Event-Triggered Diffusion Kalman Filters (1711.00493v5)

Abstract: Distributed state estimation strongly depends on collaborative signal processing, which often requires excessive communication and computation to be executed on resource-constrained sensor nodes. To address this problem, we propose an event-triggered diffusion Kalman filter, which collects measurements and exchanges messages between nodes based on a local signal indicating the estimation error. On this basis, we develop an energy-aware state estimation algorithm that regulates the resource consumption in wireless networks and ensures the effectiveness of every consumed resource. The proposed algorithm does not require the nodes to share its local covariance matrices, and thereby allows considerably reducing the number of transmission messages. To confirm its efficiency, we apply the proposed algorithm to the distributed simultaneous localization and time synchronization problem and evaluate it on a physical testbed of a mobile quadrotor node and stationary custom ultra-wideband wireless devices. The obtained experimental results indicate that the proposed algorithm allows saving 86% of the communication overhead associated with the original diffusion Kalman filter while causing deterioration of performance by 16% only. We make the Matlab code and the real testing data available online.

• The paper introduces an event-triggered mechanism to reduce communication overhead in diffusion Kalman filters for wireless sensor networks.
• It achieves an 86% reduction in communication with only a 16% decline in localization accuracy on a mobile quadrotor testbed.
• The method eliminates the need for exchanging local covariance matrices, ensuring unbiased estimation and enhanced resource efficiency.

Overview of Event-Triggered Diffusion Kalman Filters

In the domain of wireless sensor networks, the challenge of distributed state estimation is compounded by the constraints of limited computational resources, communication bandwidth, and energy efficiency. Traditional distributed Kalman filters, particularly those utilizing diffusion strategies, require extensive message exchanges and computational efforts, which may not be sustainable for resource-constrained sensor nodes. This paper presents an innovative approach to address these challenges by introducing the event-triggered diffusion Kalman filter (ET-DKF).

Key Contributions

The ET-DKF leverages an event-triggered mechanism to regulate communication and computation based on a predefined local error signal. The algorithm aims to optimize resource usage without significantly compromising estimation performance. A notable aspect of the proposed solution is its ability to operate without necessitating the exchange of local covariance matrices among nodes, thus reducing the communication overhead significantly. The paper outlines the application of ET-DKF to solve the distributed simultaneous localization and time synchronization (D-SLATS) problem, demonstrating a high degree of efficiency and effectiveness in practical scenarios.

Experimental Setup and Results

The proposed ET-DKF was tested using a physical testbed comprising a mobile quadrotor and custom ultra-wideband (UWB) wireless devices. The experiments demonstrated that by adopting the ET-DKF, approximately 86% of the communication overhead observed in standard diffusion Kalman filters can be saved, with only a minor deterioration of 16% in performance in terms of localization accuracy.

Theoretical Insights and Practical Implications

The theoretical analysis within the paper proves that the ET-DKF remains an unbiased estimator under the considered conditions and highlights the intricate relationship between the global and local error covariance matrices. The innovative design ensures that local estimates account for neighborhood measurements without direct covariance matrix exchange, preserving network resources.

Practically, the ET-DKF offers substantial improvements for applications such as interconnected IoT devices, where maintaining an efficient yet effective estimation process is paramount. By dynamically triggering measurement and information exchange events based on local error metrics, systems can achieve desired accuracy levels while considerably conserving energy and reducing communication load.

Conclusion and Future Directions

The ET-DKF represents a significant step towards developing energy-efficient and computationally feasible state estimation techniques for distributed systems. Future research may explore exploring adaptive thresholds for triggering events, expanding the application scope across more complex dynamic environments, and further optimizing network topology for more comprehensive resource savings. The availability of the Matlab code and real testing data via the provided GitHub repository enhances transparency and encourages further experimentation and validation by the research community.

In summary, the ET-DKF provides a promising alternative to conventional decentralized estimation methods, striking a balance between performance integrity and resource limitations, which is crucial for the advancement of sensor network applications.