ALLPCB
Introduction
Within a sensor network, radar—especially millimeter-wave radar—is becoming a primary sensing modality for many intelligent systems. From predicting risks for autonomous driving to detecting presence in smart appliances, these systems rely on two key capabilities: range resolution along the line of sight and angular resolution across the field of view. These two metrics determine how well a radar can separate closely spaced targets in complex environments.
Range Resolution
Imagine two cars travelling in the same lane, one behind the other. Whether the radar recognizes them as two distinct targets rather than a single blurred return depends on its range resolution. Range resolution is the minimum separation along the line of sight at which two targets can be distinguished.
The underlying physics is simple: resolution is determined by the bandwidth of the transmitted signal. The relation ΔR = c / (2B) (c is the speed of light, B is bandwidth) shows that wider bandwidth yields finer resolution (smaller ΔR). For example, in the automotive 77 GHz band, a radar with 4 GHz bandwidth has a theoretical range resolution of about 3.75 cm. That level of resolution can distinguish vehicles from each other and even resolve limbs on a pedestrian, providing centimeter-level input for automatic emergency braking systems. By contrast, earlier 24 GHz radars with narrower bandwidth could have range resolution on the order of 60 cm.
It is important to distinguish range resolution from range accuracy. Resolution relates to the ability to separate two objects: a radar that can resolve two targets 5 cm apart is superior to one that can only resolve 10 cm. Accuracy refers to how close a measured distance is to the true distance: for a 1.00 m target, a measured 0.99 m corresponds to 1 cm accuracy. Modern high-integration millimeter-wave radar chips can achieve centimeter-level range resolution without hardware changes. In point cloud displays, a human target appears as a compact cluster of detection points, and nearby objects are clearly separated.
Angular Resolution
When two cars drive side by side in adjacent lanes at similar ranges to the radar, angular resolution is needed to distinguish them. Angular resolution is the minimum angular separation at the same range that the radar can resolve.
This capability depends on the radar's antenna array. Reflections arriving at different antennas have slight path differences that convert into phase differences. By decoding these phase differences, the radar estimates the target azimuth. Theoretical angular resolution follows roughly Δθ ≈ λ / (L * cosθ), where a shorter wavelength λ and a larger effective aperture L improve resolution.
Therefore, the core ways to improve angular resolution are:
- Increase the number of antenna elements to enlarge the effective receive aperture.
- Use higher frequencies (for example, 77 GHz versus 24 GHz) to obtain shorter wavelength.
- Use MIMO (multiple-input multiple-output) techniques. With careful transmit and receive timing, a relatively small number of physical antennas can synthesize many virtual antennas, significantly increasing resolution. For example, a 4-transmit, 4-receive physical array can form 16 virtual channels.
In real products, angular resolution varies widely by design target: high-performance 4D imaging radars can reach about 1 degree, while some basic models may be in the 15°–30° range. Angular accuracy, the precision of a single target angle measurement, is typically within ±1° near the boresight for higher-end systems.
Algorithmic Enhancements
Radar perception goes beyond antenna count and bandwidth. Advanced signal processing can extend performance beyond classical limits.
Doppler separation: Even if two targets overlap in range and angle, different radial velocities produce distinct Doppler shifts. Radar can separate such targets using Doppler information, which is highly valuable in traffic scenarios.
Beamforming and adaptive algorithms: Beyond standard FFT-based beamforming, advanced algorithms such as Capon and MVDR can nearly double angular resolution on the same antenna hardware, producing denser and clearer point clouds.
System-level approaches: Industry practices like multi-chip cascades and synthetic aperture techniques further push resolution while balancing cost and form factor.
Comparison with Cameras and LiDAR
Compared to cameras, radar offers direct distance and velocity measurements and is far less affected by lighting and many weather conditions. Cameras provide high angular resolution but do not directly measure range or speed. Compared to lidar, lidar provides very high angular and range resolution for detailed 3D point clouds but at higher cost and with degraded performance in rain, snow, or fog. Millimeter-wave radar tends to offer lower cost, better environmental robustness, and direct velocity sensing.
The core advantage of radar is fused sensing: it provides simultaneous range, angle, and velocity information. This multidimensional data supports reliable detection and tracking in many applications.
Conclusion
Range resolution and angular resolution together form the effective measuring scales of millimeter-wave radar. From distinguishing adjacent lanes in autonomous driving to detecting small motions in smart home devices, the precision of these metrics directly determines how clearly an intelligent system understands its environment. For further reading, see the articles "预告 多目标追踪24G雷达传感" and "开启指尖配置智能感知新时代".
