Logged Data and Simulation Exploitation for ADAS Validation

Son Tong･Balakrishnan Ayyanar･Kohei Noma･Reiji Takeuchi･Theo Geluk･Daisuke Nagahata (Siemens Digital Industries Software)

Recently, the ADAS industry focuses on scenario-based approach with V&V development cycle (MiL-HiL-real vehicle) for safety validation. In this work, we present our ADAS scenario-based testing process with focus on how to reconstruct and adapt virtual test cases via exploiting real logged vehicle sensors data. The adaptation process helps to make the scenario more realistic and close the gap between virtual and real test cases. In addition, a procedure to reduce the number of scenarios while moving from MiL to HiL and real vehicle testing using transfer learning technologies will also be discussed. Several ADAS use cases will be demonstrated.

Navigating the Global Evolution of Autonomous Mobile Robots in Public Areas

Tim Brockmeyer･Rolf Adomat･Sebastian L. Weiss･Stephan Cieler (Continental Automotive Technologies)･Hatef Tahvildari･Yuanrun Teng (Continental Automotive)

The global adoption of Autonomous Mobile Robots (AMRs) into public spaces presents unique challenges, facing inconsistent regulations in various regions. Moreover, multiple standards complicate integration and interoperability. We'll explore how regulatory frameworks for AMRs compare with those from automated driving, and the implications for industry players. We propose a structured approach to categorize AMR and their deployment strategies, to ensure safe and publicly accepted operation. A concrete example of interaction with humans will illustrate the need for standardization and explain how the proposed approach effectively structures the world of robotics, providing guidance and adaptability to local regulations.

Path Planning for Ultra-Compact Mobility Based on Pedestrian Sensitivity and Risk Potential Models

Toshiyuki Sugimachi･Kotaro Nishihata･Yuichi Sueshige･Hideo Miyachi･Toshiaki Sakurai･Tetsuo Maki (Tokyo City University)･Yoshihiro Suda (The University of Tokyo)

In recent years, area management of new road spaces in which ultra-compact mobility and pedestrians coexist has been considered. It is required that ultra-compact mobility and pedestrians cooperate in the same space where they coexist. In this study, we propose a method to generate the cooperative trajectory based on a pedestrian sensitivity model and a risk potential model with respect to the relative distance between the ultra-compact mobility and pedestrians, and evaluate its effectiveness by simulation experiments.