The cloud-based platform for developers to get insights on your system.
It assesses, improves and validates your AD/ADAS by using data.
Accelerate your validation
Save engineering time by getting instantaneous access to relevant and challenging events in simulated logs and real-road recordings.
Quantify your safety performance behavior
Get deep insights and understanding of the safety performance of your system with metrics. Map the performance of your perception, path-planning and behaviour stack. Measure the evolution of your system over several software updates.
Scale your validation
Scale your validation to new ODDs by efficiently processing petabytes of data overnight. The highly optimized cloud-based analytics pipeline can process one year of driving data in a few days since it processes object lists approximatively 100x faster than realtime.
Reduce your validation costs
Reduce storage and infrastructure cost by having challenging situations at hand, while less-meaningful data on cold storage. Reduce on-road testing by making better use of recorded data.
4 use-cases:
Identify challenging scenarios
“Our OEM logs all our AV fleet’s real-road perception data. This is close to 400 hours of data recording per day. The Safety Analytics helps my team to identify which are the challenging and complementary scenarios to focus on, so we can transfer those to the re-simulation team.”
– Laura
Measure system performance
“Our company fleet has recorded a lot of real-world data over the years. We re-run many scenarios at every AD system update. the most and how the AD system evolved over several updates.” The Safety Analytics metrics measure where the system failed.”
– Robert
Benchmark perception
false positive and false negative of the obstacle detection. “We have to constantly improved our perception stack, finding Platform provides a ground-truth enabling us to benchmark and The off-line perception running in the Safety Analytics improve our run-time perception.”
– Thomas
Reduce storage and infrastructure costs
“Our company fleet has recorded around 35 petabytes of data across the last years. Maintaining and storing this data Platform made us realize that 57% of this data is meaningless brings a considerable and growing cost. The Safety Analytics for day-to-day analysis. This enabled us to move most of the data to lower cost cold storage groups.”
– Robert
Scenario – based testing
One of the trending approaches for validation of automated systems (L2+ to L4), is to perform scenario- based validation. This means that engineers validate their automated system by (1) collecting many miles of real- road data, (2) identify a set of challenging scenarios by generating new variations of those real- road miles in a virtual environment, and (3) resimulate those challenging scenarios with a new system update and assess the behavior with pass/fail criteria.
At IVEX, our Safety Analytics Platform greatly contributes in finding the challenging scenarios, and we also provide performance metrics that can be used in the pass/fail reporting.