Abstract

High-speed, precise 3D sensing is essential for autonomous driving, robotics, and remote sensing, with light detection and ranging (LiDAR) as the leading technology. Nevertheless, single-channel LiDAR is bottlenecked by the light’s round-trip delay, limiting its point acquisition rate (PAR). While parallel detection can overcome this, current solutions either require multiple detectors—adding complexity—or are hampered by electronic encoder speeds, restricting scalability and resolution.

Here, we propose a parallel anti-interference LiDAR architecture based on code-division multiple access (CDMA) with all-optical encoding. A broadband source is split by wavelength-division multiplexing (WDM) into two spectral channels, each encoded with an orthogonal sequence via fiber splitters. A single-pixel photodetector captures mixed echoes, and matched filtering separates the channels for precise time-of-flight (TOF) extraction—achieving parallel detection with minimal hardware. Experimentally, operating at a 25.6 ns emission period, the system attains a dual-channel PAR of 78 MHz (ambiguity distance of 3.84 m) with ~2 mm ranging precision (standard deviation) under motion.

Furthermore, even at an extremely low signal-to-interference ratio (SIR) of −13 dB, the scheme retains over 50% valid demodulated points, validating its superior interference resilience. This low-complexity, highly integrable architecture establishes a critical foundation for ultrafast, high-precision, anti-interference 3D imaging in autonomous vehicles, drones, and robotics.