.svg)
Safety is paramount for all of us here at Tensor. As we innovate forward and progress towards a future where self-driving cars become an integral part of our daily lives, we bear a profound responsibility to ensure the safety of all in and around our vehicle including occupants of vehicles, and all others including vulnerable road users.
Meet the Tensor’s Safety Case. We share our approach to achieving safe autonomous vehicle testing and deployment and we present a brief description of our Safety Case Framework which is a comprehensive, systematic, and evidence-based blueprint that underlines our dedication to safety. The Tensor Safety Case unequivocally demonstrates our meticulous effort to address every critical development step, and prove that our vehicles are acceptably safe for public roads.
The safety case serves as the cornerstone of our entire vehicle life-cycle of planning, development, testing and validation, manufacturing, deployment, field monitoring, maintenance, and continuous improvement until the vehicle decommission. It is our pledge to the public that safety is not merely a priority but a fundamental principle ingrained in every facet of our operations.
Safety case is a structured argument, supported by a body of evidence, that provides a compelling, comprehensible and valid case that a system is safe for a given application in a given environment.
Tensor’s Safety Case Framework adopts the C-A-E notation, which provides a clear, logical, and easily maintainable format for presenting our safety case. The claims are statements about the safety of the system or its components, the arguments are the logical reasoning that connects the claim to the evidence, and the evidence consists of objective, verifiable artifacts that support the claims and arguments.
We decompose the safety case framework to establish sub-claims and sub-arguments, until the claims and arguments are clearly interpreted, and the corresponding evidence can be provided.
Throughout the framework, we utilize the Safety Performance Indicators (SPIs) to support our safety claims and arguments. SPIs provide measurable data that show how our system is performing in terms of safety. The data collected from these indicators forms a critical part of the evidence base for our safety assertions and helps to validate the effectiveness of our risk mitigation strategies.
This safety case framework forms a foundation for a safety case that supports our confidence in the AV’s ability to operate safely, responsibly, and predictably within its defined ODD. We remain committed to transparency and continuous improvement as part of our mission to earn and maintain public trust in autonomous vehicle technology.
As we develop and deploy autonomous vehicle technology, it is crucial to establish clear and comprehensive conformance targets that align with regulatory standards and requirements; industry standards, guidelines and best practices; and our own rigorous safety requirements and objectives.
Some of the conformance targets are listed below:
All vehicle safety regulations applicable in the markets of deployment
Standards:
· UL 4600: 2023
· Functional Safety: ISO 26262:2018
· SOTIF: ISO 21448:2022
· Cyber Security: UN R155, UN R156, ISO 21434:2021
· ODD definition: ISO 34503:2023
· Fault Management (MRM): ISO/DIS 23793
· Test scenarios: ISO 34504:2024, ISO 34502:2022
· Road Testing: SAE J3018
· Safety Arguments: ISO 15026-2:2022
Guidelines and Best Practices:
· NHTSA AV Guidelines
· NHTSA AV STEP
· NHTSA Cybersecurity Best Practices
· NIST Cybersecurity Framework
· AVSC Best Practices
In this Safety Case Framework, we will provide some examples of the detailed claims to demonstrate how our autonomous vehicle meets or exceeds these conformance targets.
The Tensor’s Safety Case is established towards our top safety claim that:
Claim: Tensor Autonomous Vehicle (AV) is absent of unreasonable risk in its mission capability and related operational design domain (ODD).
We chose this top claim to address the following aspects:
The claim that an AV is free from unreasonable risk in its mission capability and related operational design domain (ODD) is an appropriate target for autonomy safety. This claim directly addresses the core concerns of safety and capability in AV development.
The safety case will identify potential risks and set clear boundaries for acceptable risk levels with regard to regulations, standards, and industrial best practices to ensure that the AV operates within defined parameters and manages risks effectively.
The mission capability refers to all of the Dynamic Driving Tasks (DDT) that our vehicle is capable of executing, such as acceleration, braking, and steering. Meanwhile, our ODD specifies the operational scenery, environmental conditions, dynamic elements and operation status.
This systematic approach facilitates the identification and mitigation of potential hazards, ensuring that the AV does not expose road users to unreasonable risk. Consequently, the safety case aligns with fundamental risk management and safety engineering principles, providing a robust framework for safe AV operation.
We decompose the top claim to subsequent levels of claims to demonstrate how we build our safety case, and ensure that high-level safety goals are systematically broken down into manageable and verifiable subclaims. This is achieved by using the logical argument methods including decomposition, substitution, evidence incorporation, concretion and calculation of proof, to be further augmented and supported by appropriate evidence.
AV: Autonomous Vehicle
AVSC: Automated Vehicles Safety Consortium
AV STEP: ADS-equipped Vehicle Safety, Transparency, and Evaluation Program of NHTSA
C-A-E: Claim-Argument-Evidence
DDT: Dynamic Driving task
ISO: International Organization for Standardization
NHTSA: National Highway Traffic Safety Administration
NIST: National Institute of Standards and Technology
ODD: Operational Design Domain
SAE: Society of Automotive Engineers
SPI: Safety Performance Indicator
SOTIF: Safety of the Intended Functionality
UL: The Underwriter's Laboratory
Latest news, announcements and
updates direct to your inbox.
.webp)
