New Lidar Chips for Self-Driving Vehicles are Smaller Than a Dime, Cost $10 to Manufacture
Please consider the IEEE Spectrum article MIT and DARPA Pack Lidar Sensor Onto Single Chip.
Light detection and ranging, or lidar, is a sensing technology based on laser light. It’s similar to radar, but can have a higher resolution, since the wavelength of light is about 100,000 times smaller than radio wavelengths. For robots, this is very important: Since radar cannot accurately image small features, a robot equipped with only a radar module would have a hard time grasping a complex object. At the moment, primary applications of lidar are autonomous vehicles and robotics, but also include terrain and ocean mapping and UAVs. Lidar systems are integral to almost all autonomous vehicles and many other robots that operate autonomously in commercial or industrial environments.
Lidar systems measure how far away each pixel in a 3D space is from the emitting device, as well as the direction to that pixel, which allows for the creation of a full 3D model of the world around the sensor. The basic method of operation of a lidar system is to transmit a beam of light, and then measure the returning signal when the light reflects off of an object. The time that the reflected signal takes to come back to the lidar module provides a direct measurement of the distance to the object. Additional information about the object, like its velocity or material composition, can also be determined by measuring certain properties of the reflected signal, such as the induced Doppler shift. Finally, by steering this transmitted light, many different points of an environment can be measured to create a full 3D model.
Most lidar systems—like the ones commonly seen on autonomous vehicles—use discrete free-space optical components like lasers, lenses, and external receivers. In order to have a useful field of view, this laser/receiver module is mechanically spun around, often while being oscillated up and down. This mechanical apparatus limits the scan rate of the lidar system while increasing both size and complexity, leading to concerns about long-term reliability, especially in harsh environments. Today, commercially available high-end lidar systems can range from $1,000 to upwards of $70,000, which can limit their applications where cost must be minimized.
Our work at MIT’s Photonic Microsystems Group is trying to take these large, expensive, mechanical lidar systems and integrate them on a microchip that can be mass produced in commercial CMOS foundries.
Our lidar chips are produced on 300-millimeter wafers, making their potential production cost on the order of $10 each at production volumes of millions of units per year. These on-chip devices promise to be orders of magnitude smaller, lighter, and cheaper than lidar systems available on the market today. They also have the potential to be much more robust because of the lack of moving parts. The non-mechanical beam steering in this device is 1,000 times faster than what is currently achieved in mechanical lidar systems, and potentially allows for an even faster image scan rate. This can be useful for accurately tracking small high-speed objects that are only in the lidar’s field of view for a short amount of time, which could be important for obstacle avoidance for high-speed UAVs.
Another problem bites the dust: cost.
Fully driverless vehicles, with no steering wheel, will be on US roads by 2021 in the US, sooner elsewhere. Fully driverless trucks will not only be on highways, but will be the primary mode of long-haul trucking in a 2022-2024 time frame.
For further discussion, including a rebuttal to the often stated claim that driverless vehicles cannot work in snow, please see Uber Offers Driverless Rides This Month! What About Snow, Rain, Pigeons, 80-Year-Olds on Roller Skates?
Mike “Mish” Shedlock