Session
Fast and Fiducius: Achieving Efficient Autonomy with Confidence
Event TypeResearch Manuscript
Autonomous Systems (Automotive, Robotics, Drones)
Autonomous Systems
TimeTuesday, July 12th3:30pm - 5:30pm PDT
Location3000, Level 3
DescriptionAutonomous systems need to be trusted on one hand, and efficient, i.e., achieving high performance with limited resources, on the other, to be widely deployed. This session addresses a number of fundamental problems in autonomy along these two directions. The first three papers focus on perception efficiency and safety. The fourth paper addresses critical timing aspects in ROS2 and provides timing predictability for robotics. The last paper targets electric systems and investigates battery voltage prediction. The theoretical breakthroughs of these papers could have significant impacts in automotive, robotics, and drone applications.
Presentations
| 3:30pm - 3:54pm PDT | Enabling Efficient Deep Convolutional NeuralNetwork-based Sensor Fusion for AutonomousDriving | |
| 3:54pm - 4:18pm PDT | Zhuyi: Perception Processing Rate Estimation for Safety of Autonomous Vehicles | |
| 4:18pm - 4:42pm PDT | Processing-in-SRAM Acceleration for Ultra-Low Power Visual 3D Perception | |
| 4:42pm - 5:06pm PDT | Response-Time Analysis for Deadline-Based Scheduling of ROS2 | |
| 5:06pm - 5:30pm PDT | Voltage Prediction of Drone Battery Reflecting Internal Temperature |


