Wireless blind-spot monitoring system with real-time motion detection and risk estimation using yolov11 and optical flow on nvidia jetson orin nano for toyota hilux 2.4 fx 4x2 mt (w/rear a/c)/
Al-hajri, Sarah Bint Mubarak F.
Wireless blind-spot monitoring system with real-time motion detection and risk estimation using yolov11 and optical flow on nvidia jetson orin nano for toyota hilux 2.4 fx 4x2 mt (w/rear a/c)/ Sarah Bint Mubarak F. Al-hajri, Elizhea C. Camonias, Jason D. Del Rosario, Alexandre Jr. Dela Cruz, and Ivan James Gabriel R. Yray.-- - Manila: Technological University of the Philippines, 2025. - xvii, 177pages: 29cm.
Bachelor's thesis
College of Engineering.--
Includes bibliographic references and index.
This study presents the development and evaluation of a Wireless Blind-Spot
Monitoring System with Real-Time Motion Detection and Risk Estimation using
YOLOv11 and Optical Flow on the NVIDIA Jetson Orin Nano, designed specifically for
utility vehicles such as the Toyota Hilux 2.4 FX 4x2 M/T. The system aims to enhance
driver awareness and safety by providing real-time object detection and motion tracking in
the vehicle’s blind spots through a non-intrusive Heads-Up Display (HUD). It integrates
four ESP32-CAM modules for video capture, wirelessly transmitting data to the Jetson
Orin Nano for processing. YOLOv11 is utilized for object detection while a CUDA-
accelerated optical flow algorithm, in conjunction with OBD-based vehicle speed input, is
employed for estimating the relative motion of surrounding objects.
Experimental results demonstrate strong object detection performance with a mean
Average Precision ([email protected]) of 0.78 and consistent tracking of moving objects. Real-
time processing was achieved with an average inference time of 0.06 seconds per frame,
though memory constraints on the Jetson Orin Nano presented limitations. Speed
estimation tests yielded mean absolute percentage errors ranging from 4.4% to 69.1%,
depending on object distance and motion complexity. A bird’s-eye view (BEV) map was
integrated into the HUD for intuitive spatial awareness.
The system was deployed and evaluated across 30 road trips in urban environments,
with user surveys indicating improved driving confidence and reduced distraction. This
study concludes that the proposed system offers a scalable, affordable, and real-time
solution to blind-spot monitoring, contributing to safer road conditions and intelligent
vehicle systems in developing transportation infrastructures.
Real time motion detector
Bird's-eye view (bev)
Monitoring system
BTH TK 870 / A45 2025
Wireless blind-spot monitoring system with real-time motion detection and risk estimation using yolov11 and optical flow on nvidia jetson orin nano for toyota hilux 2.4 fx 4x2 mt (w/rear a/c)/ Sarah Bint Mubarak F. Al-hajri, Elizhea C. Camonias, Jason D. Del Rosario, Alexandre Jr. Dela Cruz, and Ivan James Gabriel R. Yray.-- - Manila: Technological University of the Philippines, 2025. - xvii, 177pages: 29cm.
Bachelor's thesis
College of Engineering.--
Includes bibliographic references and index.
This study presents the development and evaluation of a Wireless Blind-Spot
Monitoring System with Real-Time Motion Detection and Risk Estimation using
YOLOv11 and Optical Flow on the NVIDIA Jetson Orin Nano, designed specifically for
utility vehicles such as the Toyota Hilux 2.4 FX 4x2 M/T. The system aims to enhance
driver awareness and safety by providing real-time object detection and motion tracking in
the vehicle’s blind spots through a non-intrusive Heads-Up Display (HUD). It integrates
four ESP32-CAM modules for video capture, wirelessly transmitting data to the Jetson
Orin Nano for processing. YOLOv11 is utilized for object detection while a CUDA-
accelerated optical flow algorithm, in conjunction with OBD-based vehicle speed input, is
employed for estimating the relative motion of surrounding objects.
Experimental results demonstrate strong object detection performance with a mean
Average Precision ([email protected]) of 0.78 and consistent tracking of moving objects. Real-
time processing was achieved with an average inference time of 0.06 seconds per frame,
though memory constraints on the Jetson Orin Nano presented limitations. Speed
estimation tests yielded mean absolute percentage errors ranging from 4.4% to 69.1%,
depending on object distance and motion complexity. A bird’s-eye view (BEV) map was
integrated into the HUD for intuitive spatial awareness.
The system was deployed and evaluated across 30 road trips in urban environments,
with user surveys indicating improved driving confidence and reduced distraction. This
study concludes that the proposed system offers a scalable, affordable, and real-time
solution to blind-spot monitoring, contributing to safer road conditions and intelligent
vehicle systems in developing transportation infrastructures.
Real time motion detector
Bird's-eye view (bev)
Monitoring system
BTH TK 870 / A45 2025