2023 IISW : Samsung Research Team Proposes Efficient LiDAR System for Mobile Applications with Enhanced Ambient Light Tolerance

Home » Blog » Industry Trends/ Developments » 2023 IISW : Samsung Research Team Proposes Efficient LiDAR System for Mobile Applications with Enhanced Ambient Light Tolerance

Highlights

The authors propose a LiDAR system for mobile applications that is power efficient and can tolerate high ambient light.
The system uses a 940nm scanning laser source and a 192×144 SPAD array receiver with on-chip time correlated histogramming.
An adaptive single-pass histogramming architecture saves power by staying in coarse resolution mode when returned laser pulses are weak.
Novel signal processing methods allow depth resolution beyond the coarse bin size.
The system can measure depth images at 30 fps with up to 10 meters range, 1% range error, and only consumes 12mW optical power.
Simulations show the system achieves a good balance between power efficiency, ambient light tolerance, range, and resolution.
The authors conclude the system is suitable for mobile depth sensing applications.

Background

LiDAR (Light Detection and Ranging) sensors have become important for depth sensing in applications like augmented reality, autonomous vehicles, and surveillance. Among different LiDAR approaches, direct time-of-flight (dToF) is advantageous because of its immunity to multi-path interference and ease of integration with vision systems. However, designing a dToF system that balances power efficiency, ambient light tolerance, range, and resolution is challenging. Increasing laser power or reducing histogram bin size often improves performance but also increases power consumption and memory requirements.

Results

In this work, the authors propose a dToF LiDAR system optimized for mobile applications. The system uses a 940nm scanning laser source and a 192×144 single-photon avalanche diode (SPAD) array receiver with on-chip histogramming. A key feature is an adaptive single-pass histogram architecture that operates in coarse resolution mode when returned signals are weak, saving power. Novel signal processing methods, including pulse collision recovery and conditional peak enhancement filtering, help improve range resolution beyond the coarse histogram bin size. Simulations of the system demonstrate 1% range accuracy up to 10 meters, 30 frames per second operation, and only 12mW optical power consumption.

Methods

System Design:

Employed a scanning LiDAR architecture with 940nm laser source transmitter and 192×144 SPAD receiver array.
Flexible addressing scheme in SPAD array to match scanning patterns and avoid missing signals.
On-chip processing including counter circuits and TDCs to enable real-time histogram building.

Adaptive Histogramming:

Single-pass histogram architecture that adapts from coarse to fine resolution based on signal strength.
Saves power compared to two-pass approaches by avoiding discarding weak signals.

Signal Processing:

Pulse collision recovery algorithm to regain signals lost due to pulse merging.
Conditional FIR filtering for accurate enhancement of histogram peaks.

Performance Evaluation:

Monte Carlo simulations to estimate measurement error under different conditions.
Assessed key metrics like range, precision, frame rate, power consumption.

Conclusions

The adaptive single-pass histogramming provides superior performance compared to two-pass approaches under limited optical power budgets. By recovering signals lost to pulse collisions and accurately enhancing peaks, the processing methods help maximize range resolution. The system strikes a good balance of power efficiency, ambient light tolerance, range, and resolution. According to the simulation results, the system is well-suited for mobile depth sensing applications requiring low power consumption.

Keywords: LiDAR, SPAD, dToF

Figure 1.This work proposes a LiDAR system with a 940nm infrared transmitter (TX) and addressable SPAD array receiver (RX).

Figure 5.Sensor chip (RX) architecture.

References:

https://imagesensors.org/Past%20Workshops/2023%20Workshop/2023%20Papers/P14.pdf