Support Center
CASE STUDY
Industry: Autonomous Vehicles
Product: 7.1 MP Atlas® 5GigE cameras
SDK: LUCID's Arena SDK
Autonomous Vehicle Integrates Atlas 5GigE Cameras in aUToronto Autodrive Competition
aUToronto, a student design team from the University of Toronto, has been competing in SAE’s Autodrive Challenge II for the last three years, with plans to participate again in next year’s competition. This competition features 10 university teams from the United States and Canada competing to develop an SAE Level 4 autonomous vehicle. Each year, teams are challenged to iteratively test their vehicle’s abilities, providing students with valuable learning experiences to sharpen their engineering, teamwork, and leadership skills. Students also gain hands-on experience working with industry-standard products, such as LUCID’s Atlas 5GigE cameras.
LUCID is pleased to provide sponsorship to aUToronto, the University of Toronto's self-driving car team, as they compete in SAE’s Autodrive Challenge II.
Challenge
The competition provides each team with a Chevy Bolt EUV vehicle from GM, one of the event’s organizers. Teams must modify the vehicle to incorporate their own autonomous driving technology. While sponsors provide some hardware, such as Intel-based computer platforms and Cepton LiDARs, teams must independently source other crucial components like cameras. This is where LUCID’s support has been pivotal for aUToronto and allowed them to gain a competitive edge.
A robust perception system is essential for an autonomous vehicle. While each visual sensor provides important information, the cameras play a central role in capturing data for processing. LUCID’s sponsorship filled a critical gap by supplying high-quality cameras to the team.
The team must modify a Chevy Bolt EUV with their own autonomous driving technology.
Solution
LUCID sponsored aUToronto by providing four Atlas 7.1 MP GigE Vision cameras. These cameras feed images directly into the team’s deep learning inference engine, where deep neural networks perform 2D object detection. The detected objects are then processed by an object tracker engine, fusing the 2D object detection data with 3D object detection data from LiDAR. LUCID’s camera system provides a reliable and consistent input for these processes, ensuring fast and accurate object tracking.
The Atlas 7.1 MP cameras, which feature Sony’s IMX420 global shutter CMOS sensors over a 5GigE interface, are used for the comprehensive perception system. The Atlas cameras are strategically placed to give the vehicle left, right, and forward views, increasing the system’s coverage. In the forward orientation, one wide-angle camera and one long-range camera provide redundancy in the field of view.
Object and light detection with aUToronto's perception system.
aUToronto's perception system uses four Atlas IP67 5GigE cameras featuring 7.1 MP Sony IMX420 global shutter CMOS sensors.
LUCID’s Arena SDK has streamlined the development of aUToronto’s custom ROS2 camera node. The ease of use and continuous reliability of the Atlas cameras enable the team to focus more on development, rather than troubleshooting hardware issues. Additionally, the IP67 water-resistant rating of the Atlas cameras allows the team to test the vehicle in challenging weather conditions, such as rain and snow, which was crucial for a Canadian team. According to Chad Paik, current perception lead of aUToronto, “As a fun fact, we secured first place in our vehicle presentation last year by pouring water on our sensor suite during the presentation, impressing the judges with the water resistance of the Atlas cameras.”
Conclusion
aUToronto’s experience with LUCID’s cameras has been overwhelmingly positive. The sponsorship has played a critical role in enabling the team to develop a highly capable perception system. The reliability and ease of use of the Atlas cameras have been instrumental in aUToronto’s progress in the competition. LUCID’s support has not only enhanced the team’s performance but also provided them with the opportunity to gain hands-on experience with industry-standard technology. aUToronto expressed their appreciation for LUCID’s contribution and looks forward to future collaborations.
Learn More:
Visit aUToronto, the University of Toronto’s self-driving car team.
Visit the Atlas 5GigE camera product page.