PRODUCT & SERVICES
Hologram complements traditional simulation and road testing of perception systems. It intelligently tests perception software against adversarial examples, which it generates from your own sensor data. Hologram identifies risks that are difficult to find with other types of testing and analysis. This shows how your perception software reacts in realistic “what if” scenarios.
Use Hologram to construct strong safety cases, comply with standards, and maximize investments in on-road testing. It helps you to coordinate key roles across autonomy development: designing and training machine learning algorithms, managing fleet deployments, and performing system-safety programs.
“a heavy tail distribution of surprises from the real world could make it impossible to use a simplistic drive/fail/fix development process to achieve acceptable safety”
MISSA’s data-driven safety methodology builds safety cases more quickly and easily than traditional, checklist-based methods. The implication of design changes on the safety case can be immediately communicated to developers so that intelligent trade-offs can be made.
A common frustration with traditional safety processes is that safety cases become outdated by the time they are finished. This is especially true in complex, fast-moving fields like autonomous vehicle development. MISSA’s iterative nature means that new features with safety- or mission-critical implications can be easily incorporated by refining the existing model. There is no need to go through the time and expense of re-creating manual fault trees as behaviors and identified risks evolve.
If you have already worked through a preliminary hazard analysis, ECR can take you through the functional hazard analysis, the verification plan, and beyond. While every system is different, we estimate that a safety program using MISSA is 5-6x faster than a traditional safety process. Rather than being a box-checking solution to standards compliance, MISSA’s output actually helps developers build safer systems on time and on budget.
Our decades of experience show that a focused safety concept can save significant development
expense while cutting risk. Using MISSA, our services team helps customers iteratively construct a safety concept that’s based on a clear accounting of deployment hazards and the safety requirements necessary to avoid them.
History shows that most of the cost of developing safety-critical systems is spent on verification and validation. MISSA helps to optimize your system architecture to isolate safety-criticality in a way that meets your safety concept using minimal technology. MISSA’s model-based tools for assessing functional hazards help you to quickly evaluate design trade-offs from a risk perspective.
Using MISSA, our services team helps you integrate your system’s safety requirements with our Hologram and Switchboard tools to perform robustness testing. We also help customers plan and conduct standards-based verification activities called for in MIL-STD-882E, ISO 26262, and other standards.
The goal of any safety program is building a better, safer product. Unfortunately, many safety programs are structured as “boxes to check” that don’t add inherent value to the product itself. As a result, “safety” is siloed from “development.” MISSA bridges this gap to achieve continuous integration for your safety case. This helps you to improve both product performance and product safety.
The Switchboard stress test engine performs random, intelligent fuzzing that complies with your system’s defined interfaces. Switchboard’s novel process accelerates your ability to find and fix realistic edge case bugs before products are launched. Switchboard intelligent automation empowers development teams and eliminates concerns of testing bias. Developers simply define undesirable outcomes in Switchboard and run tests to isolate the circumstances that surround failures.
Connect with Us
Michael Wagner, CEO, Statement on the Publication of UL 4600, The Standard for Safety for the Evaluation of Autonomous Products
As UL 4600 is now officially published, I would like to take this opportunity to provide perspective on why we felt it was essential to lead this effort, as not many startups invest time in writing standards. Edge Case is not your typical startup; we are on a mission...
What is Safety? When will AI be good enough to remove the safety drivers from autonomous vehicles? Who decides, and how do they do it? From U.S. Navy subs to Carnegie Mellon, the National Robotics Engineering Center and now Edge Case Research, Dr. Phillip Koopman has...