VTechWorks staff will be away for the Thanksgiving holiday beginning at noon on Wednesday, November 27, through Friday, November 29. We will resume normal operations on Monday, December 2. Thank you for your patience.

Browsing by Author "Weaver, Ethan"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Torc Robotics: Rain Detection Solution "RainSense AI"
    Lowe, Alyssa; Holland, Darius; Weaver, Ethan; Andrade, Jason; Samuel, Jonathan; Cho, Minho (2023-12-06)
    Torc Robotics makes autonomous vehicles, specifically self-driving trucks. As such, they hope to make their vehicles perform as safely and efficiently as possible. This task becomes more difficult when the vehicles are subjected to adverse weather conditions, such as rain. The first step in making these vehicles perform better in rainy conditions is to give the vehicle the ability to determine whether it is currently “raining” or “not raining”. This information could then help in informing the vehicle to adjust its maneuvers based on the presence of rain. The machine learning solution we developed can make highly accurate predictions as to whether there is rain present in the vehicle’s localized environment. “RainSense AI” is an ensemble learning system that takes in data from various sensors already present on the vehicle. Namely, it incorporates image data from the vehicle’s front-facing camera and LiDAR point-cloud data from the vehicle’s two LiDAR sensors. This data is passed into our system, which consists of two main parts: a Convolutional Neural Network that processes the image data, and an Anomaly Detection algorithm that processes the LiDAR data. Each of the two processes can make individual predictions regarding the presence of rain. Our system uses both of these predictions to generate a final prediction as to whether there is currently “rain” or “no rain”. This system can run in real-time, taking in the most current data from the vehicle’s sensors and continuously generating predictions.