Applied Intuition has partnered with Ouster to create, test, and release synthetic models of Ouster lidars. These validated lidar models empower Ouster’s and Applied’s customers to generate synthetic lidar data that accelerates the deployment of lidar-based machine perception systems (Figure 1).
Ouster develops high-performance digital lidar solutions for automotive, industrial, robotics, and smart infrastructure applications, with offerings including both spinning and solid-state lidars. Over 600 customers across 50 different countries have chosen Ouster for the superior size, weight, form factor, power efficiency, and durability its digital lidars offer over legacy analog lidar sensors.
System integration can present a barrier to adoption of lidar-based machine perception systems. Challenges like physically determining the optimal mounting location, which can require laborious testing and engineering effort, slow down system development. Further, machine learning (ML) models are often subject to overfitting, with characteristics tailored to a specific make and model of sensor. Teams might be unable to unlock the performance of an Ouster lidar unless they collect additional training data and adapt their ML models to this new type of sensor.
Applied’s sensor simulation software enables customers to accelerate the deployment of new Ouster lidars through faster iteration cycles and rapid sample data generation:
“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,” said Mark Frichtl, CTO at Ouster.
“We are excited to partner with Ouster to accelerate the deployment of its lidar models,” said Peter Ludwig, CTO and Co-Founder of Applied Intuition. “Through this partnership, we are developing state-of-the-art lidar models that faithfully represent Ouster hardware, ensuring that our customers can generate impactful synthetic data and trust the results when developing and testing in simulation.”
Contact Applied to learn how your team can use sensor simulation and synthetic lidar data to train and evaluate ML-based perception systems.