Safe Autonomy – Quantitatively
Sensmetry is an R&D company focused on autonomous driving safety and standards development.Our vision
Safety via Rigorous Quantification
We believe that the road to autonomous vehicle safety is via rigorous analysis & quantification. Sensmetry employs mathematical modeling techniques and large-scale data analytics to evaluate the perception, planning and control layers in the AV/ADAS stack.
Safety via Verifiable Specifications
Our goal is to help companies and institutions transition from lax, convoluted safety definitions to formally expressed, modular system requirements which can be independently analyzed and measured using statistical tools and big data software.
Safety via Transparency
Trust is crucial in establishing safety. At Sensmetry our mission is to bring transparency and trust to the public by providing unambiguous, interpretable safety arguments that both our clients and the end-users can rely on.
Services
Sensmetry offers consulting services and software tools for accelerating safe autonomous system development. All of our work processes and tools derive from our proprietary Probabilistic Modular Safety methodology.
ODD Modeling & Coverage Analysis
We help our clients build up a probabilistic Operational Design Domain (ODD) model. The ODD model is used to formulate a data collection & evaluation strategy and is used as an input in our statistical tools for analyzing performance.
Autonomous System Safety & Performance Evaluation
Our software tools combine state-of-the-art statistical modeling, probabilistic programming and machine learning techniques to perform an extensive analysis of different AV/ADAS tasks (localization, perception, planning, control).
Requirement Formulation & SOTIF Consulting
Our methodology provides a formal language and a process for guiding companies through the ISO 21448 'SOTIF' framework. We offer help in refining business requirements, analysing performance limitations, and preparing a safety argument.
