Lidar’s Critical Role in Next Gen Autonomous Driving

The auto industry finds itself at a crossroads. According to a recent CarGurus study, 33% of respondents are excited to see the advancement of self-driving vehicle technology. However, respondents' trust in current technology shows that the excitement only goes so far, with 53% of respondents wanting to continue to maintain some control over the vehicle. Citing safety (at 51%) as the most concerning aspect of autonomous vehicles, this disconnect between expectation and reality shows that technology has its work cut out.

Safety is unmistakably the next frontier for differentiation among automotive OEMs, particularly as the industry moves further along the road to autonomous driving. Advanced Driver Assistance Systems (ADAS) will play a pivotal role in delivering a new wave of safety experiences that drivers will expect and that OEMs will rely on to stand out. Beyond simple lane keep assist or low-speed autopilot, ADAS for use in autonomous vehicles will need to anticipate dangerous scenarios as they develop in real time and enable vehicles to take rapid, proactive action faster than a human ever could. Lidar and its accompanying perception software is poised to enable existing ADAS to do all that and more.

What is lidar?

Lidar (light detection and ranging) uses lasers to create a three-dimensional image of the environment or an object, whether moving or stationary. Lidar has evolved rapidly and significantly, with more compact hardware and more advanced technology, to become an important advancement in the automotive industry’s quest for autonomous vehicle safety. Lidar works by creating a point cloud that maps a vehicle’s surroundings, identifying obstacles and assisting a vehicle’s ADAS to avoid crashes.

What are the different types of lidar and how is it different from radar?

There are many different types and uses of lidar, with applications in a number of industries outside of automotive. The type that auto manufacturers focus on for use in the development of autonomous vehicles is “terrestrial, mobile” lidar. In a terrestrial lidar system, hardware is mounted on a vehicle and surveys the ground and surrounding objects. The system sends out laser pulses and then processes that information to “see” obstacles, pedestrians, other vehicles, and more. Lidar provides a vehicle’s ADAS with near-instant readings for distance and object measurements.

As far as the differences between lidar and radar, it primarily comes down to light versus radio waves. Lidar systems use many laser pulses per second, measuring the time they take to bounce back from the object. This data is used to calculate an accurate distance between itself and the object. Radar works similarly but uses radio waves and an antenna. Radar can be used over long distances but has much lower resolution than lidar, thus making objection detection less effective.

Are cameras as effective as lidar for autonomous driving?

The answer goes back to safety.

Camera-based systems take pictures of the surrounding environment and send those pictures through an AI algorithm that learns to identify objects over time with the expectation that this system can better handle unpredictable environments. However, some important downsides to using camera-based systems in ADAS and autonomous vehicles are the inability to perform at highway speeds, difficulty performing in rain and fog, unreliability at night, and the significant computing effort required to aggregate multiple camera feeds. Additionally, current camera module technology is not suitable for long ranges and wide fields of view.

Lidar systems perform more accurate 3D mapping of a vehicle’s surroundings and process this information quickly and predictably. Lidar works well in different lighting conditions, even in direct sunlight and complete darkness, and excels at evaluating distance regardless of road composition. And with a nod to future tech, it's possible that one day a city’s smart infrastructure system could talk directly to an autonomous car’s lidar-enhanced ADAS system, communicating things like traffic management and construction detours.

What are the different levels of driving automation as it pertains to the progression of lidar?

The Society of Automotive Engineers (SAE) defines six levels of driving automation, as shown below. The international standards they have identified serve as benchmarks in developing autonomous vehicles. Levels 0-2 require drivers to actively participate in the driving experience, and the technology functions as support to the driver. Level 3 is a hybrid with some automated features but the driver ultimately in control when the ADAS requests it. Levels 4 and 5 require no human interaction to perform.

Source: SAE Levels of Driving Automation?

Many OEMs have already met the criterion for safety standards for Levels 0-2 in generations of vehicles. Now, manufacturers have begun partnering with lidar developers to move their ADAS to Level 3, inching closer to true autonomous driving and looking to the future of meeting the standards for Levels 4 and 5. Even as critical decisions are made by OEMs on how to deliver safe next generation autonomous driving experiences, the question remains: what does it take to achieve high performance at highway speeds? MicroVision operates where the advancement of lidar plus perception technology can make the biggest impact.

What’s the connection between lidar and the success of self-driving cars?

?The success of self-driving cars lies at the intersection of safety at high speeds and seamless integration, including cost factors, for OEMs. Lidar’s exceptional ability to accurately map surroundings allows autonomous vehicles to avoid obstacles and collisions, including dangerous conditions posed by pedestrians, moving and stationary vehicles, bikes, and animals. By combining faster processing times to detect obstacles with the ability to identify the variety of obstacles in a cost-effective system that can perform at highway speeds, the dream of Level 5 autonomous vehicles comes much closer to becoming a reality.

What is the growth potential for lidar?

Growth potential for lidar as the way forward for autonomous vehicles is exponential. According to Global Market Insights, automotive lidar market size surpassed $150 million in 2017 and is anticipated to grow at over 40% CAGR from 2018 to 2024. GMI credits an increase in research activities in autonomous vehicle technology as fueling the industry growth. What about 2025 to 2030?

Why is MicroVision’s lidar superior?

?MicroVision approaches next generation use of lidar through its proprietary lidar technology based on known silicon, laser, and sensing technologies.?

?MicroVision has been at the forefront of MEMS-based laser beam scanning technology for the last 25 years. This breadth and depth of experience has resulted in an established semiconductor partnership network and a strong intellectual property portfolio, including over 430 patents in MEMs-based technology. As a leader in the field, we’re proud to have delivered technologies to some of the leading global tech companies along with the U.S. Military. This is all to say that we have a demonstrated track record for commercializing our innovations.

Today, MicroVision is focused on bringing our expertise to the exciting field of automotive lidar. Our lidar hardware and perception software will enable automotive OEMs to deliver a new generation of ADAS safety features at highway speeds. MAVIN, the latest integrated lidar system from MicroVision, will see farther, with greater clarity, and respond faster to emerging situations.

Our commitment to OEMs doesn’t stop at enabling their ADAS with ultra high-resolution point cloud data. Built with materials known to OEM supply chains, MicroVision hardware is scalable, sourceable, and supports a lower cost structure allowing OEMs to bring next gen ADAS features to more drivers in more cars.

“In the future, MicroVision expects to offer its Level 3 ADAS solution with two lidar sensors, a reduced number of long-range radar and cameras, with our new, dynamic-view system able to send this fused data to the domain controller. This eliminates the need for high computing power and allows for lower latency and faster calculations. The result is better data processing speeds, sleeker hardware, and a more affordable lidar system for our customers,” said Dr. Thomas Luce, Vice President of Business Development at MicroVision.

Collaboration: MicroVision aims to provide OEMs with a fully-integrated lidar solution that will enable the next wave of ADAS, defining the next era of differentiation in the auto industry. This approach allows OEMs to maintain control over the driving experience with the ability to leverage the point cloud to develop innovative new safety features as a basis of differentiation.

?Authentic Results: Delivered through a mature OEM partnering philosophy that emphasizes collaboration and joint development over empty promises of “turn-key” products, our test track performance shows that we are ready for highway speeds.?

As autonomous driving comes ever closer to reality, lidar systems are proving to be a critical component. Automotive manufacturers will need to choose a partner that progresses OEMs’ ADAS into the next generation with a system that delivers safety, accuracy, and reliability from our city streets to interstate highways.