Fueling LiDAR Technology: 3 Startups Aim to Fast-track Autonomous Driving
As the battle for autonomous driving heats up, more companies are investing in LiDAR. From improved sensing to a smaller design, three startups are adding fuel to the LiDAR fire.
From autonomous driving to surveying and beyond, light detection and ranging (LiDAR) have become integral in many applications. Many companies are thus fighting to create the best LiDAR implementations in their given industries. Accordingly, analysts expect the market to eclipse $2.9 billion by 2027. That forecast has enticed numerous new startups to enter the marketplace to keep the momentum rolling of LiDAR’s immediate relevance.
Proposed LiDAR market growth to 2025. Image used courtesy Global Marketing Insights
However, not all companies or missions are created equally. LiDAR solutions alone aren’t compelling unless they capture customer interest or bring technical improvements to the forefront. This article will round up three companies that do just that and are making waves in the LiDAR industry.
Aurora Tech and its FirstLight LiDAR
Fresh off of a merger agreement with Reinvent Technology Partners Y and a public debut, Aurora, an autonomous driving company at heart started in 2017, has signaled its intention to become a significant LiDAR player, especially with its first autonomous product release in 2023.
Scanning speed, accuracy, and distance are key performance indicators for LiDAR systems. A sensor must pick out objects and continuously determine their positioning (and relevance to the road). This positioning is essential in autonomous driving, where identifying other cars and pedestrians is paramount.
The company hopes to scale its driver technology within the trucking industry. This autonomous system relies on an integrated hardware-software stack, relying heavily on Aurora’s FirstLight LiDAR technology.
Aurora's FirstLight LiDAR. Image used courtesy of Aurora
Aurora plans to implement these units in Class 8 trucks and eventually passenger vehicles. It claims that improvements in object tracking, distance calculation, and speed will help bring this vision to life.
FirstLight LiDAR relies on long-range, multimodal sensing. While humans rely on their senses, LiDAR systems draw on various cameras and sensors to operate effectively. These components might include imaging devices, inertial measurement units, and global satellite navigation systems. Each vehicle in Aurora’s fleet will capture real-time data from multiple sources and expose it to computerized algorithms. Machine learning and AI will then help create a sense of environmental awareness for the vehicle in question.
AM vs. FM LiDAR. Image used courtesy of Aurora
One helpful aspect to FirstLight is frequency modulation. Frequency modulation allows FirstLight to send waves outward continuously, thus analyzing objects up to 300 meters away. It can also measure approach speed and angles of incoming objects and then determine if they’re noteworthy.
Aurora also claims its technology is less afflicted by common issues caused by bright sunlight, sensor crosstalk, and self-interference. By partnering with Volvo and Toyota, Aurora aims to outperform amplitude modulation (AM) LiDAR technologies in the coming years.
Though Aurora is aiming to innovate LiDAR technology, another company AEye Inc is trying up the ante even further.
Founded in 2013 and focused on autonomous driving, AEye aims to boost response and driving safety with its proprietary LiDAR system. The company’s intelligent detection and ranging (iDAR) system states to excel at doing more with less by identifying and prioritizing critical data while capturing four to eight times more of it than traditional LiDAR alternatives.
The four steps of iDAR. Image used courtesy of AEye Inc
While fixed scanning patterns fail to assign priority to key pixels, iDAR could react more effectively to sudden changes as measured by its sensor stack.
AEye even claims that iDAR slashes irrelevant data consumption by up to 90%. It does this while reducing power consumption by five to ten times. These claims offer massive improvements over older solutions, where 75 to 95% of incoming data could be meaningless.
The company has thus designed its system to emulate the human visual cortex and its ability to assess our surroundings. The system is therefore much more adaptive. Since the reception channels and transmission are separated, iDAR can immediately expose incoming sensor data to its AI platform. Real-time environmental scanning can happen much quicker as a result.
From a hardware standpoint, iDAR’s implementation is relatively straightforward. The system uses a solid-state adaptive LiDAR unit, a low-light HD camera, a receiver, and SoC. Its overall packaging remains fairly compact in both the company’s 4SIGHT A and 4SIGHT M (which is preparing for volume production) units. The iDAR leverages RGB and XYZ sensor data to read a scene, blending 2D, 3D, and 4D information with less latency and bandwidth usage.
With its promising products reliant on MEMS technology, AEye hopes to become an autonomous force in mobility, trucking, transit, construction, aerospace, and more.
As both AEye and Aurora have shown, LiDAR startups are trying to push the boundaries of LiDAR technology and integration. One final company that has been standing out is Sense Photonics.
Last but not least, Sense Photonics emerged in 2016, intending to democratize 3D vision, primarily within the automotive industry.
The company produces flash LiDAR technology for OEMs while equipping each vehicle with a wide, reactive field of view during autonomous operation. Sense Photonics debuted its illuminator LiDAR floodlight in 2018 and soon paired it with Sense Silicon just two years later. According to Sense Photonics, this component is the world's "highest resolution 3D CMOS SPAD Imager."
Accordingly, its LiDAR unit combines this single-photon avalanche diode (SPAD) technology with vertical-cavity surface-emitting lasers (VCSEL) emission. SPADs operate with minimal crosstalk and can detect events (incoming cars or pedestrians) based on light reactions reflected back to the vehicle's sensors.
Meanwhile, VCSEL lasers have a narrow beam divergence and minimal shift as temperatures fluctuate. They also have a high pulse frequency, pair well with dotted spectral patterns in the car's field of view, and are highly efficient, which helps limit motion blur and data gaps.
Example of one of Sense Photonics' LiDAR systems. Image used courtesy of Sense Photonics
Many have seen just how large traditional LiDAR systems can be. Mounted on vehicle rooftops, they can typically span multiple feet in size while weighing a ton. On the other hand, Sense Photonics' LiDAR platform offers a much more compact approach to autonomous sensing. The unit is designed to fit behind the windshield or within vehicle headlamps. Interestingly, the company likens its system to the one found within current iPhones.
Currently, Sense Photonics offers four LiDAR units and touts their adaptive response time, even within advanced driver assistance systems (ADAS). The company has partnered with a major "Detroit-based automotive OEM," possibly General Motors, to eventually bring its products to mass-market cars. Sense Photonics aims to make its short-and-medium range sensors cost-effective for suppliers worldwide.
Importance of Startups
Recently, startups have been a major focal point in the tech industry. Often, fresh and innovative ideas can come out of these young, aspiring companies, which can benefit the industry as a whole, and larger companies looking for the next boost to their technology. These startups provide new ideas and become a shortcut to R&D for major companies, opening up the floor to newer and better tech.
Interested in even more tech startup news? Find out more in the articles down below.
Arm’s Startup Day Shows its Support for Future Hardware Startups
Blending AI and Hardware: 3 Startups Reach for AVs and the Cloud
AI on Demand: SambaNova Hopes to Bridge the AI Integration Gap