Perception, Deep Learning, and ADAS Workflows for Formula Student Driverless Vehicles
Overview
This webinar covers perception workflows for Formula Student Driverless vehicles, including cone detection using images and point clouds in MATLAB, data labeling, model training, and GPU deployment.
This webinar focuses on the development and use of perception and deep learning workflows and models for the Perception Task at Formula Student competitions. This session will discuss the task of cone detection, starting with using Deep Learning to identify the cones from LIDAR data, and then move on to introducing the entire workflow from data labeling, model building, training, and model deployment to GPUs.
This webinar will also provide an overview and demo of integrating a Simulink Automotive model with ROS and Unreal Engine, along with a brief overview of latest MathWorks tools for Automated Driving, and RoadRunner for Scenario and Scene design.
Highlights
- Detecting cones using LIDAR data
- Object Detection workflow using deep learning
- How to label data
- How to design and train a deep learning network
- How to deploy a deep learning model to a GPU
- What’s new in Automated Driving
- Designing scenes and scenarios with RoadRunner
- Integrating a Simulink model with ROS2 and 3D visualization in Unreal Engine
Who Should Attend
Student competitions, people working on perception, automated driving, Formula Student driverless teams, deep learning models
About the Presenter
Liping Wang
Liping Wang is an Education Program Engineer at MathWorks. She holds a Ph.D. from KTH Royal Institute of Technology in Sweden. Previously, she worked on the research and development of 4G and 5G communication systems at companies including Ericsson and Huawei. Her expertise spans wireless communications, signal processing, and artificial intelligence.
Akshra Narasimhan Ramakrishnan
Akshra Narasimhan Ramakrishnan is a Senior Student Programs Engineer at MathWorks. She is the technical lead for automotive and ADAS competitions, along with supporting FSAE North America MathWorks activities. Her expertise includes sensor data processing, sensor fusion algorithm development, MIL testing in Simulink and MATLAB, along with Scenario and scene design using RoadRunner.
Agenda
| Time | Title |
09:30 - 10:15 |
Using Deep Learning to identify cones from LIDAR data, introducing the entire workflow from data labeling, model building, training, and model deployment to GPUs |
10:15 - 11:00 |
Demo of integrating a Simulink Automative model with ROS and Unreal Engine, overview of latest MathWorks tools for Automated Driving |
Product Focus
This event is part of a series of related topics. View the full list of events in this series.
