WaveSense Uses GPR for Autonomous Vehicle Location

WaveSense uses ground penetrating radar – GPR to create a reference map of the subsurface conditions below the pavement of a highway. This can be accomplished because the return GPR signal creates a unique electromagnetic signature.

From an article in Motortrend.

image of WaveSense Uses GPR to Develop Unique Highway Subsurface highway and GPR Signal

WaveSense Uses GPR to Develop Unique Highway Subsurface Signature

Waymo, which recently launched its autonomous ride-hailing service in the Phoenix area, is betting big on traditional sensing technologies—including cameras, GPS, radar, and those super-expensive spinning lidar sensors on the roof.

These Waymo cars are essentially beta-testers, which means they are not fail safe. (Remember Uber?) If Waymo happens to have an accident and the fault is found to be with the sensor array, it’s a safe bet Waymo won’t be able to blame a blizzard or pea-soup fog.

An MIT spinoff startup called WaveSense hopes to hasten the arrival of autonomy in regions with, you know, weather, by providing an additional “leg” of the environmental perception stool that is utterly weatherproof: ground-penetrating radar (GPR). There are numerous applications for this technology in use today. Law enforcement locates buried booty or bodies with it, road commissions use it to assess road-bed fitness, it helps utilities locate pipes, and archaeologists rely on it to find the next King Tut’s tomb. Although most of these applications use a much lower-frequency radar than the forward-looking automotive kind—1-3 gigahertz (billion cycles per second) versus 77 GHz—WaveSense uses a frequency that’s way lower still—100-400 megahertz (million cycles per second).

The higher-frequency GPR provides super-high resolution but can’t measure as deep and suffers from “blur” at higher vehicle speeds. It’s also more susceptible to things like trash on a roadway, “thermal drift” as temperatures change, and the inevitable variation in the height of the sensor off the road that comes with vehicle pitch, roll, and payload variation. A 100-400-MHz system avoids these problems and can detect, record, and analyze underground features buried 6-10 feet deep. It also requires just 40 microwatts total, of which only 4 “leak” into the surrounding air. Higher-frequency GPR consume 1,000 times as much power. This radar senses differences in the electromagnetic properties of objects such as pipes, roots, and rocks in the surrounding dirt—all of which tend to be extremely stable over time.

As with camera- and lidar-based positional sensing, the road network must first be mapped by vehicles using essentially the same hardware, correlating GPR imagery with GPS location data. It takes a few passes to ensure the 5-foot-wide beam generates full coverage of a lane. The raw map data can be used for location-correlation, providing lateral/longitudinal accuracy within about an inch, and it works at highway speeds. Sensors that can “see” through 10 feet of dirt are unfazed by a layer of grunge on the lens, and at scale the sensor should cost $100 per vehicle. Another bonus: They mount underneath, so they don’t mar the vehicle’s styling.

For more information click here.

For more information on WaveSense 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 3D Modeling, AI, artificial intelligence, Autonomous vehicles, Government, Ground penetrating radar, Infrastructure, intelligent cities, Mapping, military, remote sensing, Research, Sensors, Technology 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.