9:00 am - 12:00 pm Workshop A: Mile by Mile: Setting Up Your Test-Bed
Safety of passengers is highly prioritised in the development of Autonomous Vehicles. Test-bed aids the facilitation of more accurate research, subsequently ensuring functional and operational safety and security in the driverless cars. Pioneering test-bed acts as a pilot programme and can gradually expand to operational services. The aim is to gain public and transport regulators trust in self-driven road vehicles.
- Defining private, semi-public, public test-beds and considerations involved in expansion of test area
- Exploring dual function of testing the most critical functionality in the real world, while at the same time validating that the simulation in the virtual world is correct
- Identifying scenario-based methodology to analyse country’s traffic conditions and driver behavior
Mahesh ShindeGM, HEAD - ERC (Indoor testing)
Tata Motors Limited
1:00 pm - 4:00 pm Workshop B: Onward to Full Autonomy: Exploring Methods to Increase Robustness of Perception and Address Safety of Autonomous Vehicles
Perception acts like the eyes of a self-driving car, as such, a major component in ensuring safety. AV perception fails sometimes and there is a need to make it more robust. Research have shown that deep learning algorithms for computer vision can be fooled easily so, new methods has been developed to make them more robust. This will directly impact AV to achieve full autonomy and other mission-critical systems that rely on AI.
- Applying new methods to make deep learning more robust by allowing systems to understand the physical world
- Identifying the hidden variables in modelling to allowing accurate inference
- Testing complex systems and exotic test cases through a digital twin
Dr Justin DauwelsAssociate Professor of the School of Electrical and Electronic Engineering
Nanyang Technological University (NTU)