Next-generation self-driving cars: convergence of physical AI and integrated digital twins
Next-generation self-driving cars: convergence of physical AI and integrated digital twins
Danoosh Nozari1 Roghayeh Mirzaei2 Armita Yazdani3
1) Master s degree in urban and regional planning, PhD in urban planning, founder of the Urban Tech Center and CEO of Danamond Holding.
2) Master of Urban and Regional Planning, Tarbiat Modares University and Director of Scientific and Specialized Content Development at the Urban Tech Center
3) Bachelor s degree student in Mechanical Engineering at Sharif University of Technology and responsible for technology monitoring and future research at the Urban Tech Center.
IOI Code
:
ABGZ-ACZGDZ
Publication :
The third international conference on new approaches in engineering and basic sciences(icnabs.ir)
Abstract :
With the development of autonomous systems, future vehicles are transitioning from purely data-based systems to physical AI systems. These systems are able to perceive the environment, estimate the situation, make decisions and plan, and directly influence the physical environment through control and actuators. In the meantime, integrated digital twins can serve as an interface layer between the physical and virtual space, enabling real-time synchronization of the vehicle, environment, infrastructure and AI models. Focusing on the convergence of physical AI and digital twins, this research presents a conceptual framework for next-generation autonomous vehicles in which sensor data, physics-based and data-based models, simulation, state estimation, planning and control are integrated in a two-way feedback loop. In this framework, in addition to supporting training and transferring simulation to reality, the digital twin can be used to evaluate driving scenarios, optimize decisions, predict maintenance, and monitor safety. Also, extending this architecture from a single-vehicle digital twin to integrated vehicle and urban infrastructure twins paves the way for collective learning and decision-making among autonomous vehicles. However, realizing this vision requires solving challenges such as real-time latency, uncertainty, continuous validation, interoperability, scalability, and cybersecurity.
Keywords : physical ai self driving cars integrated digital twins cyber physical systems autonomous decision making