The road to implementing autonomous driving (AD) functions requires the development and integration of brand new technologies in vehicles, and sensors are a fundamental part of it. But what exactly is the role of Lidars in this roadmap?
Although some autonomous vehicles such as the Google Car are in operation today, the automotive industry agrees that the technology used on these vehicles is still far from being fit for consumer products (the Lidar scanner protruding from the top of the Google Car is worth around $70,000 alone!). While showing tremendous potential, these sensors clearly do not make the cut yet: too bulky, too expensive, not robust enough.
So then, is Lidar the right choice, or will it be out-staged by other, more optimized technology? Clearly, our recent discussions with automotive industry experts and tier-1 players, including during the Autonomous Cars Silicon Valley conference in February, re-confirm the strategic role of the Lidar in implementing ADAS and AD functions.
According to Frost and Sullivan analysts, Lidar will be the preferred sensor suite for autonomous vehicles. This vision is also being shared by Goldman Sachs Global Investment Research, which forecasts that while Lidars have a slower path to market than other, more mature sensing technologies, it is expected at term to become the largest revenue generator of the sensor segment surpassing cameras and radars from 2030 on.
Source: Goldman Sachs Global Investment Research.
So, what exactly does optical detection and ranging and Lidars bring to the table that other sensors can’t match? In a post-conference report from Autonomous Cars Sillicon Valley just released, Amine Taleb, advanced projects, comfort and driving systems at Valeo states that “LiDAR technology holds many of the aces needed to reach the autonomous goal”. These aces include wider field of view, longer range, superior distance measurement accuracy and simultaneous tracking of objects. “These are very powerful characteristics of such a sensor that you cannot find in other sensors” concludes Taleb.
Not your Grandpa’s car
Not to say that Lidar will replace the other sensing technologies already used in ADAS applications. Everyone agrees that the successful road towards more autonomy will rely on a combination of sensors complementing each other to provide the best possible “multi-spectrum” sensing, leveraging light as well as radio and ultrasonic signals. The following chart demonstrates the arsenal of sensing technologies developed for automotive applications. These will also include a more integrated use of GPS information and advanced mapping applications, as well as communication between vehicles (V2V) and between the vehicle and external systems and infrastructures (V2X).
Needless to say that, in order to work together and with extremely high reliability, these components require the development of powerful signal processing software and complex decision making algorithms.
Leddar’s answer to autonomous driving challenges
It is with great momentum and anticipation that a LeddarTech delegation left for the recent Autonomous Cars Silicon Valley, which gathers some of the key think tanks, influencers, visionaries, and developers of autonomous driving technologies on the west coast.
LeddarTech was one of the select few vendors that had the opportunity to display and discuss its novel and upcoming solutions in front of this audience. This allowed us to demonstrate how Leddar technology enables automotive-grade Lidars to easily be integrated into standard headlamp and tail light assemblies in order to provide highly reliable detection and ranging capabilities.
This fixed-beam Lidar solution already generates a palpable interest, being affordable (with price points equal or inferior to those of radars), packaged in a compact solid-state design (no moving parts), while delivering reliable detection over a broad field of view.
This was also the opportunity to reinforce Leddar’s unique capability to bridge some key technology gaps that exist for the deployment of Lidars by Tier-1 automakers in mass-produced vehicles at prices that would make them attractive for the average car buyer.
LeddarTech’s VP Sales Dakx Turcotte at the Autonomous Cars Silicon Valley event.
Autonomous cars: the long road ahead
Besides the technological advancements required to enable the functional and safe commercial deployment of autonomous cars, many other very important questions must be answered before we can see widespread consumer autonomous vehicles deployments on our roads.
Legally, a plethora of things are getting stirred up, especially regarding liability: who’s responsible for your security, and who’s going to be liable if something goes wrong? The component manufacturer, the system integrator, the software developer, the car maker? Or is it the driver? As well, the very concept of “driver” has to be thought over (some laws already recognize a computer system as a legal driver). These are key questions that legislators in all countries and the insurance industry are mulling over.
Psychology will also be looking for some answers: researchers also admit that we still know very little about the relationships between a human and semi-autonomous or fully autonomous systems, especially when our safety is in play. How will we behave, interact, and react to various situations? How to share driving duties and responsibilities? When/how the human driver takes command, under which circumstances? Will he/she be alert and ready to take control? What about misuses of the systems and how should we engineer around it?
Will we get to see autonomous cars out of the assembly line hitting our roads anytime soon? Not everybody sings the same song about the timeframe: while the agile, techno-centric Tesla announced they would produce such a vehicle by 2018, most observers and Tier-1 car makers estimate that it will take another 10-15 years for the industry to be ready for it.
The road from the first ADAS solutions to full autonomy may be a long one: as demonstrated by the following table, inspired by recent data from the Evercore ISI analyst firm, we are still years away from the mass adoption of semi-autonomous systems. In the meantime, Lidar will become even more instrumental to AD implementation, and we are excited to be contributing in making AD a reality for the benefit of all drivers, passengers, and road users.
Source: Evercore ISI, Dec. 2015