Laser range-finding, or LIDAR, is a way of measuring distance, directly analogous to radar except using radiation in the near infrared range of the electromagnetic spectrum rather than the radio waves used in radar. LIDAR techniques are used for a variety of final applications including terrain-mapping for geology, urban planning and archaeology; distance measurement for surveying, golf, hunting, military applications, and docking of large ships; and speed measurement for traffic monitoring and speed limit enforcement.
Semiconductor laser diodes, along with silicon and InGaAs PIN photodiodes and Avalanche Photodiodes (APDs), have served these applications well in traditional, hermetic metal “transistor outline” package formats known as TO-cans since the 1970s, and continue to do so for applications where the environment is challenging and/or the ultimate in performance is needed. Lasers and APDs are used in space and on aircraft for extreme long-range applications. More recently, range-finding has moved closer to the consumer where production in high volume at low cost becomes a key driver, while keeping performance loss to a minimum.
To support these new applications, technologies have been developed that allow component packaging to go from labor-intensive and complex hermetically-sealed TO-cans toward packages based on a ceramic or printed circuit board (pcb) with epoxy over-moulding. These provide the protection necessary to ensure high levels of performance and reasonably long lifetimes.
The most recent laser and APD designs feature low-cost packaging techniques suitable for surface mounting using solder reflow, removing the need for the LIDAR module supplier to manually solder leads into holes on a circuit board.
Autonomous Vehicles – No Longer Fiction
The latest application for this technology is developing from automotive driver-aiding safety equipment and, fusing with cameras, radar and ultrasonic systems, is allowing manufacturers-both traditional and disruptive-to work toward fully autonomous vehicles.
Over the last few years, the level of available safety equipment and driver aids has risen at an incredible rate. All new cars have air bags, pre-tensioning inertia reel seatbelts, passenger safety cages, anti-lock brakes, and many more features. And even mainstream cars can use sensors and actuators to detect parking spaces and then park the car with no steering input needed from the driver.
Radar has been used on cars to provide automatic cruise control and collision avoidance, while cameras and image-processing software help keep cars in their lane and inform the driver of speed limit signs. However, the recent, tragic death of a Tesla driver using a semi-automated driving system, based on radar and cameras, that was unable to detect a large truck crossing its path shows that a much greater level of situational awareness is needed before cars need no human action to get safely from point A to point B.
LIDAR Technology is Gaining Importance
This is why LIDAR is now of such great interest to both traditional car manufacturers and high-tech companies. Systems are already in use that detect objects in front of the car to provide more intuitive and capable reactive cruise control and collision avoidance. While they are more precise and capable of detecting smaller obstacles than radar-based systems, these early generation systems are based on small arrays of discrete edge-emitting 905nm lasers and silicon APDs, and are limited in field of view and range. To achieve fully autonomous vehicles, some high-tech companies based outside of the traditional automotive world have looked at different ways to incorporate LIDAR into their vehicles.
Google has had cars driving autonomously for millions of miles, using scanning LIDAR systems to map the area around and ahead of the vehicle to allow the on-board computer systems to use the data gathered and compare it with location-based data to work out where the car is and what is happening around it, and to react accordingly. These scanning LIDAR systems were based on traditional 905nm edge-emitting lasers and silicon APDs. Other new companies are focusing on their core competencies in LIDAR and software, and partnering with traditional automotive suppliers to bring similar scanning systems to vehicle manufacturers. Additionally, traditional Tier-1 automotive suppliers are working to have this technology under their own in-house control. Tens of thousands of surface-mount lasers and APDs have been used to support these activities.
Some Challenges Still Lie Ahead
The real challenges now facing developers and integrators of LIDAR systems are field-of-view, range, reliability, and aesthetics.
Field-of-view is being addressed by the use of laser and detector arrays. Companies are working closely with key customers on the development of compact, individually addressable laser arrays capable of delivering pulses of a few nanoseconds’ duration at a repetition rate of more than 100kHz. Customized versions with laser driver circuitry within the package have already been delivered, and a stand-alone version was launched early this year. The company is also working to take its epitaxial APD technology from single discrete detectors to arrays, and coupling this with its surface-mount, low-cost packaging to meet the technical, volume and price expectations of the automotive market.
To address range, LIDAR manufacturers are turning to longer wavelengths, where lasers can be run at higher powers before eye safety is affected. Here, InGaAs APDs are already being used on the road for cars, where autonomy at high speeds makes range important, and they are being designed into autonomous truck systems with their longer braking distances and reduced maneuverability, where massive safety improvements and gains in operating efficiency are the key benefits.
Aesthetics, in an age where form is just as important as function, means that the bulky and unattractive pods on top of vehicles will not be part of those systems that reach the market in vehicles for sale. Rotating scanning modules are not ideal as a solution to reliably handle arctic and desert, tarmac and trail terrains, so the next generation will feature solid-state scanning and fixed APD arrays. Excelitas 905nm lasers are already being used to develop waveguide-based systems to enable solid-state scanning, and new laser and detector arrays will also help to enable development in this area.
As is often the case, using existing technologies in an innovative way opens up new markets. When the remaining challenges can be satisfactorily overcome and driver safety can be sufficiently assured, the use of LIDAR technologies to enable autonomous vehicles will present tremendous new opportunities.