Performance Evaluation of the Highway Radar Occupancy Grid.

Area of Science:

• Robotics and Autonomous Systems
• Environmental Perception for Vehicles
• Sensor Fusion and Mapping

Background:

• Accurate environmental perception is vital for automated vehicle safety and functionality.
• Occupancy grid mapping algorithms are commonly used to fuse sensor data for environmental representation.
• Evaluating the quality of these maps, especially radar-based ones for highways, is critical.

Purpose of the Study:

• To develop and describe an effective method for measuring the quality of radar-based occupancy grid maps.
• To focus specifically on highway driving scenarios.
• To introduce a versatile assessment technique applicable without detailed ground truth.

Main Methods:

• Utilized a novel grid pole-like object analysis method for quality assessment.
• Evaluated the proposed assessment method using both simulation and real-world highway experiments.
• Focused on radar sensor data for occupancy grid generation.

Main Results:

• The grid pole-like object analysis method provides an effective means to measure occupancy map quality.
• The assessment is versatile and does not require extensive ground truth information.
• Validation through simulation and real vehicle highway tests confirmed the method's applicability.

Conclusions:

• The proposed quality assessment method is suitable for radar-based occupancy grids in highway applications.
• This technique enhances the reliability of environmental perception for automated driving systems.
• The versatility of the method allows for broader application in autonomous vehicle development and testing.