“By working with Applied Intuition to provide high-fidelity sensor simulation to our customers, we have greatly simplified the sensor integration process and ultimately accelerated a customer’s time to autonomy.”
Mark Frichtl CTO at Ouster
The sensor suite performance of advanced driver assistance systems (ADAS) and autonomous vehicles (AVs) depends on many variables. This makes selecting sensor hardware, optimizing its mounting position and bracket, and testing against requirements difficult.
There are hundreds of sensor options and vendors on the market
Creating test rigs and setting up real-world tests is time and cost-intensive
Each change to a vehicle’s chassis, operational design domain (ODD), or capabilities impacts the optimal configuration
It is impossible to deterministically test against faults, adverse conditions, and edge cases for every potential sensor suite
Applied’s physically accurate, real-time sensor simulation solution models real-world sensors, enables teams to test thousands of potential configurations, and analyzes performance to identify sensor suites that optimize performance while minimizing costs.
Select validated, hardware-specific sensors from Applied’s pre-configured library, or flexibly model new sensors
Build hypotheses by visualizing sensor and sensor suite field of view coverage, overlap, and blindspots in 3D
Run high-fidelity simulations to quantitatively evaluate and compare sensor suite performance
Scale simulations with an optimization toolkit to explore new promising options efficiently while avoiding excess computation
Design and validate an optimal sensor suite
Define sensor suite requirements
Define requirements based on upcoming vehicle features and perception needs. Tie requirements to metrics or concrete test cases.
Experiment with candidate sensors and packaging locations
Model your vehicle and experiment with different sensors,sensor mounting positions, and brackets. Visualize fields of view coverage, overlap, and blindspots to intuitively understand performance and make design tradeoffs.
Execute physically accurate sensor simulations to evaluate thousands of sensor suites across a deterministic battery of static and dynamic tests. Efficiently explore the solution space with algorithms that automatically suggest promising sensor suites.