intelligent driving coach that moves beyond simple telemetry dashboards and provides actionable feedback that helps drivers understand their behavior and improve over time.
1 Dec 2026
