Automotive and Industrial RF Automotive Radar Informational

What is the role of 4D imaging radar in next generation automotive sensing?

4D imaging radar plays a critical role in next generation automotive sensing by simultaneously measuring four dimensions: range (distance to target), radial velocity (Doppler speed), azimuth angle (horizontal position), and elevation angle (vertical position) of every detected object in the scene. This produces a dense 3D point cloud with per-point velocity information, bringing radar perception closer to the scene understanding capability previously achievable only with LiDAR while maintaining radar's inherent advantages of all-weather operation, direct velocity measurement, and lower cost. 4D imaging radar uses large MIMO antenna arrays (typically 12+ TX and 16+ RX channels creating 192+ virtual elements) to achieve angular resolution below 2 degrees in both azimuth and elevation, enough to distinguish pedestrians from vehicles, detect lane markings through enhanced reflectivity processing, and map the driving environment in three dimensions. The elevation dimension is critical for distinguishing overhead objects (bridges, signs, traffic lights) from road-level obstacles, solving one of traditional radar's biggest limitations for autonomous driving. Key 4D imaging radar products include Continental ARS540 (the first production automotive 4D imaging radar), Arbe Phoenix (2304 virtual channels), and ZF PREMIUM radar.

Category: Automotive and Industrial RF

Updated: April 2026

Product Tie-In: Radar ICs, PCB Materials, Antennas

4D Imaging Radar Technology for Autonomous Vehicles

4D imaging radar represents a paradigm shift from traditional automotive radar, which detected a sparse set of object-level targets, to a sensor that produces point-cloud representations of the entire driving scene. This advancement is critical for the perception pipeline of Level 3+ autonomous driving systems.

What Makes It "4D"

Range: Distance to target from FMCW beat frequency, resolution 4-15 cm depending on bandwidth
Velocity: Radial velocity from Doppler shift, resolution 0.1-0.5 m/s. Unique to radar (LiDAR cannot directly measure velocity)
Azimuth angle: Horizontal position from MIMO virtual array processing, resolution 1-2 degrees with 192+ virtual channels
Elevation angle: Vertical position from MIMO virtual array processing in the orthogonal dimension, resolution 3-10 degrees

Comparison with Traditional Radar and LiDAR

Traditional automotive radar produces 10-100 detections per frame with no elevation information. 4D imaging radar produces 1,000-10,000+ detections per frame with full 3D spatial information plus velocity. LiDAR produces 100,000-300,000+ 3D points per frame with superior angular resolution but no direct velocity measurement and degraded performance in rain, fog, and dust. 4D imaging radar is increasingly seen as a cost-effective complement or potential replacement for LiDAR in production autonomous vehicles.

AI/ML Processing

The dense point cloud from 4D imaging radar enables deep learning-based perception algorithms similar to those used for LiDAR processing. Point cloud networks (PointPillars, PointNet, CenterPoint) can perform object detection, classification, and tracking directly on the radar point cloud. The additional velocity dimension actually improves classification accuracy in some scenarios compared to LiDAR-only approaches.

4D Imaging Radar Performance Metrics

Point cloud density: N_points ~ N_targets x (R_max/delta_R) x (theta_az/delta_theta_az) x (theta_el/delta_theta_el)
4D imaging radar: ~1,000-10,000 points per frame at 20-30 Hz
LiDAR: ~100,000-300,000 points per frame at 10-20 Hz
Radar advantage: velocity accuracy ~ 0.1 m/s per point

Common Questions

Frequently Asked Questions

Will 4D imaging radar replace LiDAR?

4D imaging radar is unlikely to fully replace LiDAR for Level 4/5 autonomy due to its coarser angular resolution (1-2 degrees vs. 0.05-0.1 degrees for LiDAR). However, it provides a strong complementary sensor and may replace LiDAR in cost-sensitive applications (Level 2+ and Level 3). Tesla's camera-and-radar approach, and several Chinese OEMs' adoption of 4D imaging radar without LiDAR, suggest a viable path without LiDAR for some autonomy levels.

What is the cost difference between 4D imaging radar and LiDAR?

Current production 4D imaging radar modules cost $50-150, compared to $500-3,000+ for automotive LiDAR (2024 prices). This 5-20x cost advantage makes 4D imaging radar attractive for mass-market vehicles. Both technologies are trending downward in cost, but radar is expected to maintain a significant cost advantage due to mature semiconductor fabrication.

Can 4D imaging radar detect lane markings?

Yes, emerging 4D imaging radar systems can detect lane markings through their electromagnetic reflectivity contrast with the road surface. At 77 GHz, painted lane markings have a different backscatter coefficient than asphalt, which can be detected in the radar's reflectivity map. This capability is still developing and works best in specific conditions (dry road, fresh markings).

Keep Reading