FURI | Summer 2020

Object Detection With Sensor Fusion and Embedded Systems

Data icon, disabled. Four grey bars arranged like a vertical bar chart.

As the technology available for autonomous vehicles advances, car manufacturers are looking for ways to implement Advanced Driver-Assistance Systems (ADAS) economically. Both in terms of cost and energy used. This is the motivation that leads into investigating embedded systems to create an ADAS using sensor fusion and deep learning. A system using a NVIDIA Jetson TK1, a Basler camera, a LeddarTech lidar sensor, and an implementation of the YOLOv3 algorithm was created to simulate an ADAS. While its power consumption is low, more sophisticated boards must be investigated to run detection algorithms at real time.

Student researcher

Yoga Mahartayasa

Electrical engineering

Hometown: Aurora, Illinois, United States

Graduation date: Spring 2022