On the 19th of October 2023, in the frame of The European Conference on EDGE AI Technologies and Applications AI4CSM and EDGE AI projects’ organized a joint event to share their results, expertise and insight about Intelligent Mobility systems. At the event, AI4CSM partners were honored to present AI4CSM-related results in two sessions dedicated to Intelligent Mobility Systems and Autonomous Vehicles Domains Applications.
AI4CSM coordinator Jochen Koszescha from Infineon Technologies, in the presentation “AI for Connected and Shared Mobility“ gave an overview of AI4CSM project, which focuses on the design of highly optimized and automated components for digital mobility that allows smart and energy efficient fleet operations in terms of smart routing and platooning achieving low latency in the vehicle intelligence and perception systems. This means not only smart sensor fusion, but also reducing the level of complexity by using functional architectures and trustable automotive AI, to foster Europe’s competitiveness in the global automotive market.
Gerhard Dorn from Virtual Vehicle Research GmbH gave a presentation titled “Edge Perception in Autonomous Driving”. The talk provided introduction to their current work on the autonomous driving Robotaxi within the EU project AI4CSM and focussed on the development and integration techniques of edge perception algorithms using Julia for machine learning tasks. Also, challenges regarding data quality, data sources, uncertainties, and run-time issues, as well as opportunities regarding fast prototyping and integration, were discussed.
Ovidiu Vermesan from SINTEF gave a presentation “Embedding ROS and Edge AI-Based Perception Capabilities in AV Platforms“. Autonomous vehicles operating in fleets are cognitive agents using perception modules to process images, GNSS positions or LIDAR scans for autonomous system decision-making, resulting in actions, such as actuator commands or C2X messages. The high degree of interdependencies between many functional components of autonomous vehicles requires the implementation of new system architectures and new underlying software frameworks. The presentation described the concept for embedding Robot Operating System (ROS) into a compact, scalable, AI-based perception and localization system for autonomous vehicles used for the transport of goods applications. The system is composed of a LiDAR, a stereo camera, an inertial navigation system, and an AI-based embedded computing unit.
In the presentation “Reward Shaping for Safe Reinforcement Learning Controllers“ Lorenz Klampfl from AVL List GmbH gave an overview of formal specifications to automatically provide locally shaped rewards. The advantage of such an approach is that we can address all the above-mentioned challenges efficiently. The performance of controllers based on reinforcement learning heavily depends on designing appropriate reward functions. With poorly designed rewards, an agent can behave unsafely, non-optimal, learn how to obtain the reward without achieving the desired goals, or take too long to learn a policy.
In the presentation “Connectivity and Cognitive Communication for Future Shared Mobility“, Gerald Fritz-Mayer from TTTech Auto AG talked about the current developments w.r.t. cognitive communication, which enables AI-powered methods to access data from both the edge and the cloud, providing fast, reliable, and low-latency data connections. The presentation showcased the respective demonstrators of a proof of concept communication platform and novel technologies to prove wireless data transmission on 28GHz mmW frequency.
Finally AI4CSM SC8 leader and standardization expert Erwin Schoitsch from AIT Austrian Institute of Technology presented the poster about AI/IoT-Technology and Standards for a Sustainable, Resilient and Human-centered Society 5.0. The poster presented joint standardization activities in AI4CSM, A-IQ Ready and AIMS5.0 projects. Standardization following the European vision supports Sustainability, Resillence&Digital Sovereignty as well public acceptance by addressing different domains, the GreenDeal and the UN Sustainable Development Goals.