Federated Data-Driven Kalman Filtering for State Estimation

2024-10-02

IEEE International Workshop on Multimedia Signal Processing

AI Summary

This work addresses the high communication overhead and poor robustness of real-time V2X-dependent approaches in high-precision autonomous vehicle localization. We propose FedKalmanNet—the first distributed Kalman filtering framework integrating federated learning (FL) with KalmanNet. Adopting a “adapt-then-aggregate” paradigm, it enables inter-vehicle collaborative training of data-driven uncertainty estimation (e.g., Kalman gains) without sharing raw sensor data or real-time measurements, thereby achieving V2X-free state estimation. Our key innovation lies in the first incorporation of a recurrent neural network (RNN) into FL to model the time-varying parameters of an extended Kalman filter (EKF), and we empirically demonstrate that collaborative estimation—unlike collaborative decision-making—achieves superior localization accuracy with significantly reduced communication cost. Evaluations in CARLA show a 32% reduction in localization error and a 90% decrease in communication overhead.

Technology Category

• Intelligent Robots: State Estimation
• Machine Learning: Distributed Machine Learning & Federated Learning
• Computer Vision: Generative Adversarial Networks (GANs) for Vision

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

This paper proposes a novel localization framework based on collaborative training or federated learning paradigm, for highly accurate localization of autonomous vehicles. More specifically, we build on the standard approach of KalmanNet, a recurrent neural network aiming to estimate the underlying system uncertainty of traditional Extended Kalman Filtering, and reformulate it by the adapt-then-combine concept to FedKalmanNet. The latter is trained in a distributed manner by a group of vehicles (or clients), with local training datasets consisting of vehicular location and velocity measurements, through a global server aggregation operation. The FedKalmanNet is then used by each vehicle to localize itself, by estimating the associated system uncertainty matrices (i.e, Kalman gain). Our aim is to actually demonstrate the benefits of collaborative training for state estimation in autonomous driving, over collaborative decision-making which requires rich V2X communication resources for measurement exchange and sensor fusion under real-time constraints. An extensive experimental and evaluation study conducted in CARLA autonomous driving simulator highlights the superior performance of FedKalmanNet over state-of-the-art collaborative decision-making approaches, in localizing vehicles without the need of real-time V2X communication.

Problem

• Federated learning for autonomous vehicle localization
• Improving state estimation with FedKalmanNet
• Reducing V2X communication dependencies in localization

Innovation

• Federated learning for localization
• Adapt-then-combine FedKalmanNet
• Distributed training without V2X