Why do you need a safety case to test an AV?
A safety case is a structured argument supported by body of evidence that demonstrates all the safety risks have been identified and appropriate controls have been put in place to minimise the risk of harm.
It is fast becoming less appropriate to consider vehicle safety by breaking it down into discreet sub-systems and a holistic approach is now commonly required. Safety cases have been used for decades in other safety critical industries as best practice for assuring system safety to stakeholders, including regulators.
TRL have technically authored BS PAS 1881, which specifies the requirements for safety case for automated vehicle trials that reflects good practice. This specification will evolve and be updated over time as the industry and technology matures. Being able to interpret and implement the requirements of this specification in a safety case will give trial organisations an advantage on their path to commercial exploitation, assure relevant stakeholders and ensure safety on the UK road network is maintained. The BSI PAS 1881 is available to download for free here.
For examples of a Safety Case created by TRL, here are abridged versions produced for the general public who were participants in two automated vehicle trials:
- Streetwise Trials Abridged Safety Case
- SMLL Abridged Safety Case for Shared Research Programme Trials
The anatomy of a safety case1. Safety case development
A safety case has to demonstrate that all the risks associated with the vehicle trial have been identified and appropriately mitigated. The safety case also demonstrates compliance with all relevant standards, guidance and legislation.
TRL has created a framework approach which includes the system as a whole – vehicle, control and infrastructure. Thanks to our experience of CAV trials to date, we have a comprehensive understanding of the many complex issues and multifaceted expertise involved in managing safety of trials. The TRL framework covers
- Operational design domain and test objectives
- Risk assessment, evaluation and mitigations
- Operational guidance and training
- Route selection and assessment
- Reporting, monitoring and continuous improvement
- Compliance and insurance
- Stakeholder consultation and permissions
2. Risk assessment
TRL has a strong background in conducting risk analysis for a number of transport related applications including road worker safety, new driver safety, HGV training, pedestrian routes and cycle highways. TRL has worked on the earliest CAV trials in the UK and is currently the lead safety advisor on the most high-profile CAV trials. As a result of our experience in this field we have developed a proven framework for undertaking a comprehensive risk assessment, which is central to the safety case.
3. Risk mitigation
A successful trial is always one conducted without incident or unexpected outcomes.
TRL has been designing and conducting trials on public roads for decades, and we absolutely understand the dynamic between the need for data and the safety of everyone involved.
As a result of our experience, we have a comprehensive understanding of the mitigation measures which are appropriate, proportionate and can quickly make suggestions to improve the safety of a trial without compromising your research objectives.
4. Independent Safety Assurance
Just as the best engineering follows the principle of “measure twice, cut once”, it is expected that a safety case will be reviewed by the all project stakeholders. Where TRL is not the originator of a safety case for a trial, or when stakeholders do not have the expertise in house to interpret the information in a CAV trial safety case, TRL provides independent safety assurance of that safety case. Through the application of carefully constructed evaluation criteria, TRL reviews safety cases, identifies any gaps and provides advice on bridging those gaps. This offers an additional layer of diligence and reassurance for stakeholders and for insurers in particular.
CredentialsTRL have been involved in issues of safety in transport for over 80 years, in many guises. For many decades TRL led all the research into the material design of vehicles which has ultimately led to Global NCAP standards for all new vehicles. TRL wrote the guidance for roadworker safety and traffic officer safety, and for workplace transport. The driver hazard perception test was designed and introduced by TRL to improve the performance of young and novice drivers. We wrote the textbooks on designing streets to make them safe for all road users. We continue to lead research and advise on how to design the safe flow of traffic and maintenance vehicles through roadworks on the SRN. On behalf of the Dept for Transport we collect and analyse vehicle collision data and all fatal collision data to feed back into recommendations for safety improvements of vehicle design, driver assessment, and road design.
TRL have been involved in most of the critical research projects into ADAS and CAV trials since 2010 and have built up the most comprehensive understanding of the many complex issues.
We have advised and supported the following organisations in the development of standards, policy and guidance documents:
Dept for Transport - Code of Practice for CAV Trials
Transport for London - guidance on AV trials in London
Zenzic - Safety Framework for common development of safety cases at all the CAM TestbedUK trial sites
Here is a summary of all current relevant regulations and guidance on conducting safe automated vehicle technology trials.
Listen here to a discussion between Camilla Fowler and David Hynd about the issues which mean that concerns about safety are dictating the pace of development and adoption of automated vehicle technologies in the UK, and who decides: How safe is safe enough?
Chris Lodge, Safety Consultant at TRL, explains on the video below how safety assurance fits into the development and deployment of connected & automated vehicle technologies, and why public confidence is so vital to the successful mass adoption of these new mobility options.