Class of 2025Richmond Hill, Ontario
- "Assessing the performance of AV perception algorithms in adverse snowy weather conditions" with mentor Noah (Jan. 25, 2024)
Assessing the performance of AV perception algorithms in adverse snowy weather conditions
Started May 2, 2023
Abstract or project description
Within the last 20 years, we have seen a leap in the research towards autonomous vehicles and intelligent driver assistance overall. We have seen this via different controls and features being added to current vehicles, but we hope to see fully autonomous vehicles soon. This field of autonomous vehicles has gained a lot of attention from the public and specifically those in the field of engineering and computer science. There are many different scenarios that every current driver must go through on the daily like pedestrians, road signs, traffic, and on top of that, harsh weather conditions. When looking at harsh weather conditions like a heavy fog in the morning, or heavy snow on the roads that makes it difficult to see lines, it is hard for your average driver to control the car in these situations, let alone a fully autonomous vehicle. These are some very common road conditions around the world, but today’s best lidar sensors, cameras, automotive radars, and GPS’ fail exceptionally in these circumstances. The most important use of autonomous driving in our world today is in harsh weather conditions, so this gap in our current options can be proven fatal. This is true not only for typical cars and taxis, but even 50-ton 18-wheeler trucks! There are very little studies into the effect that weather has on our current types of sensors and systems that are being used in the prototypes of autonomous vehicles, so this is a deep dive into the hardware and computing that are obscured by the severe weather conditions around the world. I will investigate the impact of the different state-of-the-art sensors and systems in our current self-driving cars such as lidar, GPS, camera, radar and more!