Metaheuristic for Optimal Dynamic K-Coloring Application on Band Sharing for Automotive Radars.

Area of Science:

• Automotive Engineering
• Signal Processing
• Artificial Intelligence

Background:

• Vehicle-based radar systems are rapidly increasing, projected to be in 50% of cars by 2030.
• Current radar specifications lack mandatory waveform and channel access policies, increasing interference risks.
• Effective interference mitigation is crucial for reliable radar operation and Advanced Driver-Assistance Systems (ADAS).

Purpose of the Study:

• To develop an optimal resource-sharing strategy for vehicular radars to minimize harmful interference.
• To reduce the frequency of resource changes required by radars while maintaining interference minimization.
• To explore the utility of a metaheuristic approach for interference management in complex radar environments.

Main Methods:

• A centralized metaheuristic algorithm was developed to optimize time-frequency resource allocation among radars.
• The algorithm considers the relative positions of vehicles to assess line-of-sight (LOS) and non-line-of-sight (NLOS) interference risks.
• The approach uses known past and future vehicle positions for system-wide optimization.

Main Results:

• The metaheuristic effectively minimizes radar interference by organizing the radar band into non-interfering time-frequency resources.
• The study demonstrates a method to balance interference reduction with the number of radar resource reconfigurations.
• The proposed algorithm provides near-optimal solutions suitable for simulations and machine learning data generation.

Conclusions:

• Organizing radar resources into non-interfering time-frequency blocks significantly reduces interference.
• A metaheuristic approach offers an effective, albeit computationally intensive, method for optimizing radar resource sharing.
• This work provides a foundation for developing robust interference mitigation strategies for future automotive radar systems.