Image Courtesy: Amy Pan and Sampson Wilcox
Researchers at the Massachusetts Institute of Technology have developed a new silicon-photonics lidar chip that overcomes one of the technology’s most persistent limitations, potentially paving the way for smaller, cheaper, and more reliable sensors for autonomous vehicles, drones, and industrial mapping systems.
The team created a novel antenna architecture that significantly reduces interference between components on a lidar chip, allowing for a much wider field of view without sacrificing measurement accuracy. The breakthrough addresses a challenge that has limited chip-based lidar systems for years and could accelerate the adoption of solid-state sensors with no moving parts.
Lidar, short for Light Detection and Ranging, works by emitting pulses of laser light and measuring how long they take to bounce back from surrounding objects. The technology enables machines to create detailed three-dimensional maps of their environment and is considered a critical component in autonomous navigation systems.
While traditional lidar systems often rely on rotating mechanical assemblies to scan their surroundings, silicon-photonics lidar replaces moving components with integrated optical phased arrays that steer light electronically. This approach promises lower costs, greater durability, and easier mass production, but engineers have struggled to expand the technology’s field of view without introducing unwanted distortions.
The MIT researchers tackled the problem by designing three different types of antennas that can be placed much closer together while minimizing crosstalk, the interference that occurs when neighboring antennas interact. In conventional systems, tightly packed antennas can severely distort outgoing light beams, forcing engineers to space them farther apart and limiting scanning performance.
Tests showed the new design reduced coupling between antennas from roughly 100 percent in comparable conventional systems to about 1 percent while maintaining a single, highly precise beam. The approach also eliminated grating lobes, duplicate beams that can confuse sensors, create false readings, and reduce overall efficiency.
The advance could have significant implications for autonomous vehicles, aerial surveying, construction monitoring, robotics, and other industries that rely on accurate environmental sensing. Wider viewing angles and improved beam quality could enable lidar systems to detect hazards more effectively while reducing hardware complexity.
The research also highlights the growing importance of silicon photonics, a field that uses semiconductor manufacturing techniques to manipulate light on computer chips. As demand rises for AI-powered machines capable of understanding and navigating complex environments, compact optical sensing technologies are becoming increasingly valuable.
The findings were published in the journal Nature Communications, and the researchers say they are now exploring ways to expand the field of view even further while investigating additional methods for wide-angle beam steering.
