Solid-State LiDAR

Overview

Solid-state LiDAR refers to laser radars without macro-scale mechanical moving parts. Compared to traditional mechanical rotating LiDARs, solid-state LiDARs offer higher reliability, smaller form factors, and lower cost potential. They are considered the key direction for LiDAR technology transitioning from R&D to large-scale mass production.

Mechanical vs. Solid-State: Why Solid-State Is Needed

Dimension	Mechanical Rotating	Solid-State
Moving Parts	Motor-driven rotation	No macro-scale moving parts
Lifespan	~10,000 hours	~100,000 hours
Size	Larger (rotation space needed)	Compact
Weight	500g - 2kg	100g - 500g
FOV	Horizontal 360 deg	Limited (60-120 deg)
Reliability	Vibration/shock sensitive	High reliability (automotive grade)
Cost (Mass Production)	High (precision machining)	Low (semiconductor process)
Automotive Certification	Difficult	Easier to achieve

FOV Limitation

The biggest limitation of solid-state LiDARs is the inability to achieve 360-degree omnidirectional scanning. Solutions include combining multiple solid-state LiDARs for full surround coverage, or using a single sensor for specific applications (e.g., forward perception).

Solid-State LiDAR Technology Routes

graph TD
    A[Solid-State LiDAR Technology Routes] --> B[MEMS Micro-Mirror]
    A --> C[OPA Optical Phased Array]
    A --> D[Flash LiDAR]
    A --> E[Prism/Wedge Mirror Rotation]
    A --> F[FMCW Solid-State]

    B --> B1[MEMS mirror controls laser direction<br>RoboSense RS-M1]
    C --> C1[Electronically controlled beam steering<br>Quanergy, Analog Photonics]
    D --> D1[Area emission/reception<br>Ibeo, Continental]
    E --> E1[Small optical element rotation<br>Livox Mid-360]
    F --> F1[Frequency modulation + coherent detection<br>Aeva, SiLC]

MEMS Micro-Mirror Scanning

Principle: Uses MEMS-fabricated micro-mirrors to deflect the laser beam direction, achieving 2D scanning.

\[ \theta(t) = \theta_0 \sin(2\pi f t) \]

Where \(\theta_0\) is the maximum deflection angle and \(f\) is the mirror resonant frequency.

Technical Features:

Feature	Description
Scanning Method	1D or 2D MEMS mirror rapid oscillation
Advantages	Mature technology, controllable cost, mass-producible
Disadvantages	Mirror is still a moving part (microscale), limited shock resistance
FOV	Typically 60-120 deg
Representative Products	RoboSense RS-M1, Innoviz InnovizTwo

Workflow:

1. Laser emits a pulse
2. MEMS mirror deflects the laser to the target direction
3. Reflected light returns through the same MEMS mirror
4. Detector receives and measures time of flight
5. MEMS mirror oscillates rapidly to cover the entire FOV

OPA Optical Phased Array

Principle: Similar to phased-array radar, uses phase differences among multiple emitting elements to control the laser beam direction, achieving purely electronic beam steering.

\[ \theta = \arcsin\left(\frac{\lambda \cdot \Delta\varphi}{2\pi d}\right) \]

Where:

• \(\lambda\) is the laser wavelength
• \(\Delta\varphi\) is the phase difference between adjacent elements
• \(d\) is the element spacing

Technical Features:

Feature	Description
Scanning Method	Purely electronic phase adjustment
Advantages	Truly no moving parts, extremely fast scanning, random access
Disadvantages	High technical difficulty, limited power, sidelobe issues
Status	R&D stage, few commercial products
Representative	Quanergy (bankrupt), MIT/Caltech research

OPA Challenges

OPA is theoretically the ideal solid-state solution, but faces technical challenges in element count, emission power, and sidelobe suppression. Commercialization progress has been slow, and large-scale deployment is unlikely in the near term.

Flash LiDAR

Principle: Similar to a camera's "flash," illuminates the entire scene at once, using an area detector array (e.g., SPAD array) to simultaneously receive reflected signals from all directions.

Technical Features:

Feature	Description
Scanning Method	No scanning, parallel area detection
Advantages	No moving parts, high frame rate, simple structure
Disadvantages	Short range (energy dispersal), resolution limited by array size
Suitable Scenarios	Close range (<30m), blind-spot coverage, gesture recognition
Representative	Ibeo, Continental, Apple iPad LiDAR

Flash LiDAR's distance limitation stems from energy conservation:

\[ P_{\text{per pixel}} = \frac{P_{\text{total}}}{N_{\text{pixels}}} \]

Covering more pixels or longer distances requires more emission power, constrained by eye safety standards.

Prism/Wedge Mirror Rotation (Livox Approach)

Principle: Uses one or more small rotating prisms to change the laser direction, producing a unique non-repetitive scan pattern.

Technical Features:

Feature	Description
Scanning Method	Small prism rotation (semi-solid-state)
Advantages	Low cost, higher reliability, non-repetitive scanning improves equivalent resolution
Disadvantages	Strictly speaking not purely solid-state, still has small moving parts
Representative	Livox Mid-360, HAP, Avia

Non-repetitive scanning coverage:

\[ C(T) = 1 - e^{-\lambda T} \]

Where \(C(T)\) is the FOV coverage within integration time \(T\), and \(\lambda\) is the scan density parameter. Coverage exponentially approaches 100% with increasing time.

FMCW Solid-State LiDAR

Principle: Combines FMCW ranging with solid-state beam control (OPA or MEMS) for coherent detection.

Technical Features:

Feature	Description
Ranging Principle	Frequency-modulated continuous wave (coherent detection)
Advantages	Simultaneous distance + velocity, strong anti-interference, high sensitivity
Disadvantages	Highest technical complexity, high cost
Representative	Aeva Aeries II, SiLC Eyeonic

Key advantage -- Instantaneous velocity measurement:

\[ v = \frac{f_{\text{Doppler}} \cdot \lambda}{2} \]

Traditional ToF LiDAR can only estimate velocity through consecutive frame differencing, while FMCW can directly obtain radial velocity from a single measurement.

Livox Mid-360 in Detail

Livox Mid-360 is currently one of the most popular solid-state (semi-solid-state) LiDARs in robotics.

Specifications

Parameter	Value
Range	40m (@10% reflectivity), 70m (@80%)
FOV	360 x 59 deg (-7 to +52 deg)
Point Rate	200,000 pts/s
Scanning Method	Non-repetitive rotating prism
Accuracy	+/-2cm (@0.2m-10m)
Returns	Dual return
Built-in IMU	6-axis, 200Hz
Interface	100M Ethernet
Size	63.18mm diameter x 47.98mm
Weight	~265g
Power	5.5W (typical)
Protection	IP67
Operating Temperature	-20 to +55 deg C
Price	~$1,099

Non-Repetitive Scanning Pattern

graph LR
    subgraph "Integration Time Effect"
    A["50ms<br>Low Coverage"] --> B["100ms<br>Medium Coverage"]
    B --> C["200ms<br>High Coverage"]
    C --> D["500ms<br>Near-Full Coverage"]
    end

As integration time increases, the scan pattern gradually fills the entire FOV, with equivalent resolution continuously improving. This is the core advantage of Livox over traditional line-scanning LiDARs.

Typical Applications

• Quadruped robots: Lightweight, 360-deg FOV, built-in IMU, suitable for FAST-LIO2/Point-LIO
• Drone mapping: Lightweight, low power
• Service robots: 3D perception and obstacle avoidance
• Low-speed autonomous driving: Campus logistics vehicles, delivery robots

Technology Route Comparison

Technology Route	Maturity	Cost	Performance	Mass Production Difficulty	Main Bottleneck
MEMS Micro-Mirror	4/5	Medium	Good	Medium	Mirror reliability
Prism Rotation	4/5	Low	Good	Low	Not purely solid-state
Flash	3/5	Low	Medium (short range)	Low	Range limitation
OPA	2/5	High	High potential	High	Power, sidelobes
FMCW	3/5	High	Excellent	High	System complexity

Future Trends for Solid-State LiDAR

1. Cost reduction: As production scales up, automotive-grade solid-state LiDAR prices will drop to \(200-\)500
2. Chipification: Integrating emitter, scanner, receiver, and processor into a single chip (SoC LiDAR)
3. FMCW adoption: 4D LiDAR (range + velocity) will become the next-generation standard
4. Camera fusion: Integrating LiDAR and camera in the same sensor module
5. 1550nm wavelength: Eye-safe + higher power = longer range

References

• Livox Technical White Paper: https://www.livoxtech.com
• LiDAR Technologies and Systems - SPIE
• Aeva FMCW Technical Documentation
• RoboSense RS-M1 Product Specifications
• Hesai FT120 Technical Data

---

评论