ASD - Initiative on Autonomous Systems Design

Start
End
Organiser
Rolf Ernst, Technical University Braunschweig, Germany
Organiser
Dirk Ziegenbein, Robert Bosch GmbH, Germany
Organiser
Selma Saidi, Technische Universität Dortmund, Germany
Organiser
Sebastian Steinhorst, Technical University of Munich, Germany

DATE 2025 Initiative on Autonomous Systems Design (ASD)

Fueled by the progress of artificial intelligence, autonomous systems become more and more integral parts of many Internet-of-Things (IoT) and Cyber-Physical Systems (CPS) applications, such as automated driving, robotics, avionics, industrial automation and smart systems in general. Autonomous systems are self-governed and self-adaptive systems that are designed to operate in an open and evolving environment which is not completely defined at design time. This poses a unique challenge to the design and verification of dependable autonomous systems. Following the successful editions from previous years, DATE is again hosting the Initiative on Autonomous Systems Design. The initiative will include peer-reviewed papers, invited contributions and interactive sessions.

Organizing Committee

  • Rolf Ernst, Technical University of Braunschweig, Germany
  • Dirk Ziegenbein, Robert Bosch GmbH, Germany
  • Selma Saidi, Technical University of Braunschweig, Germany
  • Sebastian Steinhorst, Technical University Munich, Germany

 

Full Program 2025


Overview for Monday/Tuesday 31 March-01 April


Monday 31 March 2025

ASD01: 11h-12h30 (Regular Session)

Session Title: Enhancing Dependability and Efficiency in Automotive and Autonomous Systems


ASD02: 14h-15h30 (Focus Session)

Session Title:  Cybersecurity Challenges of Autonomous Systems


ASD03: 16h30-18h00 (Focus Session)

Session Title: Dynamic, Multi-Agent Sensing-to-Action Loops in Distributed Autonomous Edge Computing Systems: Opportunities and Challenges


Tuesday 01 April 2025

ASD04: 08h30-10h00 (Regular Session)

Session Title: Novel Safety Metrics, Adaptive Patterns for Resilience, and Legal Frameworks in Autonomous Systems Design


ASD05: 11h-12h30 (Focus Session)

Session Title: Teleoperation as a Step Towards Fully Autonomous System


13h15-14h00: ASD Keynote

Title: AI/ML at the Forefront of Semiconductor Evolution: Enhancing Design, Efficiency, and Performance

Speaker: Yankin Tanurhan, Synopsys, United States


 14h-18h: ASD Workshop: How to supervise autonomy?


 

Detailed Program


ASD01 Regular Session “Enhancing Dependability and Efficiency in Automotive and Autonomous Systems” (Monday 31.03, 11h00-12h30)

Chairs: 

Selma Saidi, TU Braunschweig

Dirk Ziegenbein, Robert Bosch GmbH

Abstract: This session explores advancements in automotive and autonomous systems, focusing on achieving predictability, reliability, and efficiency. The session begins with a proposal on extending the AUTOSAR Adaptive standard using the System-Level Logical Execution Time (SL-LET) paradigm to ensure determinism, critical for the predictability of modern automotive systems. The second presentation demonstrates noise perturbation attacks on image segmentation, a core perception component of safety-critical autonomous systems, and how they can be predicted and mitigated. Finally, a framework designed to optimize the efficiency of 3D object detection in autonomous vehicles through pattern pruning and quantization is presented, significantly enhancing real-time performance and energy efficiency on resource-limited platforms.

Talk 1: Modeling the SL-LET paradigm in AUTOSAR Adaptive,

Davide Bellassai (Evidence S.r.l., Scuola Superiore Sant’Anna), Gerlando Sciangula (Huawei and Scuola Superiore Sant’Anna), Claudio Scordino (Huawei), Daniel Casini (Scuola Superiore Sant’Anna) and Alessandro Biondi(Scuola Superiore Sant’Anna). 

Talk 2: Generating and Predicting Output Perturbations in Image Segmenters,

Matthew E. Bozoukov (Miramar college), Nguyen Anh Vu Doan (Infineon Technologies AG & Technical University of Munich) and Bryan Donyanavard (San Diego State University)

Talk 3: UPAQ: A Framework for Real-Time and Energy-Efficient 3D Object Detection in Autonomous Vehicles, 

Abhishek Balasubramaniam (Colorado State University), Febin Sunny (AMD) and Sudeep Pasricha (Colorado State University)


ASD02 Focus Session “Cybersecurity Challenges of Autonomous Systems” (Monday 31.03, 14h-15h30)

Organizer: Sebastian Steinhorst (Technical University of Munich, Germany)

Abstract: With the recent dramatic increase in performance of artificial intelligence and related computing systems, together with advanced sensing, connectivity, and technological platforms, autonomous systems are poised to enter many application domains such as transportation and manufacturing. However, as autonomy increases, the risks of cybersecurity threats are equally rising, requiring the development of sophisticated methods on all layers of autonomous systems architectures. In this session, five experts from different areas of cybersecurity research in industry and academia will present challenges ranging from the physical layer to the system of systems layer of autonomous systems. Using the example of autonomous vehicles to highlight current developments, this session will discuss the efforts necessary to achieve secure and safe autonomous systems. 

The session will comprise individual 10-minute presentations of the five speakers, followed by a panel discussion that will involve the audience and further deepen the exchange.

Talk 1: Physical Layer Integrity Checks, Mridula Singh, CISPA Helmholtz Center for Information Security, Germany

Talk Abstract:  As sensor technology continues to evolve, we can expect even more sophisticated features to make driving more convenient and significantly reduce the risk of accidents and vehicle theft. However, when using insecure sensors, these systems inadvertently create new vulnerabilities. Many of these security vulnerabilities cannot be fixed at the data layer and require physical layer integrity checks to enable secure sensing. In this talk,  I will use the Passive Keyless Entry and Start System (PKES) as an example to illustrate the importance of physical layer security and how the lessons learned from designing secure PKES can be applied to other sensors and communication systems.

Talk 2: Ethernetification of CAN, Alexander Zeh, Infineon Technologies AG

Talk Abstract: We propose CAN adaption layer called CANAL to harmonize data and control plane protocols with existing Ethernet (security) protocols for In-Vehicle Networks (IVNs).

Talk 3: Self-sovereign Identities for Software-defined Vehicles, Christian Prehofer, DENSO Automotive

Talk Abstract: Software-Defined Vehicles (SDV) predominantly enable their functions and features through software. We address the security and trust challenges for software-defined vehicles, based on several example use cases. We elaborate requirements and challenges, including identity, data integrity and trust relationships. Then we argue why the new W3C standard for self-sovereign, decentral identifiers addresses the needs of software-defined vehicles.

Talk 4: Can we achieve acceptable security for autonomous systems?, Mikael Asplund, Linköping University

Talk Abstract: In this short talk I will highlight some of the issues of how we approach cybersecurity when developing new systems in general and safety-critical autonomous systems in particular. Drawing on examples from recent incidents as well as insights from the research frontier in verifiable security, I argue that we need to take a different approach than what is currently the state of practice. This requires questioning some of the fundamental assumptions of modern system engineering where market forces and competition is currently prioritized over transparency, robustness and verifiability.

Talk 5: Managing Cybersecurity in the Autonomous Vehicle Mobility-as-a-Service System-of-Systems, Tobias Löhr, P3 automotive GmbH

Talk Abstract: The deployment of SAE Level 4 (L4) Autonomous Vehicles (AVs) for Mobility as a Service (MaaS) ride-hailing services in the US, Europe, and China has underscored the significant technological advancements achieved in autonomous driving over the past two years. Unlike traditional automotive solutions, these systems feature a complex architecture comprising a retrofitted 'vehicle operating system', a 'self-driving stack', and a 'passenger operating system', which serves as the gateway to the MaaS platform and manages passenger interactions within the car. This leads to an interconnected, interdependent, 'multimodal' architecture; however, it also engenders new challenges in ensuring cybersecurity due to an expanded lifecycle perspective that extends from the development phase through the operational phase to the end of service for AVs. The talk explores strategies to manage the extended supplier & partner networks and apply a unified cybersecurity development and validation approach. It further examines methods to secure multiple entry points or gateways within this system of systems architecture. It introduces methodologies for developing and assuring robust cybersecurity measures across the entire supplier value chain.

Panel discussion with all speakers


ASD03 Focus Session “Dynamic, Multi-Agent Sensing-to-Action Loops in Distributed Autonomous Edge Computing Systems: Opportunities and Challenges” (Monday 31.03, 16h30-18h00)

Organizers: 

Amit Ranjan Trivedi, University of Illinois at Chicago
Saibal Mukhopadhyay, Georgia Institute of Technology
 

Abstract: Autonomous edge computing in robotics, smart cities, and autonomous vehicles depends on seamlessly integrating sensing, processing, and actuation for real-time decision-making in dynamic environments. At its core is the sensing-to-action loop, which continuously aligns sensor inputs with computational models to drive adaptive control. These loops enhance responsiveness by adapting to hyper-local conditions but face challenges like resource constraints, synchronization delays in multi-modal data fusion, and the risk of cascading errors. This focus session examines how proactive, context-aware sensing-to-action and action-to-sensing adaptations can improve efficiency by dynamically adjusting sensing and computation based on task demands, such as selectively sensing a small part of the environment and predicting the rest. Action-to-sensing pathways improve task relevance and resource use by guiding sensing through control actions but require robust monitoring to prevent cascading errors. Multi-agent sensing-action loops extend these benefits through coordinated sensing and actions, optimizing resources via collaboration. Additionally, neuromorphic computing, inspired by biological systems, enables spike-based, event-driven processing that conserves energy, reduces latency, and supports hierarchical control—making it well-suited for multi-agent optimization. Finally, the session highlights the importance of co-designing algorithms, hardware, and environmental dynamics to improve throughput, precision, and adaptability, ultimately advancing energy-efficient edge autonomy in complex environments.

Talk 1: Speculative Edge-Cloud Decoding for Fast and Reliable Decision-Making in Autonomous Systems, Priya Panda, Yale University

Talk 2: Filling in the Sensing Blanks with Generative AI: Ultra-Frugal LiDAR Perception Using Masked Autoencoders for Autonomous Navigation, Amit Ranjan Trivedi, University of Illinois at Chicago

Talk 3: RoboKoop: Efficient Visual Control Representations for Robotics via the Koopman Operator, Saibal Mukhopadhyay, Georgia Institute of Technology 

Talk 4: Neuromorphic Navigation in the Real World: Integrating Real-Time Event-Based Vision with Physics-Guided Planning, Kaushik Roy, Purdue University

Panel Discussion with all speakers


ASD04 Regular Session “Novel Safety Metrics, Adaptive Patterns for Resilience, and Legal Frameworks in Autonomous Systems Design” (Tuesday 01.04, 08h00-09h30)

Chairs: 

Dirk Ziegenbein, Robert Bosch GmbH

Rolf Ernst, TU Braunschweig

Abstract: This session discusses key design aspects for safety, adaptability, and legality of autonomous systems. First, a framework that utilizes cross-channel safety performance indicators (SPIs) to identify and tackle hazardous driving scenarios for automated vehicles and corresponding evidence from a proof-of-concept implementation is presented. The session continues with the introduction of the Reflex pattern, an innovative approach inspired by biological reflexes that enhances system resilience by dynamically responding to fluctuating resources, demonstrated through a drone image processing scenario. Lastly, the role of legal considerations in the design of automated vehicles is explored, especially those intended to transport intoxicated individuals, underscoring the need for a multidisciplinary collaboration among management, marketing, engineering, and legal teams to ensure the development of functionally robust and legally sound systems.

Talk 1:  Identification of Hazardous Driving Scenarios using Cross-Channel Safety Performance Indicators

Caspar Antonius Jacobus Hanselaar (Eindhoven University of Technology), Murali Manohar Selva Kumar (NXP), Yuting Fu (NXP), Andrei Terechko (NXP), Ranga Rao Venkatesha Prasad(Technical University of Delft) and Emilia Silvas  (Eindhoven University of Technology).

Talk 2: Designing Resilient Autonomous Systems with the Reflex Pattern

Julian Demicoli (Technical University of Munich) and Sebastian Steinhorst (Technical University of Munich)

Talk 3: Law as a Design Consideration for Automated Vehicles Suitable to Transport Intoxicated Persons

Marilyn Wolf (University of Nebraska) and William H. Widen (University of Miami)


ASD05 Focus Session “Teleoperation as a Step Towards Fully Autonomous System (Tuesday 01.04, 11h-12h30)

Chair: Marilyn Wolf, U Nebraska-Lincoln

Organizers: Frank Diermeyer, TU München; Rolf Ernst, TU Braunschweig

Abstract: In the foreseeable future, highly automated mobile systems, such as vehicles, robots, UAVs, or trains, will be confronted with difficult situations that require external support. The availability of such external support corresponds to level 4 driving automation and is an essential feature in current robotaxis and automated public transportation. While the first generation of level 4 prototypes relied on safety driver support, commercial systems are gradually moving towards support by teleoperation. Designing teleoperation support for level 4 systems is an end-to-end problem involving two main research and practical challenges, the teleoperation function defining the remote human interface with its scene representation and available control functions, and the real-time communication channel involving wired and wireless segments, which must provide reliable end-to-end data transport. 

Talk 1: Automated vehicle teleoperation – vision and challenges. Frank Diermeyer, TU München 

Talk Abstract: This talk will introduce the fundamentals and challenges of vehicle teleoperation support in the context of highly automated driving. It will explain the system architecture and the safety concept in the context of limited communication guarantees. It will conclude with upcoming trends of teleoperation using the rapidly growing vehicle sensing capabilities.

Talk 2: The role of remote operation in enabling autonomous driving business cases 2025+. Arwed Schmidt, EasyMile 

Talk Abstract: tbd

Talk 3: Reliable real-time communication for teleoperation. Selma Saidi, TU Braunschweig 

Talk Abstract: This talk approaches reliable communication operation from two perspectives. The first one is the application perspective requiring reliable end-to-end connection over heterogeneous networks including handover scenarios under low-latency constraints. The second one is the network perspective addressing suitable wireless resource assignment in cellular (5G and 6G) and WLAN V2X networks as a basis for such applications. 

Panel discussion with all speakers


ASD Workshop “How to supervise autonomy?” (Tuesday 01.04, 14h00-18h00)

Abstract:  Autonomous systems are on their way from an exotic system species to a mainstream technology, where they even reach safety critical and high assurance applications. Yet, efficient design concepts providing the required behavioral guarantees while keeping the benefits of autonomous intelligence are still an open topic, in theory and even more so in engineering practice. Some approaches rely on centralized guidance via infrastructure; others extend individual component capabilities by protective, often model based, functions. Another differentiation is the use of human support in unclear situations, such as in level 4 vehicle automation, vs. independent management with function degradation (e.g. safety layers). The workshop plans to provide examples from very different areas, such as road traffic, UAVs, human assistive robotics, and facility management. The design concepts have a high societal and economic relevance, including legal aspects, such as certification and liability.

Workshop Organization:

After an introduction, the workshop will start with introductory talks by experts from industry and academia from various domains working on different autonomous systems solutions. The session will continue with a panel discussion with these and further experts reflecting the different approaches to autonomy supervision.

Session 1: Introductory talks

Talk 1:  Safeguarding Autonomous Systems: Industrial Practices and Challenges

Presenter: Dirk Ziegenbein, Bosch Corporate Research

Talk 2: Simulation Platform for Agile Governance of Mobile Robots and Drones in University Campuses

Presenters: Hiroyuki Tomiyama, Ritsumeikan University, and Norikazu Hayashi, NEXTY Electronics

Talk 3: Legal Outcomes as System Design Features and the Need for Automation Specific Law Reform

Presenter: William Widen, School of Law, University of Miami  

Talk 4: Autonomous systems for the visual impaired – the supervision challenge

Presenter : Aman Malhotra, CTO & Co-Founder, NorrSpect AI

 

Session 2: Panel

- Dirk Ziegenbein, Bosch Corporate Research

- Bill Widen, School of Law, University of Miami  

- Hiroyuki Tomiyama, Ritsumeikan University

- Marilyn Wolf, U Nebraska-Lincoln

- Arwed Schmidt, EasyMile

 

 


 

Call for Papers for ASD 2025

Areas of Interest

We welcome your contributions within the scope of the event. The main areas of interest include but are
not limited to the following:

  • Concepts, Algorithms and Formal Methods for Autonomy
  • Platforms for Autonomous Systems
    • Architectures (from Chip Level to System Level) for Autonomous Systems
    • Middleware and Frameworks (e.g., ROS) for Autonomous Systems
    • Models and HW/SW Mechanisms for Self-Awareness and Self-Adaptation
  • Design of Autonomous Systems
    • Mastering Emergent and Evolving Behavior (Goals, Constraints, ...)
    • Design, Verification and Test of High-Assurance Learning-Enabled Systems
    • Design Automation and Methodologies for Autonomous Systems
  • Dependability and Trustworthy Autonomy
    • Functional Safety and Assurance Cases for Evolving and Learning-Enabled Systems
    • HW/SW Mechanisms for Resilient Systems
    • Autonomous Systems Security
  • Applications and Case Studies of Autonomous Systems

Types of Contributions

Conference Papers:

The special initiative features regular sessions for presenting novel technical contributions. Submitted papers will undergo a peer-review process and accepted papers will appear in the DATE conference proceedings. All manuscripts should be submitted in pdf format not exceeding 6 pages, following the DATE submission instructions. Detailed submission instructions can be found here: https://www.date-conference.com/call-for-papers#Autonomous-Systems-Design

  • Abstract Submission: Saturday, 09 November 2024 AoE
  • Full Paper Submission: Saturday, 16 November 2024 AoE
  • Acceptance Notification: Tuesday, 17 December 2024 AoE

Special Sessions / Panels / Interactive Sessions:

We additionally solicit proposals for special sessions, panel discussions, interactive sessions and invited talks, targeting emerging challenges in the field of autonomous systems, especially that are of interest to the DATE conference participants. Sessions on industrial perspectives, prototypes and demos are also highly welcome. Proposals must consist of an extended summary of up to 1,500 words in a PDF file, describing the topic, the authors/speakers, and the intended format. 

Submission deadline is Saturday, 16 November 2024 AoE.

Submission Instructions

All submissions should follow the DATE proceedings submission instructions. Please submit your contribution at https://softconf.com/date25/special_ASD/

Program Committee

Marc Zeller, Siemens
Rasmus Adler, Fraunhofer IESE
Paolo Burgio, UNIMORE
Chung-Wei Lin, NTU
Mohammad Hamad, TUM
Jong-Chan Kim, Kookmin U
Nikil Dutt, UCI
Bryan Donyanavard, SDSU
Huafeng Yu, DoT
Jyotirmoy Deshmukh, USC
Marilyn Wolf, UNL
Ichiro Hasuo, NII
Bardh Hoxha, Toyota
Mario Trapp, Fraunhofer
Masaki Gondo, eSOL