SPIE Proposes Automotive Lidar BenchmarkingTests

Photo of SPIE to Benchmark Automotive Lidars

SPIE to Benchmark Automotive Lidars

Working with experts like you, SPIE is helping to establish standard tests to benchmark automotive LiDAR performance. Please comment on this draft in the text field below. Or download the PDF and send comments via email to LidarTest@spie.org so that Paul McManamon and his team can revise, update, and improve the LiDAR test plan.

Purpose: To develop standard benchmarking tests for auto lidar performance, and to employ these tests on available auto lidars.

Procedure:
These tests will occur on Monday and Tuesday of the SPIE Defense and Commercial Sensing (DCS) Symposium. We have allocated 4 hours for set up and expect to test 2 lidars per hour. We are planning to test a lidar for 15 minutes per lane. We can test up to 2 lidars at the same time, so if we keep to this schedule we could test 4 lidars per hour. This is aggressive. It will be one of the things we learn this year.

We have allocated 2.5 hours for tear down of the test hardware. The facility will be available from 8 AM to 6 PM on both days, with round-the-clock security. This should allow at least 20 lidars to be tested. If more than 23 vendors sign up, this plan will be revisited. All Lidars will be tested using the identical setup. We have planned to gather comments at panel discussions on Wednesday in the exhibit hall. While the full analysis will not be complete by then, we expect to provide some preliminary data during those meetings. An archival publication will be submitted for publication within a few months after the test. The analysis will be performed by the test panel, with companies having no access until the results are published.

We will test a single lidar of each model, even though there could be variation form one unit to another of the same model lidar. Companies can submit different models for testing. For example, it would be good to test both the Velodyne HDL64 and HDL128. While it would be useful to test several lidars of the same model in order to examine variations, this would be an unreasonable burden on both the companies and testing team. Therefore, for this initial test, we are asking companies to provide the best lidar of a given model for the tests. This is an item that will be discussed during the panel discussions.

Background: Autonomous vehicles were given their start by the DARPA Grand Challenges, however, recent work has been dominated by commercial interests. The DOD is still very interested in autonomous vehicles, particularly for reducing casualties from IEDs. The hope is that autonomous vehicles will reduce risks to life for transit through dangerous places. While this effort is focused on auto lidars, these devices should be useful for other forms of autonomous vehicles that could be used by the military or commercial interests. DCS is an ideal conference for this topic.

Schedule:
Testing scheduled for April 27 and 28 at DCS in Anaheim. While we hope we could test up to 4 lidars per hour, for planning purposes we will plan on 2 per hour. We hope this is conservative. This is a first time, so we need to be conservative. Below is a tentative schedule.

For the complete document click here.

Note – If you liked this post click here to stay informed of all of the 3D laser scanning, geomatics, UAS, autonomous vehicle, Lidar News and more. If you have an informative 3D video that you would like us to promote, please forward to editor@lidarnews.com and if you would like to join the Younger Geospatial Professional movement click here.

This entry was posted in artifical perception, artificial intelligence, Autonomous vehicles, Certifications, computer vision, Conferences, driverless vehicles, Government, Hardware, Laser Scanning, Lidar, Orgs, Research, Sensors, Standards, Technology, The Industry, Uncategorized, Young Geospatial Professional and tagged , , , , , . Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.