CARLA ? C-Shenron based Simualtor for Sensor data generation.
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 
 

2.2 KiB

Shenron Antenna Gain & Compensation Logic

📡 Azimuth & Elevation Gain Integration

The latest iteration implements a separable antenna pattern model. This ensures that targets detected at the edges of the radar's Field of View (FOV) are correctly attenuated, matching real-world hardware behavior.

1. Implementation Details (Sceneset.py)

In get_loss_3(), the antenna gain $G_{ant}$ is calculated as the product of horizontal and vertical patterns: $$G_{ant}(\theta, \phi) = G_{az}(\theta) \times G_{el}(\phi)$$

  • Azimuth ($\theta$): Uses a Look-Up Table (LUT) for symmetric gain interpolation. Points outside the precise LUT range are clamped to the nearest edge value.
  • Elevation ($\phi$): Implements a Hard FOV Cutoff. Points beyond the mechanical vertical beamwidth receive zero gain, accurately simulating physical sensor limitations.
  • Physics Order: Gain is applied to the incident power after the path loss but before the material interaction, ensuring that the antenna pattern does not affect target reflectivity.

2. High-Level System Gain (gain = 110 dB)

In ConfigureRadar.py, the parameter self.gain is set to values like 10 ** (110 / 10).

Why 110 dB? This is a System-Level Calibration Constant that encapsulates several physical and simulation factors:

  1. Transmit Power ($P_t$): The actual chirping power emitted (typically ~10-15 dBm).
  2. Peak Antenna Gain ($G_t, G_r$): The boresight gain of the TX and RX patches.
  3. Signal Chain Amplification: Hardware-specific gains in the LNA and Mixer.
  4. Simulation Scaling: In physics-based simulations, units computed for unit scatterers ($1/R^2$) result in extremely small floating-point values. The 110 dB factor "lifts" these signals into a dynamic range that maps correctly to real 16-bit or 32-bit ADC data formats.

By tuning this single "dial," the simulation can be calibrated to match the peak Signal-to-Noise Ratio (SNR) seen in real-world hardware datasets.

⚙️ Calibration Flow

  • Standard Baseline: 110 dB (Calibrated for Iteration 16/18 physics).
  • Match Hardware: Adjust this dial ±5 dB if targets at 10m in simulation appear significantly stronger/weaker than in real sensor captures.