Compact And Cheap Lidar Could Steer Small Autonomous Vehicles
A close-up look at a miniature optical beam steering component that promises to enable lighter and less costly autonomous vehicles. (Photo: Kristinn Glyfason)
For autonomous vehicles, lidar is an essential technology to
recognize and detect surrounding objects. Researchers at KTH have taken aim at
the key component of lidar, optical beam-stearing, and developed a device that
is significantly cheaper to manufacture, lighter and more resource-efficient
than previous variations of the technology. This advance could pave the way for
smaller autonomous craft such as drones and robots and help enable better profitability
in the vehicle industry.
Current lidar (light detection and ranging) technology costs
thousands of dollars, weighs around one kilo and consumes several watts of
power.
A new version of lidar being developed at KTH Royal
Institute of Technology in Stockholm is in another league altogether.
Carlos Errando-Herranz, a postdoc in the Division of Micro
and Nanosystems, says that this version of lidar would cost about USD 10 each
given large production volumes, weigh a few grams (including peripheral
equipment) and consume some 100 milliwatts. The research was reported in the
journal Optics Letters.
Errando-Herranz says that the miniaturized beam-stearing
device measures approximately 100 micrometres, and is best observed under a
microscope.
“We use the same production techniques as for manufacturing
accelerometers and gyroscopes for smartphones,” he says. “This means the costs
can be really low on large volumes.”
Errando-Herranz says that the technology can enable more
craft, such as robots or drones, to be self-flying or self-driving for example.
The advance could also eliminate the need for remote control
on drones that are designed to deliver emergency medical equipment to remote
places, such as defibrillators, says Kristinn B. Gylfason, Associate Professor
at KTH.
“Robots and drones are absolutely possible application
areas,” Gylfason says. “Current lidar systems are also too expensive for
self-driving cars. The vehicle industry is very cost sensitive. Other
possibilities are facial recognition for smartphones, such as Apple's Face ID.”
The difference with the KTH approach to lidar is that it uses
micro-electromechanical optical beam steering.
“A traditional lidar is based on mounting an array of lasers
on a rotating tower, like the Velodyne puck,” Gylfason says. “Our approach to
lidar is based on integrated micro-opto-mechanics, where we have built a
tuneable grating into the surface of a silicon chip. By modifying the grating
period, we decide in which direction the beam should sweep.”
Optical beam steering can also be used for three dimensional
imaging in medical diagnostics, with a technique known as Optical Coherence
Tomography. With this miniaturized technology, a scanner could be inserted into
the body during keyhole surgery and used to identify changes in tissues.
