Fixes lidar pattern horizontal resolution bug

[pascal-roth]

Description

This PR fixes a bug in the lidar pattern horizontal angle calculation and enhances the test suite for ray caster patterns.

Bug Fix: The lidar pattern was generating incorrect number of horizontal angles, causing the actual angular resolution to differ from the requested resolution. For example, requesting 90° resolution for a 360° FOV produced only 3 rays (120° spacing) instead of 4 rays (90° spacing)

Test Enhancements:

• Added comprehensive parameterized tests to verify the fix
• Parameterized all tests over both CUDA and CPU devices
• Consolidated redundant tests (reduced from 24 to 18 test functions while maintaining coverage)
• Improved test efficiency with batched operations

Fixes #4430

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality - enhanced test suite)

Checklist

• I have read and understood the contribution guidelines
• I have run the pre-commit checks with ./isaaclab.sh --format
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• I have updated the changelog and the corresponding version in the extension's config/extension.toml file
• I have added my name to the CONTRIBUTORS.md or my name already exists there

[greptile-apps]

Greptile Overview

Greptile Summary

Fixed a critical bug in lidar_pattern where the horizontal angle calculation produced incorrect angular spacing. The issue was that torch.linspace(start, end, n) creates n-1 intervals, so requesting 90° resolution for 360° FOV produced only 3 rays with 120° spacing instead of 4 rays with 90° spacing.

Key Changes:

• Added + 1 to num_horizontal_angles calculation in patterns.py:157-159 to account for linspace interval behavior
• Added comprehensive test suite (test_ray_caster_patterns.py) with:
  • Parameterized test test_lidar_pattern_exact_angles specifically validating angular spacing matches requested resolution
  • Device parameterization (CUDA/CPU) across all tests via pytest fixture
  • Tests for grid, lidar, bpearl, and pinhole camera patterns
  • Validation of ray count, angular spacing, normalization, and wraparound behavior

The fix is mathematically correct and thoroughly tested.

Confidence Score: 5/5

• This PR is safe to merge with no risk
• The change is a minimal, well-understood mathematical fix (+1 to a calculation) that addresses a specific bug. The comprehensive test suite validates the fix across multiple scenarios and devices, and the change has no breaking implications.
• No files require special attention

Important Files Changed

Filename	Overview
source/isaaclab/isaaclab/sensors/ray_caster/patterns/patterns.py	Fixed critical horizontal angle calculation bug by adding +1 to num_horizontal_angles to ensure correct angular resolution
source/isaaclab/test/sensors/test_ray_caster_patterns.py	Added comprehensive test suite with parameterized tests validating the fix, including angular spacing verification and device coverage

Sequence Diagram

sequenceDiagram
    participant User
    participant LidarPatternCfg
    participant lidar_pattern
    participant torch
    participant RayCaster

    User->>LidarPatternCfg: Create config with horizontal_res=90°, FOV=360°
    User->>lidar_pattern: Call lidar_pattern(cfg, device)
    
    lidar_pattern->>lidar_pattern: Create vertical_angles via linspace
    
    lidar_pattern->>lidar_pattern: Check if 360° FOV (line 151-154)
    Note over lidar_pattern: If 360°: up_to = -1<br/>Else: up_to = None
    
    lidar_pattern->>lidar_pattern: Calculate num_horizontal_angles (FIX)
    Note over lidar_pattern: OLD: ceil(360/90) = 4<br/>NEW: ceil(360/90) + 1 = 5
    
    lidar_pattern->>torch: linspace(start, end, num_horizontal_angles)
    Note over torch: Creates 5 points: [-180°, -90°, 0°, 90°, 180°]
    
    lidar_pattern->>lidar_pattern: Apply [:up_to] slicing
    Note over lidar_pattern: For 360° FOV: removes last point<br/>Result: [-180°, -90°, 0°, 90°]<br/>Spacing: 90° ✓
    
    lidar_pattern->>lidar_pattern: Convert to radians and create meshgrid
    lidar_pattern->>lidar_pattern: Spherical to Cartesian conversion
    lidar_pattern->>RayCaster: Return (ray_starts, ray_directions)
    
    Note over User,RayCaster: Result: Correct 90° angular spacing<br/>4 rays instead of 3 (120° spacing)

Loading

[greptile-apps]

_{No files reviewed, no comments}

_{Edit Code Review Agent Settings | Greptile}

[tobiabir]

A very clean fix. Thank you @pascal-roth. This does indeed solve #4430.

nit: Long-term, I still think it would be nice to have consistency in the configuration between horizontal and vertical (i.e. number of channels vs resolution).