Distributed Clustering and Tracking Algorithm for Trustworthy MIMO Radars

As radars are being increasingly used in autonomous driving, it is important to ensure that the results delivered by the radar sensors are trustworthy. Clustering and tracking of targets is part of the signal processing in such radar systems. Widely used system-on-chip (SoC) radars have vulnerabilities that affect trustworthiness. Therefore, a newly distributed architecture for clustering and tracking algorithms is introduced in this article, which can be implemented using multiple application-specific integrated circuits (ASICs). This architecture improves the protection of intellectual property (IP) related to the algorithms and improves the functional integrity. An adversary with access to a single ASIC of the proposed system cannot understand the system’s complete functionality with limited information stored in it. Further redundancy in tracking and a method to estimate the Jaccard Index of clusters are proposed to identify failures. Additionally, features—MIMO phase hopping based on phase coding of the Tx and randomly ordered clustering—are introduced to help identify potential data manipulation or failures and minimize the compromise on the delivered results. The effectiveness of the proposed distributed algorithm against the manipulation of certain information is demonstrated with simulated data.