The auto industry is on a path to offer self-driving cars within the next several years. And with all the headlines about the rapidly advancing technology, automakers’ grand plans, and on-the-street pilot programs, it’s starting to sound as though they soon will become a reality. But a lot still needs to happen before you can pick one up at a local dealership.
Indeed, at least three developments must occur for the automakers to make good on their sensor-to-cloud vision for
autonomous vehicles. Specifically, self-driving technology must:
- become several times more reliable on the open road than humans,
- be small enough to be integrated into the car design, out of sight, and
- cost a small fraction of what they do today.
In one way or another, advancing the state-of-the-art in LIDAR laser detection systems will go a long way toward overcoming all three hurdles. Which is exactly what the risk-taking innovators at Analog Devices have set out to do. Through key discoveries from our team and targeted acquisitions – combined with an existing stable of core competencies in critical areas – Analog Devices is poised to get LIDAR to fit in the car. Further, LIDAR systems will enable autonomous machines in a host of other areas, like robots and drone delivery services to cargo ships and combines. And unimaginable 3D gaming and new user interfaces with precision gesture recognition.
Along with RADAR and camera, LIDAR rounds out the key automotive object recognition technologies. The laser detection technology fills a critical gap in the capabilities between the other two, which is why most self-driving pilot programs feature it. Due in large part to the auto industry, in fact, the market for LIDAR sensor systems is forecasted to mushroom to $944.3 million in 2020, more than three times the size of the 2015 market, which stood at $309.8 million (Grand View Research).
When it comes to sensing obstacles and hazards on the road, no one object recognition technology can do it all. With robust algorithms, cameras are unmatched for spotting and characterizing objects. With enough contrast and light, they can detect and feed analytics platforms a wealth of visual detail to fuel better decision-making. But cameras cannot accurately determine the distance to objects and are highly impacted by weather and driving conditions.
RADAR is great for assessing motion and distance to targets. It can detect objects in any light and in any weather. But current systems don’t offer up much detail about the objects they spot. RADAR, in fact, may not be able to distinguish between a tree and a pedestrian standing next to it. That’s an important distinction because unless it’s Treebeard from J. R. R. Tolkien’s Middle-earth fantasy world, there’s no danger the tree will step off the curb. The pedestrian on the other hand....
(ADI’s upcoming 77/79 GHz Automotive RADAR will help distinguish between objects, but more detail will still be necessary.)
Like RADAR, LIDAR works in the dark and in most weather conditions. Unlike camera systems, scanning lasers can offer up a very accurate picture of the size and shape of objects on the road. But LIDAR systems are currently expensive, power-hungry and generate lots of data. And they’re big and unsightly.
Indeed, most self-driving cars on the road are easy to spot because of the clunky-looking LIDAR systems perched atop the roof. Even without that visual cue, their sticker would be easy to spot at a dealership, as well: many automotive LIDAR today costs more than twice the average selling price of cars.
A big part of the problem is that most LIDAR systems were designed for autopilot systems in airplanes, and for experimental projects. The systems must be re-architected for the needs of the automobile industry.
Enter Analog Garage Innovations in LIDAR
This is just the sort of problem that the Analog Garage was designed to tackle. Analog Devices established the corporate incubator program to give entrepreneurs from inside and outside the company the help to explore new technologies, applications and business models for sensor-to-cloud systems.
Analog Devices’ LIDAR effort began with an idea from an inside innovator. The team has been hard at work for several years developing sensor-to-cloud LIDAR systems for the automotive market. The goal was simple: make LIDAR systems so that they are good enough to sense and detect objects when cameras and RADAR can’t – but compact and inexpensive enough to make them practical for commercial automobiles. We currently have working prototypes and working on ways to further enhance performance and reduce size.
Analog Devices recently acquired Vescent Photonics, a company that had developed a solid-state laser-beam steering technology. This acquisition will help Analog Devices shrink the size and cost of LIDAR systems by eliminating bulky mechanical steering while improving performance and reliability.
Analog Devices has pioneered solutions for automotive airbag, stability, and rollover systems for more than 20 years since the company introduced the first-ever commercial MEMS motion sensors. Accordingly, ADI MEMS sensors are found in one-quarter of vehicle airbag systems today. A 15-year history in 24GHz and 77GHz Automotive Radar has led to ADI technology in half of Automotive Radar modules in recent development. As well, Analog Devices has key technologies that make it uniquely qualified to develop and manufacture difficult-to-make LIDAR components in high volume.
LIDAR is a key pillar of Analog Devices’ Drive360 autonomous driving solutions. Safe, autonomous driving relies on technologies which continuously provide redundant and reliable information. Drive360 LIDAR, RADAR and Inertial MEMS will work harmoniously and form a 360-degree safety shield around the car which provides multiple data points for vehicle positioning and object recognition. In this way, autonomous vehicles will drastically reduce traffic accidents in all cases and weather scenarios.
Analog Devices is the world leader in high-speed amplifiers and data acquisition – a key technology required for LIDAR in addition to digital signal processing vital to piecing the torrent of data produced by the LIDAR system into an intelligible 3d view of the world around the car. Exactly the space where Analog Devices has been innovating for more than 50 years.