Light-Detection-And-Ranging, or LiDAR for short, is any sensor that uses light to determine an object’s distance. When it comes to building LiDAR for autonomous vehicles – whether it’s for the backcountry, the city, mine sites, or in space – performance for LiDAR still comes down to the three Rs.


The system must detect objects of low reflectivity at distances ranging from directly in front of the sensor to hundreds of meters down the road.


The system must produce high-resolution, 3D characterization of objects in a scene without significant artifacts or noise. Because far-away objects appear smaller, high-resolution is needed to classify objects at range.


The system must be able to withstand rugged conditions, endure constant use, and be made of cost-effective and readily available components.

The search for rugged high-performance LiDAR has challenged the autonomous driving industry for years. As it stands, the most defined demands come from the trucking, industrial, and mining industries which have fixed-routes, small maps, require slower speed but these vehicles operate over long time periods and in tough conditions.

Baraja’s Spectrum-Scan™ LiDAR is built using mature wavelength-tuning technology. It leverages proven off-the-shelf telecom components and allows for a flexible, software-defined LiDAR that exploits the wavelength property of light for opto-electrical beamsteering.

I. The beginning

Baraja Spectrum-Scan™ emerged from the minds of Federico Collarte and Cibby Pulikkaseril in 2015. The two Australian founders observed how the telecommunications industry had leveraged the physics of light to develop a fiber-optic laser to facilitate the communications-wave during the 1990s. However, the light of the telecom laser was “trapped” into a fiber optic, so the light-bulb moment came as they imagined a new application where the light is set free from the fiber and shot through free-space prism-like optics as a steering mechanism to perform LiDAR measurements.

From a Sydney garage, the two began experimenting with sending lasers through prisms to see whether they could develop a LiDAR system optimized for autonomous vehicles and perception. The objective was to design a system which could produce an exceptionally high-resolution pointcloud, with no moving parts that covered a wide field of view.

Founders Federico Collarte & Cibby Pulikkaseril in Baraja’s first Sydney HQ (Federico’s garage)
Founders Federico Collarte & Cibby Pulikkaseril in Baraja’s first Sydney HQ (Federico’s garage)

II. History of LiDAR


Light detection and ranging (LiDAR) is one of the primary sensors being used in autonomous vehicle development. At its core, LiDAR provides 3D pointclouds of the environment, where each point is a true measurement of distance and position

Early LiDAR provided simple low-resolution distance data, allowing vehicles to range and classify objects with greater accuracy and speed than other sensors alone. Deploying LiDAR to automate a vehicle was a breakthrough and, compared with cameras producing image data, LiDAR worked at any time of day thanks to its active sensing.

But while it was clever, early LiDAR had limited range, low-resolution and poor reliability.

Because autonomous vehicles need to see all around the vehicle, a LiDAR sensor must tackle the problem of moving a laser beam, which points in a single direction, across a field of view. Rotating the lasers mechanically was the obvious solution: engineers could strap a collection of lasers to a turntable and have them spin around like a radar on a battleship. Unfortunately, this means rotating active electronics and optics, which are often large, heavy, power-hungry, and delicate.

Other approaches are a bit smarter; they aim a stationary laser at a mirror, then move the mirror quickly back and forth, scanning the laser across the field of view. These approaches either use big mirrors (often moved by devices called galvanometers or ‘galvos’) or small mirrors (called Micro-Electro-Mechanical Systems or MEMS).

But this often ran into reliability problems. Moving parts were likely to burn out with constant use, particularly in gruelling automotive and industrial environments. Sensors must last for years of operation while surviving temperature fluctuations, vibration, and g-forces which ultimately kill all but the most rugged electronics.

Additionally, early sensors lacked resolution. Controlling mechanical steering mechanisms precisely enough for high-resolution at long range is a significant challenge. Every small error in beamsteering angle is compounded over distance, placing stringent requirements around measurement and control accuracy that are tough to meet.

And lastly, early LiDAR also had a problem with range. Because LiDAR operates by measuring returned light reflected back by an object, “brighter” targets that reflect more light are easier to detect. Conversely, “darker” objects – such as tires, black cars, and even asphalt – are much harder to detect since they reflect less light back to the sensor. As an object gets further from the sensor, less light makes it all the way out and back, making dark, far-away objects very hard to detect. So, while many LiDAR sensors can see bright targets far away, few maintain this range for dark targets. Because an autonomous vehicle must be able to avoid all obstacles of any brightness, this limitation presented a real problem for perception software that could not confidently see a clear path far enough ahead.

Fortunately, the telecommunications industry was facing the same fundamental problem: they needed to move light over long distances, without breaking parts, and maintain a clear signal.

Rather than reinvent the wheel, Baraja founders Federico Collarte and Cibby Pulikkaseril decided to learn the lessons of the telecommunications industry.

III. Lessons from the Telecom Industry (use the wavelength of light )

As communication around the world increased, telecoms were looking into high-speed optics to facilitate the onslaught of real-time, simultaneous messages. They needed to get multiple data paths through a single cable in the same way LiDAR needs to sample multiple points in space.

Initially, telecom fiber optics started with mechanical switching methods, like galvanometers that used bigger mirrors, on either end of the cable. Just like LiDAR did.

Then, the industry miniaturized this to small mirrors, like the Micro-Electro-Mechanical-Systems or MEMS. Just like LiDAR did.

But these mechanical methods were slow, unreliable, and expensive, so telecommunications researchers began using wavelength-tunable lasers to transmit information to different devices based on different wavelengths of light.

Dense wavelength division multiplexing (DWDM) was born, and enabled large numbers of signals to transmit down a single fiber without slowing down and creating bottlenecks. The signals were fast, clear and brought telecom costs down.

The result was explosive, and during the 1990s the telecommunications industry experienced profound growth thanks to widely tunable semiconductor lasers.

Federico and Cibby experienced how the telecom industry moved from big, clunky mechanical systems to wavelength technology. They watched how this successful innovation meant the manufacturing and installation costs came down, and the off-the-shelf parts used were readily available anywhere in the world.

Using this mature, tested and globally available wavelength technology, the two engineers applied it to the LiDAR problem and began building Baraja Spectrum-Scan™.

We grew out of the garage and into our new Sydney based HQ at CSIRO
We grew out of the garage and into our new Sydney based HQ at CSIRO

IV. Introducing Baraja Spectrum-Scan™

Federico and Cibby saw the opportunity to bring these two innovations together: pairing a wavelength-tunable laser with prism-like optics.

The result is beamsteering on one axis without any mechanical motion.

Meet SpectrumHD
Meet SpectrumHD

Imagine the way a prism emits different wavelengths of visible light, or colors, at different angles. Baraja Spectrum-Scan™ steers infrared light through a prism in the same way, creating a spectrum spread vertically over the field of view as it changes wavelengths. Pairing these optics with a single telecom-grade laser source allows for high-power coaxial detection, enabling unprecedented and far-reaching range.

Low-divergence optics, plus the unmatched beam quality emanating from single-mode fiber-coupled laser, combined with fine wavelength tuning also result in thousands of lines of high-quality resolution. In fact, this resolution is only limited by the number of discrete wavelengths of light your laser can produce and detector can discern. Baraja Spectrum-Scan™ is capable of thousands of discrete wavelengths, resulting in unmatched resolution capability.

Baraja Spectrum-Scan™ uses a telecommunications-grade laser with a passive glass element that’s ready for all kinds of extreme conditions and won’t burn out through heavy use. Unlike the mechanical approach of spinning galvos, or MEMS, this kind of static reliability means the system can operate without fear of overheating, breaking or falling out of alignment.

The end goal for the LiDAR industry has always been a robust high-performance sensor. Many of the current commercial pilots are found in highway trucking, construction equipment and mining which have high requirements for vibration, shock and vehicle longevity in rough conditions.

Baraja Spectrum-Scan™ leverages the fundamental physics of light. Rather than introduce experimental science. The team uses mature technology, proven through the success of the telecommunications industry, to build a low-loss, inertia-less, and ultimately software-configurable beamsteering mechanism.

Baraja Spectrum-Scan™ provides unprecedented pointcloud range and extremely high-resolution in a reliable and robust package.

Baraja Spectrum-Scan™ enables extremely high-resolution pointcloud data

To look at how the Baraja Spectrum-Scan™ is built, and how its components elegantly address Range, Resolution and Reliability, join the Baraja Inner Circle to stay up to date with the future of autonomous driving.