Register now After registration you will be able to apply for this opportunity online.
Life long learning control for autonomous vehicles
Design tuning, calibration, and learning as a continuous processes happening during the whole lifetime of the robot.
Define a strategy to share information learned by each vehicle with the rest of the fleet.
AVs offer a range of benefits that can positively impact various aspects of our lives from enhanced safety on the roads to improved traffic efficiency and increased accessibility.
One of the main problems of AVs is the robustness to the operating conditions because they work in unstructured uncontrolled environments. There are two types of changes in the life of AVs:
- Wear and tear, which changes the models of sensors and actuators, slowly, in time.
- Fast and abrupt changes in operating conditions. For example, going from asphalt to unpaved roads.
Models are thought to be "unique" for each vehicle and most often are manually tuned during operating downtime over the average of all conditions.
All of this will not be sufficient when AVs will operate outside of a controlled fleet, such as when they will be sold to customers, in a wide range of ODD.
AVs offer a range of benefits that can positively impact various aspects of our lives from enhanced safety on the roads to improved traffic efficiency and increased accessibility.
One of the main problems of AVs is the robustness to the operating conditions because they work in unstructured uncontrolled environments. There are two types of changes in the life of AVs:
- Wear and tear, which changes the models of sensors and actuators, slowly, in time. - Fast and abrupt changes in operating conditions. For example, going from asphalt to unpaved roads.
Models are thought to be "unique" for each vehicle and most often are manually tuned during operating downtime over the average of all conditions.
All of this will not be sufficient when AVs will operate outside of a controlled fleet, such as when they will be sold to customers, in a wide range of ODD.
- Design and implement continuous monitoring of the robot
- Design and implement continuous adaptation of the models
- Deploy custom models to tackle the different environment conditions
- Design and implement a framework to share and adapt the learned models across the fleet
- Design and implement continuous monitoring of the robot - Design and implement continuous adaptation of the models - Deploy custom models to tackle the different environment conditions - Design and implement a framework to share and adapt the learned models across the fleet
