Workshop Title:
REliable Sensing for REsilient and Sustainable Automation ((ReS)2A)
Abstract: Automated, Intelligent and Connected systems are revolutionising daily life, facilitating inclusive transportation networks, enabling safe exploration of dangerous sites for rescue/monitoring missions, empowering smart cities and housing, facilitating the optimisation of sustainable manufacturing, logistics, energy sytems. This automation revolution can enhance life quality, completely changing how we interact with the environment and the workplace, relieving humans from repetitive and hazardous tasks, promising safer workplaces and travels, more sustainable and inclusive transportation, agritech, etc. With the increased importance of different perception sensor technologies enabling automation (e.g. visible cameras, thermal sensors, event-based cameras, RADAR, LiDAR, ultrasound sensors, etc.) and the use of sensor data by Deep Neural Networks to implement human-like complex decision-making processes, this thought-provoking workshop aims to bring together esteemed sensors and machine learning experts working on automated and intelligent systems. They will explore the main challenges ofresilient and robust sensing, and perceptionand data quality for machine learningin potentially rapidly changing and unpredictable environments, from crowded urban-areas to hostile regions, with extreme temperatures, adverse weather, vibrations, particles, extreme illumination levels (e.g. tunnel, caves, etc.). Perception sensors, combined with other sensor technologies (e.g. inertial measurement units -IMUs, gyroscopes, etc.) are responsible for providing reliable and accurate data to build situational awareness, using traditional and machine learning algorithms, interacting with the environment and surrounding human beings, enabling prompt and safe decisions. However, every and each sensor adds to the complexity and consumption of the automated systems, so the workshop will also explore robust sensing under a sustainability viewpoint.
The workshop will encourage participants to explore (learning from each other and discussing from different viewpoints and different fields of expertise) the implications of perception sensors¡¯ data quality on automated systems for different applications. The focus will be to identify factors and situations severely impeding perception sensor data, discussing upcoming focal areas, technologies. Neededprocessing, and potential mitigations in hardware, software and system architectures will be included in the discussions. This focus can enable to design resilient and robust sensing and perception (using traditional and deep learning algorithms) for autonomous and automated systems in different applications, and also to include in the design aspects such as power efficiency, real-time performance, sustainability.
The objectives of workshop are three-fold:
1) Share the latest technical knowledge between members of the robotics and automation community from industry and academia, with a particular focus on specific applications, such as transportation, smart cities, manufacturing.
2) Enable idea exchange between academics (including early-career researchers and postgraduate students) and industry experts, not only through presentations, but also using question and answer sessions and panel sessions to provoke interactions, guided discussions, and proposal of new ideas, building collaborations and a common understanding of future directions for research and development. High-profile academics and industrial experts will be invited to present the latest findings related to perception sensors, robustness, machine learning, and sustainability.
3) Raise awareness of the latest trends in perception sensors technologies and the newest computer vision and machine learning techniques, also considering neural networks applied to different types of data, e.g. not only images but also 3D point clouds. such as structured tools/techniques that can support a thorough noise factor analysis of automotive perception sensors, and therefore assess the autonomous systems¡¯ safety. To this aim, a short tutorial will present existing frameworks to breakdown noise factors affecting sensors for assisted and automated/autonomous driving.
Objectives of the Workshop
Content of the Workshop
Schedule
Invited Speakers
