A student competition as been added to the upcoming AV Workshop (June 15 & 16): the Rydesafely AV Safety and Validation Pitch Competition.
Are you a university student? Pitch your solutions to uncertainty, video-understanding, and scenario generation based challenges in robotics to a leading startup in the autonomous vehicles safety space! Register by June 11, and teams will receive more details on the questions, constraints and a submission template. Teams will have 72 hours to address questions.
- Registration deadline: 11th June, 2021,
- Submission deadline: 14th June, 2021.
The top teams will present their vision as part of the UofT AV workshop, on Wednesday July 16 3:30-5pm ET. Winning teams will be recognized as Rydesafely Workshop Winners and will receive a gift card.
The challenges:
- Real-time video understanding for autonomous vehicles remains a difficult problem. Teams will propose system design architectures and training methods that can advance the state-of-the-art for video understanding and other relevant downstream tasks such as 3D pose and activity prediction for pedestrians, vehicles and other agents.
- Autonomous driving perception is dominated by learning systems that don’t differentiate between the quality of their own outputs. Teams will propose new approaches to uncertainty estimation for detection, localization, segmentation and prediction networks, exploiting different input modalities including HD maps, object priors and multiple sensor inputs.
- Being able to evaluate the safe performance of learning systems for real-world deployment can be greatly accelerated by employing simulation methods that can rapidly generate the most challenging interactive scenarios, requiring realistic variations in the environment and in behaviours of simulated agents. Teams will propose methods for automating the development of scenario variations that push learning systems to their breaking point efficiently.
Teams may submit to address multiple questions, but will only be eligible to win one prize per team. Teams will be evaluated on technical quality (10), applicability to problem (5), satisfaction of constraints (5), presentation quality (5).
If you are or know a Canadian university student interested in participating, please encourage them to form a team and register today. Register by June 11 to receive details on the questions, constraints and a submission template.
If you have questions or registration difficulty, please contact Kimberly.
About Rydesafely
Rydesafely is accelerating the last step to autonomous driving. Our testing & assessment platform is the only technical solution to hyper target safety tests specifically to client’s autonomous systems. Our Patented tech balances personalised testing with a unique cross-industry data structure that can put a quantifiable metric on client’s system safety. By acting as both driving instructor and driving license to systems, we leapfrog the autonomous industry from generic and endless mileage to decades faster deployment in 10x the time.