Photonics advance could enable compact, high-performance lidar sensors

Lidar has become a standard sensor for autonomous systems, from self-driving cars to warehouse robots, using pulsed infrared light to map 3D environments with centimeter-level precision. Traditional lidar units rely on mechanical spinning modules to sweep their field of view, but these systems are bulky, expensive, and prone to failure as moving parts degrade over time. For years, researchers have pursued chip-based solid-state lidar as a smaller, cheaper, more durable alternative, but persistent technical limitations have kept these systems from matching the performance of their mechanical counterparts.

A new advance from MIT's Photonics and Electronics Group removes one of the largest barriers to practical chip-based lidar. The team designed a novel array of integrated antennas for silicon-photonics chips that eliminates unwanted crosstalk between adjacent components, enabling wide field-of-view scanning without the noise, false positives, or beam artifacts that have limited previous designs. The work, published in Nature Communications, was covered in MIT News on May 7, 2026. It builds on years of research into integrated optical phased arrays (OPAs), the core component of solid-state lidar systems. The MIT Photonics and Electronics Group led the development, with funding from the Semiconductor Research Corporation, National Science Foundation, and other partners.

The tradeoff in existing chip-based lidar

Chip-based lidar relies on OPAs, which steer light beams without moving parts by adjusting the phase of light delivered to each antenna in an array. Each antenna is a tiny silicon waveguide with periodic corrugations that scatter light up and out of the chip. By shifting the phase of light sent to each antenna, the array can constructively interfere to point the main beam in a specific direction, scanning the environment without mechanical motion.

This approach has a fundamental tradeoff in antenna spacing. If antennas are placed too close together, light leaks between adjacent waveguides, a phenomenon called crosstalk. This coupling jumbles the phase relationships between antennas, producing noisy, imprecise beams. If antennas are spaced farther apart to avoid crosstalk, the array produces secondary beams called grating lobes, which appear at regular angular offsets from the main beam. These spurious beams can reflect off objects and return to the sensor, creating false positives that confuse 3D mapping algorithms. Grating lobes also waste optical power, reducing the sensor's effective range.

"Spacing antennas far enough to avoid crosstalk limits our field of view, so the autonomous vehicle now only knows what is in front of it for a certain angular range," says Andres Garcia Coleto, an EECS graduate student and co-author of the paper. "The grating lobes also cause false positives by confusing the sensor, which is a critical safety issue for autonomous navigation."

A new antenna geometry breaks the tradeoff

Standard OPAs use identical antennas, all with the same width, corrugation pattern, and propagation coefficient, which measures how light travels through the waveguide. Identical antennas couple strongly when placed close together, leading to high crosstalk. The MIT team instead designed three distinct antenna geometries, varying the width of each waveguide and the size, spacing, and arrangement of their corrugations. Each design has a unique propagation coefficient, so when placed close together, the antennas do not interact with each other.

"The different propagation coefficients mean each antenna doesn’t ‘see’ the antenna next to it, so it won’t couple with its neighbor," Garcia Coleto explains. In experimental tests, the team’s array reduced crosstalk from roughly 100 percent in standard designs to just 1 percent, even with antennas spaced far closer than typical OPA arrays.

Varying antenna geometries introduces another challenge: different antennas typically emit light at different angles, with different power levels, and different steering responses. For a functional lidar, all antennas in the array must emit the same amount of light, produce beams at the same angle for a given wavelength, and shift their emission angle uniformly as the beam is steered. The team solved this by first developing new fundamental electromagnetic theory describing how radiative modes couple in these waveguides, using that theory to simulate and optimize antenna designs that meet all three requirements despite their different geometries.

They fabricated the optimized array using standard silicon photonics fabrication processes, compatible with commercial CMOS foundries, which means the design can be scaled for mass production without custom manufacturing steps. Testing confirmed the array can steer a single, precise beam across a wide field of view with no detectable grating lobes, a combination that was previously impossible for OPA-based systems.

Real-world impact and future work

Jelena Notaros, senior author of the paper and Robert J. Shillman Career Development Associate Professor of EECS at MIT, says the work solves a core problem for integrated optical phased arrays. "The functionality we demonstrated in this work solves a fundamental problem for integrated optical-phased-array technology, enabling future lidar sensors that can achieve significantly higher performance than we could demonstrate previously."

The advance has immediate implications for autonomous systems. Compact, solid-state lidar with wide field of view would allow self-driving cars to detect obstacles in their peripheral vision without blind spots, eliminate false positives from grating lobes, and reduce overall system cost and size. The same sensors could be used for aerial surveying drones, construction site monitoring, agricultural robotics, and even medical imaging devices that use light-based mapping.

Joyce Poon, a professor of electrical and computer engineering at the University of Toronto and director of the Max Planck Institute of Microstructure Physics, who was not involved in the work, calls the design elegant. "This work addresses a longstanding challenge in integrated optical phased arrays: simultaneously achieving both a wide field of view, which requires dense antenna spacing, and high beam quality, which requires low crosstalk between neighboring antennas. Their innovation is an important step forward for chip-scale, solid-state beam-steering technology."

The team plans to further improve the design to expand the field of view even wider, and is exploring additional solutions for wide-FOV operation that emerged from their underlying theoretical work. The full paper, "Reduced-Crosstalk Antennas for Grating-Lobe-Free and Wide-Field-of-View Integrated Optical Phased Arrays," is available in Nature Communications. More information on the research is available from the MIT Research Laboratory of Electronics.