Co-founder and CTO, Soar Robotics
Cloud-connected robotics intelligence platform for ad hoc drone networks
Most wireless communication technologies are built on top of a very sophisticated theoretical framework whose foundations were laid more than seventy years ago by Claude Shannon. Although these theories are now the main pillars of telecommunications, they exhibit inherent limitations in their practical use. The difficulty of optimizing very complex equations in real-time, the high-dimensional nature of wireless applications, and the dependency on empirical data for creating models make wireless communications a perfect candidate for utilizing deep learning. While deep learning technologies have been successfully incorporated into countless familiar applications, their use in wireless communications is relatively new but has already demonstrated benefits. This talk will start with a discussion of the challenges of wireless communication, generally and in the context of autonomous vehicles, followed by a summary of deep learning methodologies and the most relevant artificial neural networks. Next, we will thoroughly discuss various applications of deep learning to address some of the most difficult problems in wireless communications. Furthermore, we will explore some of the proposed solutions that address the shortcomings of current communication systems used in autonomous mobility (e.g., urban air mobility, industrial drones, last-mile delivery, self-driving cars). We will conclude by presenting some open research questions whose answers may pave the way to the coming breakthroughs in wireless communication technologies.
Deniz is currently working as the technical co-founder of Soar Robotics. He is an electrical and electronics engineer specialized in embedded hardware and software of aerial robots, with 7+ years of experience in AI-powered autonomous drones research and development. He is currently developing the nex generation of telecommunications technologies for autonomous mobility at Soar Robotics. Previously, he co-founded a drone-in-a-box company called Robostate, where he led engineering teams and actively participated in development of one of the first fully-functional autonomous industrial drone systems in the world. Since 2015, he has been actively involved in the development and implementation of deep neural networks, especially in the context of industrial transformation. He assisted many big and medium-sized companies in Healthcare, Energy, Retail, Pharma, Government, Automotive, Construction and other verticals with successful end-to-end AI transformation and automation implementations.