What Challenges Does LiDAR for Self-driving Vehicle Face?

To achieve vehicle-level mass production, the most important thing is to design for manufacturability, reduce costs and improve performance. Although mechanical vehicle-mounted LiDAR is still in the mainstream of car application, solid-state LiDAR for self-driving vehicles meeting the concept of miniaturization and low cost has become increasingly popular in the industry.

All kinds of self-driving functions rely on the development of specific application algorithms. It is the ultimate goal of LiDAR application algorithm development is to accurately and quickly extract effective data from the complex LiDAR point cloud data and correctly understand and analyze useful information under the complex and changeable self-driving environment. At present, there is no unified framework and judgment standard for the application algorithm of LiDAR for self-driving vehicles. Some R&D teams have proposed many algorithmic solutions. The common technologies include point cloud segmentation technology, target tracking, and recognition technology, real-time positioning, map construction technology, etc. The key to the research on LiDAR for self-driving vehicles is how to improve the algorithm’s robustness and scalability.

LiDAR for self-driving applications still face many challenges:

1) The cost remains high

The high cost is almost the biggest obstacle to the large-scale promotion and use of LiDAR for self-driving. At present, the price of automotive LiDAR with excellent performance and suitable for high-level self-driving cars on the market ranges from several thousand dollars to tens of thousands of dollars, which is much higher than other environmental sensing sensors, and some even exceed the price of self-driving cars, which is hardly accepted by the self-driving market.

2) Difficulties in mass production of car-level LiDAR.

Although Valeo is currently the only supplier that announced their mass-produced LiDAR has passed the automotive grade, however, this mechanical rotating LiDAR is difficult to increase the number of lines, which causes huge restrictions in vertical FOV and angular resolution. Therefore, Valeo has also turned to research and development of solid-state LiDAR.
To realize mass production of automotive-grade LiDAR, it needs to meet various requirements such as performance, environmental adaptability, reliability, and product consistency, and LiDAR suppliers need to establish standardized and automated assembly lines. In addition, the effective verification method of the LiDAR for self-driving has not yet been concluded.

3) Climate and environmental influences.

The detection beam of the automotive LiDAR is affected by atmospheric absorption, scattering, and refraction. First of all, the LiDAR for self-driving vehicles is generally installed on the top of the car or embedded around the car body. The lower installation height might enlarge the echo reduction effect caused by certain gas molecules and suspended particles in the atmosphere, resulting in a worse reception effect of the LiDAR detector. Secondly, under severe weather such as rain, fog, ice, and snow, suspended matters in the air will adversely affect the laser emission, reflection, and detection processes, resulting in a decrease in the detection range and detection accuracy of the LiDAR.

4) Adaptability of application algorithms

Currently, various vehicle-mounted LiDAR application algorithms are usually only developed for a specific scenario, and the higher the accuracy, the worse the adaptability, and the limited use range. In the face of various complex and changeable self-driving scenarios, it is necessary to make the algorithm expandable and portable, and it is particularly necessary to improve the adaptability of the algorithm.

About Neuvition LiDAR

Neuvition’s solid-state HD vehicle LiDAR uses MEMS micro-galvanometer plus 1550nm laser technology, which improves reliability but cuts cost. Titan M1-Pro LiDAR has a high resolution of 480 lines, a FOV of 120 °, and an effective detection distance of 200 meters. The super performance and significant improvement on safety for autonomous driving make Titan M1-Pro the best solid-state vehicle LiDAR for self-driving cars.

Founded in 2016, Neuvition is the first professional manufacturer focusing on solid-state HD LiDAR based on the MEMS technology route. To bring better vision for smarter future, our LiDAR mainly provides smart solutions in the fields of autonomous driving, ADAS, V2X/CVIS, railway ADAS, 3D mapping, etc., to enable machines or vehicles to have perception capability and to accelerate the speed of automation processes.
Our technologies cover the world’s leading high-definition solid-state MEMS LiDAR, LiDAR data processing algorithm and software, AI, and machine learning in the application of LiDAR and other LiDAR-based software solutions. Neuvition is one of the leading companies to commercialize solid-state 3D HD video LiDARs for automated market needs with performance and parameters superior to other competitors.

To satisfy our users’ needs or drive their business growth and speed up the development of the automation world, Neuvition keeps on moving, exploring, and sharing!