SPS-ASI Webinar: Road to Cost Effective Driver and Occupant Monitoring Systems for Safe (Semi-) Autonomous Vehicles
Date: 26 March 2024
Time: 2:00 PM (CET)
Presenter(s): Dr. Jude Angelo Ambrose
The ASI Webinar Series is an event initiated by the Autonomous System Initiative (ASI) of the IEEE Signal Processing (SP) Society. The goal is to offer the SP community with free webinars looking into the future of autonomous systems. These monthly webinars are hosted on Zoom, with recordings made available in the IEEE ASI’s YouTube channel following the live events.
Abstract
Based on the recent tragic incidents with Level 2 and Level 3 autonomous cars, it is evident that camera-based Driver Monitoring Systems (DMS) are critical to monitor the driver and the passengers to improve road safety. The European Parliament has passed a law that camera-based DMS, among a range of other safety technologies, must be fitted in all new cars, vans, trucks and buses from 2024. Euro NCAP (New Car Assessment Program) is also developing protocols for the technology in cars that will achieve five-star safety ratings from 2025. It is expected that other countries such as Australia, UK and USA will follow suit in passing laws towards mandating DMS on autonomous vehicles. The global demand of driver monitoring system (DMS) is expected to reach 19 million a year from 2022 and automakers are looking at Occupant Monitor Systems (OMS) to monitor the passengers inside the vehicle as well, to better predict the events happening inside the vehicle to correlate with what is happening outside, to improve the reliability of autonomous vehicles. Hence designing cost-effective DMS and OMS systems is imperative for wider roll out of this technology in autonomous vehicles across all the countries and regions.
In this presentation, the practical challenges to design DMS/OMS systems will be explored, especially the embedding challenges to reduce the cost in terms of hardware footprint, memory and power. We will discuss the trend in the System-On-Chip (SOC) market and how that affects the packaging, optical path and cost challenges to design DMS/OMS in different type of vehicles. There are two types of DMS/OMS software setups based on the location being considered for those features in the vehicle: 1/ Standalone SOC target, where an SOC is exclusively allocated to run the DMS/OMS software, and 2/ Integrated SOC target, where the DMS/OMS application is integrated with other software stack such as the Advanced Driver Assistance Systems (ADAS) software stack. This presentation will detail comparative analysis on standalone DMS SOC targets and integrated SOC targets.
The talk will detail Seeing Machine's very own custom Neural Processing Unit (NPU), referred to as Occula, which is ultra low cost and specifically designed to execute DMS/OMS based applications. We will present comparative analysis of Occula to other NPUs in the market.
Biography
Jude Angelo Ambrose - As the Head of Advanced Embedding at Seeing Machines, Angelo is responsible for driving the Embedding activities across all three business pillars, such as automotive, aftermarket and aviation. He technically and operationally drive the product delivery through platform and project teams. As an embedding expert, who worked in multiple sectors including consumer, automotive and R&D, Angelo has over 15 years of experience in delivering hardware and software products. His primary focus has been optimising software and hardware for low-cost and low-power embedded systems. Angelo holds a PhD in Embedded Security from the University of New South Wales, MSc from University of Northumbria, and BSc from University of Peradeniya.
