NEWS & DEVELOPMENTS
Explore industry announcements, technical events and developments in automotive AI safety.
The organization reports its first ISO/PAS 8800 certification, covering management processes at Dongfeng’s product development and research center.
The advertised course addresses conditional independence, variable selection, parameter estimation and inference in Bayesian networks. The public brochure connects quantitative risk assessment to Python exercises using automotive data.
The public agenda introduces Bayesian-network modeling of SOTIF uncertainty, conditional dependencies and hazardous-event probabilities. Advertised Python/Jupyter exercises address acceptance-criteria validation and system tuning.
The advertised course addresses power-supply and electric-powertrain fault tolerance for L3–L5 systems. Its public brochure covers hazards, safe states, FMEA and safety mechanisms for HV/LV batteries, DC/DC converters, inverters and chargers.
Scheduled for 20 November 2026, this AAAM webinar plans to explain AEB operation, research-based performance evaluation and the use of simulation in development and assessment. The public programme lists four contributors and a Q&A. This is an event announcement, without a recording or evaluation results.
The public course agenda covers partitioning, program-flow monitoring, safe communication, mixed-ASIL independence and software diversity. It also addresses safety concepts for Linux/virtualized environments and tools such as ML frameworks.
The advertised course combines AI/ML safety fundamentals with a guided development-and-test workshop. The public brochure lists data engineering, architecture, performance metrics, acceptance criteria and safety mechanisms.
The advertised SOTIF course covers sensor and algorithm limitations and ML/AI integration in ADAS and highly automated driving. Its public agenda includes functional insufficiencies, foreseeable misuse, V&V evidence and release safety arguments.
The ASAM session scheduled for 4 November 2026 features ITK Engineering’s Andreas Bossert on AI development and MLOps for L3–L5 vehicles. Its published abstract connects ISO/PAS 8800 with OpenLABEL and OpenODD for data, scenario and ODD descriptions.
TÜV SÜD reports a joint inspection-organizations meta-analysis: alignment, calibration and ageing can impair functions without diagnostic fault codes. It proposes extending periodic inspection with functional and effectiveness tests.
The official program lists an AVSC overview followed by Uber, Volkswagen, Torc and Motional panelists on remote-assistant training and human factors. Workforce readiness and implementation experience are agenda topics; no full presentation or replay was verified.
NVIDIA introduces Halos SAFE, with dSPACE, IVEX and TÜV Rheinland collaborating on evaluation workflows. The framework connects simulation, software- and hardware-in-the-loop evaluation, road testing and fleet monitoring to generate traceable evidence against defined scenarios and performance criteria.
MIIT publishes five draft mandatory standards for consultation. Relevant automotive-AI items concern combined parking assistance and automated-parking safety requirements; these are consultation drafts.