Companies working on L2+, L3, and L4 automated vehicles drive thousands of miles collecting Petabytes of data. But only 5-10% of the data collected is useful for further R&D. The question is —which 5-10%? Nemo finds valuable events and traffic scenarios from driving logs.
A manual or semi-automated process of reviewing driving logs can consume up to 30% of data scientists’ time, making it a prolonged and expensive process.
Moreover, without an automated tool to separate valuable events and scenarios from boring miles, companies can end up spending $10s of millions in just data storage costs.
Nemo finds events and traffic scenarios that match your search query from all of your data.
Here's an example query: Give me "all scenarios where a truck made an aggressive cut-in at a highway exit," "pedestrian jaywalking at a roundabout during night-time," "scenarios where a bicyclist is within 20 meters on a downhill road".
Contact us to learn more
Developers can zero-in the right data they need without spending their engineering time manually reviewing the driving logs. The shortlisted scenarios can be exported for algorithm training, product testing, annotation, or simulation.
Ideally, only frequently accessed data should be kept in hot storage since it’s 10x costlier than cold storage. By automating the process of filtering and tiering the data based on which events and scenarios are useful for R&D teams, companies can save up to 85% in just the storage cost.
Provides analytics such as heat maps, time-series plots, and frequency plots, that offer deeper insights to product development and testing teams in terms of scenario coverage, distribution, and biases in training/testing dataset.
Nemo scans the sensor data (coming from Lidars, camera, radars, and CAN bus) to find relevant metadata across objects in the scene, road infrastructure, traffic interactions between different objects, and the vehicle’s driving behavior.
Developers can search/filter for scenarios using any permutation or combination of these metadata tags, including sub-properties such as distance and number associated with the metadata. Imagination is the only limit here.
The platform features micro-services such as search and query, intelligent data tiering, analytics, pre-labeling for annotation, scenario recreation, and simulation generation. This enables the many use cases for ML researchers (who are working on perception, prediction, motion planning), data infrastructure teams, and testing/validation teams.
Interested in learning more about our product or simply exploring ways to manage your growing data? We would love to help. Drop us a note for a private demo.
1189 Barber Ln
Milpitas, California, USA - 95035
© Ridecell Inc., all rights reserved.